<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Ed Khristus on Medium]]></title>
        <description><![CDATA[Stories by Ed Khristus on Medium]]></description>
        <link>https://medium.com/@edkh?source=rss-d7b2805cd5b0------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*mMKmjlBl1wB2pbcoLc8Vag.jpeg</url>
            <title>Stories by Ed Khristus on Medium</title>
            <link>https://medium.com/@edkh?source=rss-d7b2805cd5b0------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 17:35:35 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@edkh/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Which Decisions Should Stay in Place After Annual Leave?]]></title>
            <link>https://medium.com/@edkh/which-decisions-should-stay-in-place-after-annual-leave-3fb0ee73d91e?source=rss-d7b2805cd5b0------2</link>
            <guid isPermaLink="false">https://medium.com/p/3fb0ee73d91e</guid>
            <category><![CDATA[management]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[workplace]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Ed Khristus]]></dc:creator>
            <pubDate>Mon, 16 Nov 2026 09:20:44 GMT</pubDate>
            <atom:updated>2026-11-16T09:20:44.222Z</atom:updated>
            <content:encoded><![CDATA[<h4>When cover ends, sort customer promises and open decisions before handing the work back to its usual owner.</h4><figure><img alt="A timeline shows Adrian agreeing two deliveries during cover and Priya returning to an old sheet that lists Friday only." src="https://cdn-images-1.medium.com/max/1024/1*Q1Fsvh2Empw3G0FkZOZZTQ.png" /><figcaption>At a fictional ceramics studio, Adrian agrees two deliveries with the hotel while Priya is away. Priya returns to a sheet that still lists only Friday.</figcaption></figure><p>At a fictional ceramics studio, Priya comes back from annual leave and opens the order for twelve blue serving bowls. Before she left, the plan was to send all twelve to a hotel on Friday. The warehouse board now says six on Thursday and six on Friday. Priya still has her old packing sheet open. Which version should she give the team?</p><p>Adrian covered wholesale orders while Priya was away. He offered to send the first six bowls early, the hotel accepted, and its kitchen lead moved a prep shift. Priya takes charge of wholesale orders again at nine on Monday. The kitchen lead is still expecting a crate on Thursday.</p><p>When you run the return handover, keep commitments made within the cover arrangement. Close temporary instructions once their remaining work has moved. If a choice still needs an answer, give it to someone who can act before the deadline. The first thing to agree is when ordinary cover ends.</p><h3>Find out who is already working to the new plan</h3><p>Adrian used the authority the studio gave him to split an order into two shipments. The useful thing for Priya to see is the hotel’s acceptance beside the warehouse instructions. Had Adrian sent an offer and received no answer, Priya would need to know who was following up and when the warehouse needed a decision. On this order there is an answer, and the hotel has already scheduled around it.</p><p>Start with the decision that changes what someone else will do today. Ask the cover colleague where the agreement lives and which people have acted on it. Put the agreement beside the order the packing lead works from. Priya can mark her old sheet superseded so a saved copy doesn’t send all twelve bowls to Friday’s loading bench.</p><p>Give the returning owner enough reason to handle the next call. Priya needs to know that the hotel wanted the first six bowls for Thursday prep. If a kiln fault threatens that batch, she can use the normal route to agree a change with the hotel. She does not need to read Adrian’s whole week of messages to know why sending everything on Friday would cause a problem.</p><h3>Retire the instruction that belonged to cover</h3><p>During the absence, Adrian asked the studio to set new wholesale requests in a red tray. He worked on the floor in the mornings and checked the tray at two. Priya normally pulls requests straight into the wholesale queue. If everybody keeps using both routes, a request copied into the queue can still be picked up from the tray.</p><p>Before taking the tray away, Adrian and Priya look inside it. There’s one sample request left. They move it to the queue, tell the studio where the next request goes and remove the sign by the tray. Simply announcing that Priya is back would leave the paper on somebody else’s desk with no reason to move.</p><p>Before removing a workaround, ask who still expects to find work there. For help withdrawing the old direction, see the <a href="https://cooperly.ai/insights/conflicting-instructions-two-managers?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=founder_operating_systems&amp;utm_content=leave_cover_decisions">Cooperly guide to replacing a conflicting instruction</a>.</p><h3>Cover the decisions due before the usual owner is ready</h3><p>There is a second wholesale enquiry. A restaurant wants to collect a set of plates on Wednesday. Adrian has checked stock, and the customer expects a reply by noon on Monday. Priya takes wholesale back at nine, but she is booked on a kiln inspection and a supplier visit until two thirty. If the restaurant enquiry transfers with everything else at nine, nobody on the order desk has the authority to answer in time.</p><p>Even noon may be too late. In this example the packing lead needs the call by eleven thirty to keep the collection slot open. Ask the person doing the downstream work when a delay closes an option. The <a href="https://cooperly.ai/insights/decision-timing-leadership?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=founder_operating_systems&amp;utm_content=leave_cover_decisions">Cooperly guide to finding the real decision deadline</a> can help if the deadline on the request and the one in the work differ.</p><p>Before nine, the studio lead agrees that Adrian keeps authority over this one enquiry until he has logged the reply, which is due by noon. The lead checks that Adrian has time before the packing cutoff and can still reserve the plates. If the lead cannot extend that authority, someone who already holds it must make the call or agree a later answer with the restaurant. If the packing lead releases the slot at eleven thirty, the next message needs a different collection option.</p><h3>What to include in a return handover</h3><p>Put the note where the next person will look for instructions. Priya and Adrian have three pieces of work across two records, so each item gets a line in the record where it will be acted on. Put the authority change beside the live queue so anyone taking a new wholesale request after nine knows where to send it. The studio can collect those lines in a handback note for Priya to read between jobs. The warehouse and the order desk can use their own records to carry out the same instructions.</p><blockquote><strong><em>Authority.</em></strong><em> Priya resumes wholesale orders at 9:00 Monday. Adrian retains authority for the restaurant enquiry only and logs his reply by noon.</em></blockquote><blockquote><strong><em>Keep.</em></strong><em> Six blue bowls Thursday, six Friday. Hotel confirmed the split and moved Thursday prep. Use the dates on the warehouse order.</em></blockquote><blockquote><strong><em>Retire.</em></strong><em> Red tray closes at handback. Sample request is in the wholesale queue. Put new requests there.</em></blockquote><blockquote><strong><em>Open.</em></strong><em> Restaurant collection is undecided. Adrian makes the call by 11:30 and sends and logs his reply by noon.</em></blockquote><figure><img alt="Three cards show the handback decisions: keep the Thursday and Friday bowl deliveries, close the red tray, and give Adrian the collection decision due by 11:30." src="https://cdn-images-1.medium.com/max/1024/1*wIsqu5NjWMbH633xmxxKvA.png" /><figcaption>The studio carries forward the split delivery, closes the temporary request tray and assigns the still-open collection call to Adrian.</figcaption></figure><p>Write the exception before nine. The packing lead needs to see it before the morning cutoff. At eleven thirty the lead can check whether to hold the Wednesday slot, and by noon Priya has a reply to pick up when she is next at the desk. If a choice is still open when its owner changes, put that status beside the current stock or delivery information. If the queue turns an enquiry into an order as soon as someone types a date, leave Wednesday in the enquiry notes until Adrian has actually confirmed it.</p><h3>Use the next return to test one active job</h3><p>Pick an order that will move on the day someone comes back. Read the record the person doing the next step actually sees. If it says “back to normal,” replace that with the delivery they should work to or the decision they should wait for. Ask the person who is about to print the label which date they see. If the work still relies on a call to the founder, check whether the dates, the owner or the source of the agreement is missing.</p><p>Cooperly is the Human Operating System for founder-led, AI-native companies. A handback note makes the current agreement available to the people who pack, schedule or reply. Each of them can use the same instruction when the founder is elsewhere. If you need to set up the cover itself before the next break, use <a href="https://cooperly.ai/insights/employee-wont-take-time-off?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=founder_operating_systems&amp;utm_content=leave_cover_decisions">Cooperly’s guide to agreeing who can decide during leave cover</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3fb0ee73d91e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Set Standards When Two Teams Share One Specialist]]></title>
            <link>https://medium.com/@edkh/how-to-set-standards-when-two-teams-share-one-specialist-702a19349ff7?source=rss-d7b2805cd5b0------2</link>
            <guid isPermaLink="false">https://medium.com/p/702a19349ff7</guid>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[workplace]]></category>
            <category><![CDATA[management]]></category>
            <dc:creator><![CDATA[Ed Khristus]]></dc:creator>
            <pubDate>Sun, 08 Nov 2026 12:37:21 GMT</pubDate>
            <atom:updated>2026-11-08T12:37:21.534Z</atom:updated>
            <content:encoded><![CDATA[<h4>Give the specialist authority over what both teams need to share. Let each lead choose how their team does the work.</h4><figure><img alt="Two packing sheets. Who sets the rule? Felix’s events sheet lists one row per title and Serena’s subscriptions sheet lists one per edition. The founder faces a choice about who sets the shared rule." src="https://cdn-images-1.medium.com/max/1024/1*xcfSAqlmnjYXkVJwSfmtJQ.png" /><figcaption>At a fictional bookseller, two packing sheets count stock differently. The founder needs to decide who can set the shared rule.</figcaption></figure><p>Picture a small bookseller with two teams returning unsold copies of <em>Harbour Birds</em>. Felix’s events sheet gives one total for the title. Serena’s subscriptions sheet gives a count for each edition. Together they’ve returned twenty copies. The stock coordinator can add the sheets up, but she can’t tell how many hardbacks she can promise.</p><p>Serena wants both teams to use her sheet. Felix wants to keep his arranged in the order people pack the stall. Both ask Priya, the inventory specialist, to settle it. If she picks a winner, one crew inherits a layout designed for the other’s work. If she leaves the sheets as they are, the coordinator still has to ask which books came back.</p><p>Give Priya authority over the information the shared stock record needs. Let each lead arrange their sheet for the people using it at the table. When two teams disagree about a routine standard, start with the part of the work someone else actually has to use.</p><h3>Find the point where the work comes together</h3><p>Start with the two sheets and ask the stock coordinator to do her next job. She wants to add the books left after Saturday’s event to the available stock, then reserve six hardbacks for Monday’s subscription boxes. She needs an edition-specific stock code and a count for each code, recorded in individual copies. A combined count for the title can’t tell her whether she has enough of the right edition.</p><p>The events sheet lists books in stall order so the crew can find what they’re packing. The subscriptions sheet groups them by box. Those layouts solve different problems for the people holding the sheets. The stock coordinator can still enter the returns if she gets a separate count for each edition. The order of the rows makes no difference to her. The meaning of a row does.</p><p>Walk through the task with its actual recipient before deciding what must match. Ask what they enter, compare or promise from the output. Notice the places they have to call the sender. A question about whether a number represents books or bundles points to something worth specifying. A question about why the historical fiction section is at the bottom might have no bearing on stock at all.</p><h3>Who owns the shared standard?</h3><p>Priya can decide which fields both teams submit and how to identify an edition. Each lead can decide the order of the rows, where to put packing notes and how to brief their crew. If Felix wants to change how a return count is recorded, that belongs with Priya. If he wants to move the packing notes to the left, his team can try it that afternoon.</p><p>Check whether a new starter could tell what an entry means. If a sheet says “3” beside a carton, the coordinator needs to know whether that means three copies or three cartons. Specify the unit on the sheet itself, where the person doing the count can see it.</p><p>Put the split on the sheet the coordinator receives. Two lines are enough for this example.</p><blockquote><em>Shared return fields. Priya owns the edition-specific stock code and the count in individual copies. Send one count for each edition returned.</em></blockquote><blockquote><em>Local packing layout. Each team lead owns row order, packing notes and the layout their crew uses.</em></blockquote><figure><img alt="One shared rule. Two local layouts. The founder’s You mark connects to Priya, who owns edition code and count, then to an agreement giving the leads control of packing layouts." src="https://cdn-images-1.medium.com/max/1024/1*kiKwLID_s62J3XXfeJND-g.png" /><figcaption>The founder gives Priya authority over the shared stock fields. Felix and Serena keep control of their teams’ packing layouts.</figcaption></figure><p>Put a link to those lines next to the submission instructions. A new events lead should be able to find the rule before the first Saturday shift. Cooperly’s guide to <a href="https://cooperly.ai/insights/managing-former-peers-first-conversation?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=founder_operating_systems&amp;utm_content=shared_specialist_quality_standard">making a new remit work when you manage former peers</a> starts with a similar problem. An agreement needs to hold when the person asking wasn’t there for the original conversation.</p><h3>Try both layouts with a return you can count</h3><p>Give the teams different dummy returns, all in saleable condition. Events has five hardbacks and five paperbacks. Subscriptions has three hardbacks and seven paperbacks. Ask each team to submit its numbers in its own layout, then have the coordinator enter them in a copy of the ledger. She should be able to trace five hardbacks to events and three to subscriptions and reserve six from the eight. If she only gives you the total, a swapped entry could still pass the test.</p><p>If one team enters ten books under the title, check where the edition disappeared. Perhaps the packing sheet has only one line for the title, so the person counting can’t record a second figure. Priya can ask for two rows or a separate return slip that preserves the counts. The crew can still pack the stall in the order that works for them. Run the dummy return again before anyone copies the change into the live ledger.</p><p>Cooperly’s guide to <a href="https://cooperly.ai/insights/decision-timing-leadership?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=founder_operating_systems&amp;utm_content=shared_specialist_quality_standard">testing a decision before committing</a> suggests shrinking the choice into something reversible. Here, a small return can tell you whether the coordinator can use the output. Test stock that can’t go back on sale separately. The coordinator may need both teams to distinguish those copies on the return even if only one crew regularly handles them.</p><h3>Widen the rule when another crew needs to use it</h3><p>Suppose an events assistant starts covering subscription packing on Mondays. On her first shift she can read the edition codes but misses a handwritten note that changes what goes into a box. The return fields still work. A different part of the work now crosses the team boundary.</p><p>Before asking Priya to standardise both sheets, find out what the assistant needs at the packing table. If it’s a prominent place for an instruction that changes the contents of a box, the lead can try a labelled box on the subscription sheet. If assistants start moving both ways and miss the equivalent instruction at events, ask both leads and a person who covers shifts to try the same label for those instructions. Change where the note sits only if people still miss it.</p><p>Have the operating lead name who can set the shared packing convention if the trial shows both crews need it. In this case, Serena could own the wording and placement because she trains the assistants who cover packing. Priya still owns the stock fields. If the crews can make the change in their usual sheet updates, leave the timing with them. If a deadline requires one team to drop something it has already promised, the operating lead can decide what moves.</p><h3>Take one approval out of the next handoff</h3><p>Pick an output your teams will share this week. Write one line for the information its recipient needs and who can change that rule. Write a second for the choices each team can make and who owns them. Ask the recipient to do their next task using both submissions. Once you know where the split holds, take one approval off your plate. If Felix asks you to approve moving his packing notes, send the choice back to his crew as long as the coordinator still gets the fields she needs.</p><p>If you’re setting up a role that will make decisions for several teams, use <a href="https://cooperly.ai/insights/individual-contributor-career-path?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=founder_operating_systems&amp;utm_content=shared_specialist_quality_standard">Cooperly’s guide to defining a senior specialist’s remit</a> to put those boundaries into the job.</p><p><em>Cooperly is the Human Operating System for founder-led, AI-native companies.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=702a19349ff7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Your Employee Survey Score Rose. What Can You Claim?]]></title>
            <link>https://medium.com/@edkh/your-employee-survey-score-rose-what-can-you-claim-99fb7e537f6e?source=rss-d7b2805cd5b0------2</link>
            <guid isPermaLink="false">https://medium.com/p/99fb7e537f6e</guid>
            <category><![CDATA[management]]></category>
            <category><![CDATA[employee-engagement]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[workplace]]></category>
            <dc:creator><![CDATA[Ed Khristus]]></dc:creator>
            <pubDate>Sun, 01 Nov 2026 21:50:22 GMT</pubDate>
            <atom:updated>2026-11-01T21:50:22.703Z</atom:updated>
            <content:encoded><![CDATA[<h4>A better result is a reason to inspect the change. Check whose work improved before announcing a team-wide success.</h4><p>You delivered the workload change the team asked for. The next employee survey score is higher. You start writing: “The new process has solved our workload problem.”</p><p>Pause at “solved”. You have evidence that an action happened and that respondents gave a better answer. The connection between those facts still needs checking.</p><p>Before calling a survey action a success, compare what was asked, whose work the answers cover and what happened to the demand. Then write an update that separates the change you made from the effect you can support. A founder needs enough evidence to choose the next step; a rising score alone cannot supply the whole account.</p><p>This is a practical way to review an ordinary operating change, not a statistical test or a validated assessment.</p><figure><img alt="Cover: “The score rose. Check the claim.” In a hypothetical audio studio, improved workload answers and Serena’s fewer episode restarts connect to You. Julian’s unchecked coordination work has a dashed connection." src="https://cdn-images-1.medium.com/max/1024/1*3p-9y_cvAkn0HC016BI04A.png" /><figcaption>In the hypothetical studio, a better survey result and fewer production restarts leave the account lead’s coordination cost unresolved.</figcaption></figure><h3>Keep the question attached to the result</h3><p>Consider a hypothetical audio-production studio. Producers report that customer revision notes arrive throughout the week, repeatedly reopening episodes they thought were finished. The founder agrees a Wednesday-noon cutoff for the normal weekly revision batch. Later requests need an explicit decision about timing.</p><p>Serena, a producer, starts receiving one usable batch. Julian, the account lead, owns the customer arrangement. In the next team survey, the headline workload result improves. The founder wants to keep the cutoff and tell the company the problem is fixed.</p><p>First open the actual survey question. Did both rounds ask about manageable workload, or did the second ask whether the new revision process was clear? Those are different questions. A clearer process could still leave too much work.</p><p>Check the response scale and the period people were asked to consider. A question about this week cannot simply stand in for an earlier question about the previous month. Preserve the wording, response scale and reference period in the comparison. If any changed, describe the results separately.</p><p>Cooperly’s <a href="https://cooperly.ai/insights/employee-surveys-without-action?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=survey_score_success_claim">guide to following through on employee surveys</a> argues for a visible response to what people report. Once that response exists, there is another responsibility: tell the team what the evidence actually says about it.</p><h3>Check coverage without trying to identify respondents</h3><p>Who was invited to each round, and how much of that group responded? Use aggregate participation information the survey legitimately provides. Check whether the reported result covers the same team or whether a company-wide average has replaced a producer-only result.</p><p>Even with identical questions, the people answering may differ. A higher average could reflect changed experience, different participation, or both. A response count helps describe the limits; it cannot tell you why anyone answered or stayed silent.</p><p>Do not turn this into an attempt to reconstruct anonymous replies. In a small studio, breaking results down by role or shift may make a colleague recognisable. Use the survey’s existing privacy rules and reporting limits. Where a safe comparison is unavailable, say so and use a voluntary conversation about the work.</p><p>You can ask everyone affected whether the revision arrangement is usable. You do not need to ask who gave the low score. Keep those conversations separate from any attempt to match individual answers across rounds.</p><p>Cooperly’s guidance on <a href="https://cooperly.ai/insights/something-feels-wrong-in-team?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=survey_score_success_claim">checking an uncertain team signal</a> treats observations as a reason to investigate possible explanations. Apply the same restraint to welcome news.</p><h3>Look for the work the new arrangement created</h3><p>Return to the studio’s promise. Was the Wednesday batch actually delivered? Look at the normal revision record for recent episodes. Did producers receive usable notes by the cutoff? Were later requests given an agreed delivery decision, or did they return through direct messages?</p><p>If the arrangement never operated as intended, the survey cannot tell you whether that arrangement would help. Fix the delivery gap before treating the result as a verdict on the idea.</p><p>Suppose the batches did arrive and Serena reports fewer restarts. That is useful evidence about her production work. Now check Julian’s side. Did he spend his evenings turning scattered customer messages into the clean batch? Did routine account work wait while he did it?</p><p>Use authorised work records and his account of the task. There is no need to inspect private messages or collect a minute-by-minute activity log. You need to understand whether the receiving role had room for the responsibility the change created.</p><p>A producer’s improvement can be real while the wider arrangement still needs repair. Cooperly’s <a href="https://cooperly.ai/insights/team-overloaded-beyond-prioritisation?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=survey_score_success_claim">workload guide</a> includes capacity, flow and recovery in the operating picture. In this case, that means looking beyond the cleaner production queue to the work required to keep it clean.</p><p>Also ask what else changed. Perhaps fewer episodes were commissioned during the review period, or an unusually demanding customer paused work. Those possibilities do not disprove the cutoff’s value. They limit a claim that the cutoff caused the entire improvement. Check whether the arrangement remains workable when ordinary demand returns.</p><h3>Write the update at the scale of the evidence</h3><p>The studio can report progress before it has a complete explanation. The wording should leave the next decision visible.</p><p>Suppose the review also confirms fewer episodes during the period. Instead of “The new process has solved our workload problem”, an update could read:</p><blockquote><em>“We introduced the Wednesday revision batch. Serena reports fewer restarts on the episodes reviewed, and the latest workload answers were more positive. The review period also had fewer episodes. Account coordination still needs checking, so we are keeping the cutoff as a trial while Julian and I review the work it creates before the next full production week.”</em></blockquote><p>Every part of that example is hypothetical. In your own update, include only what you checked. If participation changed, state the aggregate limitation where it can be shared safely. If you have no work evidence yet, say that the better result is encouraging and name the check still to come.</p><figure><img alt="Illustration: “Observed progress. Cost still to check.” Serena’s fewer episode restarts and Julian’s unchecked coordination work feed an outline record that keeps the Wednesday cutoff as a trial, connected to You." src="https://cdn-images-1.medium.com/max/1024/1*_J9sw8k5g78dkr7yL65vVQ.png" /><figcaption>The hypothetical studio’s update can recognise production progress while keeping the revision cutoff under review until coordination work is checked.</figcaption></figure><p>To reuse this form, write the action delivered, the observation supporting progress and the limit that affects your next decision. Attach an owner and a realistic review point to the unresolved work. This is a proposed writing aid, not a way to calculate whether an intervention succeeded.</p><p>A bounded statement gives the team something they can correct. Serena can confirm which restarts disappeared. Julian can explain the coordination cost. “Workload is solved” leaves both people arguing against a conclusion the founder has already announced.</p><h3>Give the next review a decision to make</h3><p>Decide what the remaining evidence could change. If batching helps production and Julian has capacity, keep the arrangement and check it during a normal week. If coordination displaced necessary account work, change the scope or provide capacity before making that responsibility permanent.</p><p>If the evidence remains mixed, describe the trial as mixed. You can retain a useful part while investigating the unresolved cost. Neither a better score nor an incomplete explanation requires a company-wide success declaration.</p><p>Cooperly is the Human Operating System for founder-led, AI-native companies. Its People Context Layer brings working profiles, recurring signals, goals and team memory into the context available to approved AI tools. Access is selected, permissioned and read-only; people remain responsible for the interpretation and decision. That context can support review preparation. It does not establish what caused a score to rise.</p><p>If you need to see how that preparation fits your team’s tools and permissions, <a href="https://cooperly.ai/ai-context-layer?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=survey_score_success_claim">see Cooperly’s AI Context Layer and discuss your setup in a demo</a>. Before your next survey readout, underline the strongest success claim in your draft. Check whether the work evidence covers everyone named in that sentence.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=99fb7e537f6e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[After a Meeting Interruption, Check the Promise You Sent]]></title>
            <link>https://medium.com/@edkh/after-a-meeting-interruption-check-the-promise-you-sent-903344849801?source=rss-d7b2805cd5b0------2</link>
            <guid isPermaLink="false">https://medium.com/p/903344849801</guid>
            <category><![CDATA[management]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[communication]]></category>
            <category><![CDATA[workplace]]></category>
            <dc:creator><![CDATA[Ed Khristus]]></dc:creator>
            <pubDate>Mon, 26 Oct 2026 11:08:37 GMT</pubDate>
            <atom:updated>2026-10-26T11:08:37.068Z</atom:updated>
            <content:encoded><![CDATA[<h4>A private apology can settle the conversation while colleagues and customers still work from an unchecked commitment.</h4><p>You apologise to the colleague who was interrupted. You speak privately with the person who cut across them. Both conversations go reasonably well. Meanwhile, someone outside the meeting is arranging work around the date you agreed before hearing the missing information.</p><p>When a meeting interruption leaves a material warning unheard, follow the decision beyond the room. Recover the contribution, check what it changes, then correct the commitment wherever people are still relying on it. Coaching the interruption and repairing the instruction require different follow-through.</p><p>This is a practical approach for an ordinary work decision, not a tested method or a diagnosis of anyone’s motives. Its useful question is narrow: who could still act on the promise you now have reason to question?</p><figure><img alt="Cover: “Friday promised. Monday booked.” In a hypothetical guidebook order, Felix’s Friday collection email connects to You while Serena’s Monday printer booking has a broken connection, showing the unheard warning." src="https://cdn-images-1.medium.com/max/1024/1*y9HZ7IrQMe4zPXzL3OxHwA.png" /><figcaption>In the hypothetical guidebook order, the Friday promise travelled while the warning about Monday’s printer slot went unheard.</figcaption></figure><h3>Find the commitment that left the meeting</h3><p>Consider a hypothetical small company producing museum guidebooks. In a planning meeting, Serena, the production lead, begins explaining a problem with the printer’s availability. Felix, the sales lead, interrupts to discuss the customer’s opening event. You move the conversation along and agree that the museum can collect the guidebooks on Friday.</p><p>Felix emails the museum. The warehouse receives a collection note. Serena later tells you she never finished her point: the printer’s confirmation names the following Monday as the production slot.</p><p>You need to check the confirmation. But you also have a commitment already in circulation. The museum may be arranging its event around Friday; the warehouse has an instruction it cannot yet substantiate.</p><p>Start by writing the exact promise: finished guidebooks available for Friday collection. Find the message that communicated it and ask where it went next. A meeting note saying “printer issue discussed” will not tell you which work depends on that date.</p><p>Cooperly’s <a href="https://cooperly.ai/insights/employee-keeps-interrupting-meetings?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=leadership_communication_systems&amp;utm_content=meeting_warning_commitment_correction">guide to restoring an interrupted contribution</a> begins with getting the missing information heard. Here, that repair starts after the group has already issued an instruction. Your first task is to locate what the instruction set in motion.</p><h3>Check the warning before replacing the date</h3><p>Ask Serena to finish the point in the form that helps her explain it. Look together at the relevant printer confirmation. Does Monday refer to this order? Is it the current booking? Was another arrangement confirmed elsewhere? Felix may hold an update Serena has not seen. Let both accounts be checked.</p><p>An interrupted warning deserves a hearing. It does not automatically establish that the original decision was wrong. If a current confirmation supports Friday collection, explain that finding and make it available to the people doing the work. You still owe Serena the chance to contribute.</p><p>Suppose the check instead confirms that Monday is the current slot and no earlier arrangement is agreed. Friday collection now lacks a supported production plan. The founder who approved the promise needs to decide what can be offered and who can discuss it with the museum.</p><p>Do not turn Monday’s production slot into a new collection promise. Printing, finishing and transport may still need checking. Ask the authorised supplier contact to establish the options. Give the customer contact a clear interim instruction, including when a further answer can realistically be provided.</p><p>If someone may act before your check finishes, the authorised contact should flag that the date is under review and agree what affected preparation should pause. Give them a next-update point you can meet, even if the answer will still be incomplete. The missing contribution has changed what you know; it has not supplied every detail required for a replacement commitment.</p><h3>Send the correction to the people using the promise</h3><p>A private apology reaches Serena. It leaves Felix’s email and the warehouse’s collection note untouched. Even an accurate message to everyone who attended the meeting can miss the people now depending on the decision.</p><p>Cooperly’s article about <a href="https://cooperly.ai/insights/employee-taking-credit-for-others-work?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=leadership_communication_systems&amp;utm_content=meeting_warning_commitment_correction">correcting an inaccurate account where it mattered</a> concerns attribution. Apply that practical principle to the operating commitment: find each audience or record that still carries the information you need to correct.</p><p>For the guidebooks, Felix needs to correct the museum’s collection expectation through the agreed customer channel. The warehouse needs an amended collection instruction. The owner of the shared order record needs to replace Friday’s status there too. Ask those contacts whether they passed the date to anybody else who is preparing affected work.</p><p>Keep the explanation proportionate. The museum needs the current delivery position and its options. It does not need an account of Felix’s behaviour in a team meeting. Private coaching details should stay in their appropriate record.</p><p>Take responsibility for your part: you approved the date before completing the production check. Do not make Serena carry the correction because she noticed the gap, or describe the change as her request. She supplied information the decision needed.</p><h3>Leave a usable instruction in each affected place</h3><p>Use a short correction note in the existing work record. This is a suggested aid; adapt it to the commitment and the access you actually have. It should answer:</p><ul><li>Which earlier promise is being corrected, and what checked information changed its status?</li><li>What should each affected person do now, including any work that must pause?</li><li>Who owns the unresolved decision, and when can the recipient expect the next answer?</li><li>Which messages or work records need changing, and who will confirm that happened?</li></ul><p>In the hypothetical case, the internal note could say: “Friday collection is unconfirmed. The current printer booking is Monday; finishing and collection timing still need checking. Felix will correct the museum’s expectation now and agree a 15:00 status update with them, including any questions still open. Serena will check production options with the printer before then. I own the replacement delivery commitment and will confirm the warehouse record has changed before that update. The warehouse should remove Friday’s collection instruction and hold collection preparation until the revised instruction arrives.”</p><figure><img alt="Illustration: “Date in doubt. Three places to fix.” In the hypothetical guidebook order, the museum’s collection expectation, warehouse preparation and shared order record connect to You as three places requiring correction." src="https://cdn-images-1.medium.com/max/1024/1*epfbOo_nl5zqNAM-0bJl5w.png" /><figcaption>The hypothetical guidebook correction must reach the museum, the warehouse and the shared order record.</figcaption></figure><p>That note belongs with the affected order. Felix’s customer message should contain the details the museum needs, including any existing promise that is being withdrawn. Confirm the interim arrangement with the people receiving it; they may have a dependency your internal note missed.</p><p>A general “please see the latest plan” can leave two competing instructions in use. Name the superseded promise and update the place where someone will next act on it. Use authorised records and channels; tracing one commitment does not require access to everyone’s private messages.</p><h3>Check the changed work, then return to the meeting</h3><p>At the agreed follow-up, inspect the actual correction. Has the museum received a usable answer? Does the warehouse’s next action reflect the current position? Is the order record consistent with what Felix communicated? If a recipient has already incurred a cost or made another commitment, assign that consequence to someone who can respond.</p><p>Cooperly’s guidance on <a href="https://cooperly.ai/insights/early-escalation-team-communication?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=leadership_communication_systems&amp;utm_content=meeting_warning_commitment_correction">making an early warning useful to the work</a> puts an owner and follow-through behind the signal. In this case, Serena should be able to see what became of the production warning. She need not receive private customer correspondence to know which decision followed.</p><p>Return separately to the interruption. Check what Felix understood, agree how he can raise an urgent point, and decide how you will bring an unfinished contribution back. Apply that expectation to yourself too. The corrected order does not establish that the next meeting will work better; check whether relevant information is heard before the next commitment.</p><p>Cooperly is the Human Operating System for founder-led, AI-native companies. The operating aim here is fewer repeated explanations: the people doing the work can find the current agreement and its owner. Human judgement still determines the commitment and the conversation.</p><p>For the wider management response, use <a href="https://cooperly.ai/insights/employee-keeps-interrupting-meetings?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=leadership_communication_systems&amp;utm_content=meeting_warning_commitment_correction">Cooperly’s guide to handling repeated meeting interruptions</a>. Before your next check-in, open the last promise you corrected and follow it to one person outside the room. Find out which version they are using.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=903344849801" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Before You Turn an AI Hiring Test into a First-Week Target]]></title>
            <link>https://medium.com/@edkh/before-you-turn-an-ai-hiring-test-into-a-first-week-target-9e9525711f93?source=rss-d7b2805cd5b0------2</link>
            <guid isPermaLink="false">https://medium.com/p/9e9525711f93</guid>
            <category><![CDATA[hiring]]></category>
            <category><![CDATA[management]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[leadership]]></category>
            <dc:creator><![CDATA[Ed Khristus]]></dc:creator>
            <pubDate>Tue, 20 Oct 2026 08:25:12 GMT</pubDate>
            <atom:updated>2026-10-20T08:25:12.319Z</atom:updated>
            <content:encoded><![CDATA[<h4>Before carrying a work-sample deadline into the job, check which tools, information and support the new hire will have.</h4><p>A candidate finishes your AI-enabled work sample in forty minutes. Their reasoning is sound, they catch a generated error and they explain the edits. You hire them. In their first week, a similar-looking assignment takes most of a day.</p><p>Before asking why they have slowed down, put the two tasks side by side. Did the exercise contain everything they needed? Can the real supplier notes enter the tool they used? Has the company account been approved? Who must check the result before anyone uses it?</p><p>Compare the amount and difficulty of the work too, and distinguish time spent producing it from time waiting for answers or review.</p><p>The sample can remain useful evidence of judgement. Its duration needs a separate explanation. If the conditions changed, treating forty minutes as a standing output target can make the newcomer accountable for work the company has not prepared.</p><figure><img alt="Cover: ‘Fast in the test. Waiting in the job.’ A synthetic sample finished in forty minutes and a new starter awaiting tool access sit at different points on a timeline leading to You." src="https://cdn-images-1.medium.com/max/1024/1*vtwevi6Tdi5aYsRShmYbmQ.png" /><figcaption>A completed hiring sample and an access-dependent first assignment provide different evidence about pace.</figcaption></figure><h3>Find the extra work inside the first assignment</h3><p>Consider a hypothetical wholesale homeware company hiring a catalogue editor. For the hiring exercise, it supplies synthetic product sheets and permits an AI tool to help draft descriptions. The candidate must check each description against the sheets and explain any correction.</p><p>Everything is in one pack. The candidate spots a material described incorrectly, fixes it and finishes in forty minutes. That gives the panel something concrete to discuss: how the person checks a claim and decides whether a description is usable.</p><p>After joining, the editor receives supplier notes for the next catalogue. Under the company’s existing rule, these notes cannot go into the tool used during the exercise. The company-approved tool is not yet available through their work account. Some product details also need confirmation from the buying team, and the manager must review the descriptions before proofing.</p><p>The manager asks for the set by lunchtime, remembering the test. The editor is now clarifying facts, finding the permitted workflow and waiting for review as well as writing. None of those dependencies appeared in the pack.</p><p>The quality requirement still matters. A description must represent the product accurately. What needs checking is the assumed route to that result, including the manager’s part in making it possible.</p><h3>Keep the sample’s finding narrower than its finish time</h3><p>Cooperly’s <a href="https://cooperly.ai/insights/ai-assisted-job-applications?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=ai_hiring_first_week_target">guide to role-relevant AI hiring exercises</a> recommends stating tool conditions and examining the candidate’s choices and verification. Carry that evidence into the job as a useful starting point.</p><p>For the catalogue editor, the finding might be: they checked a generated description against supplied product facts, found an error and explained the correction. The panel has not observed how they resolve an unanswered supplier question or coordinate a review with this manager.</p><p>A time limit can help keep a hiring exercise proportionate. It does not establish a sustainable production rate. A short prepared task also leaves out the interruptions and competing commitments of a normal working day.</p><p>Record the capability you observed, alongside the conditions that made the observation possible. Ask the new hire which parts of the first assignment are familiar and which need company-specific context. Their answer can guide support; it should not be treated as a confession that the interview was misleading.</p><h3>Permission in an exercise does not travel with the candidate</h3><p>Allowing AI on synthetic hiring material says nothing about permission to process real work information. The same person can follow both instructions correctly: use the allowed tool in the exercise, then refrain from using it on restricted supplier notes.</p><p>The company’s <a href="https://cooperly.ai/insights/manage-shadow-ai-at-work?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=ai_hiring_first_week_target">usable rules for approved AI tools and data</a> need to reach the manager assigning the work as well as the new starter. “We’re an AI-native company” does not answer which inputs this particular workflow permits.</p><p>If approval or access is pending, the authorised owner must resolve it. The manager can arrange an already permitted method, narrow the first assignment or agree a feasible delivery time with the people waiting. An urgent catalogue deadline does not grant permission to use a personal account.</p><p>Some restrictions will remain because the real work carries information absent from the exercise. Explain that difference. If the role routinely operates under those restrictions, revise the next hiring exercise so it tests a relevant way of working.</p><h3>Agree the first output before borrowing the benchmark</h3><p>Use the existing assignment note to compare the exercise with the real task. This is a proposed preparation aid, not a validated productivity assessment. Keep four questions beside the first output:</p><ul><li>What did the sample demonstrate, and how does this assignment compare in volume, difficulty and quality requirements?</li><li>Which information and tools are available and permitted now? Who will resolve anything missing?</li><li>What must the new hire check, what needs another person’s answer, and who reviews the result?</li><li>What delivery can everyone now agree, and when will an unresolved dependency require another decision?</li></ul><p>For the catalogue, the editor and manager could choose a small set of products with confirmed facts and a permitted drafting method. The manager arranges a review time. The buying lead owns the outstanding product questions; those descriptions remain outside the agreed first set until the questions are answered.</p><p>If that smaller set cannot meet the proofing need, the manager has another decision to make about scope or timing. “Try your best” would leave the editor carrying a deadline they cannot make feasible alone.</p><figure><img alt="Illustration: ‘The task is agreed. Inputs still pending.’ The catalogue editor’s confirmed-fact set and the buying lead’s outstanding answers converge on the first-set agreement, which leads to You; the answers use a dashed connector." src="https://cdn-images-1.medium.com/max/1024/1*DjItQ8UCbeHo8J2KwN_jVg.png" /><figcaption>The first set uses confirmed product facts; the buying lead still owns the unanswered questions.</figcaption></figure><p>Invite the editor to correct the note. Keep it about the work and the agreed support. Interview impressions, private allegations and sensitive personal information do not belong in a shared task brief.</p><h3>Review what happened under the conditions you supplied</h3><p>At the agreed review, inspect the descriptions and the checks behind them. Did the editor use the confirmed facts accurately? Did the promised answers and manager review arrive? Which parts required explanation that the hiring pack had supplied automatically?</p><p>A missed company input does not erase an unrelated error in the editor’s work. An accurate finished set does not prove that the access problem has been solved either. Keep each finding attached to what happened.</p><p>Cooperly’s <a href="https://cooperly.ai/insights/why-good-employee-quits-first-90-days?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=ai_hiring_first_week_target">guide to comparing the hired role with its reality</a> puts access, support and representative first work in the entry process. Make this comparison while the person is beginning, before a delay hardens into a story about their ability.</p><p>One assignment will leave questions open. It can tell you what to clarify before the next comparable piece of work. When the team establishes a workable method, update the role briefing and hiring exercise too, so the next candidate encounters conditions the company can actually provide.</p><p>Cooperly is the Human Operating System for founder-led, AI-native companies. Its People Context Layer connects working profiles, team signals, goals and team memory. Approved AI tools can request selected, permissioned, read-only context to help prepare work; access follows the relevant user, team and integration permissions. People still grant access, agree expectations and judge the result.</p><p>If hiring and team leads need that shared context in their approved tools, <a href="https://cooperly.ai/ai-context-layer?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=ai_hiring_first_week_target">explore Cooperly’s AI Context Layer and book a demo</a>.</p><p>Before writing the first deadline, put the sample’s finish time beside the conditions that produced it. Agree the first real output from the conditions the company can provide now.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9e9525711f93" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Your Backup Is Ready. Why Does Everyone Still Ask the Expert?]]></title>
            <link>https://medium.com/@edkh/your-backup-is-ready-why-does-everyone-still-ask-the-expert-1eeb63924ee5?source=rss-d7b2805cd5b0------2</link>
            <guid isPermaLink="false">https://medium.com/p/1eeb63924ee5</guid>
            <category><![CDATA[communication]]></category>
            <category><![CDATA[workplace]]></category>
            <category><![CDATA[management]]></category>
            <category><![CDATA[leadership]]></category>
            <dc:creator><![CDATA[Ed Khristus]]></dc:creator>
            <pubDate>Mon, 19 Oct 2026 19:09:14 GMT</pubDate>
            <atom:updated>2026-10-19T19:09:14.554Z</atom:updated>
            <content:encoded><![CDATA[<h3>Before someone goes on leave, check how colleagues send requests to their cover and use the work that comes back.</h3><p>The backup can do the work. They have tried it, used their own approved access and produced a result that meets the agreed checks. Then the expert goes on leave, and colleagues keep asking the expert for help.</p><p>Before arranging another teaching session, follow one of those requests. Did it reach the backup? Did the sender ask for something the cover actually included? Could the person waiting use the answer, or did they still need a decision nobody had agreed to make?</p><p>Temporary cover depends on the people around the substitute too. A task demonstration answers a useful question about capability. It leaves another question open: can this team get a real request through the arrangement it has made?</p><figure><img alt="Cover: ‘The backup is ready. Requests go elsewhere.’ A trained backup and a request still sent to the expert connect to You; the request route is dashed." src="https://cdn-images-1.medium.com/max/1024/1*GECdolHHLoN_8hFmdV0evw.png" /><figcaption>A request can reach the backup while still asking the absent expert to make the next decision.</figcaption></figure><h3>Find the request that bypassed the cover</h3><p>Consider a hypothetical training company. Its course coordinator will be away for a week. A colleague has practised updating the internal course timetable from confirmed teaching assignments. The timetable passed its checks, and another commitment has moved to make room for the cover.</p><p>The colleague can enter confirmed room and tutor changes. They cannot decide which of two courses should get a room both have requested. The operations lead owns that choice. These are illustrative boundaries for ordinary internal coordination.</p><p>On the first morning of cover, a course lead emails a room-change request to the usual coordinator. They copy the backup, then ask the coordinator to confirm that the change is acceptable. The backup sees the message but waits for the requested confirmation. The timetable stays unchanged.</p><p>It would be premature to conclude that the backup lacks confidence. The request still names the absent person as the decision route. Another lesson in editing the timetable would leave that route intact.</p><p>Cooperly’s <a href="https://cooperly.ai/insights/key-person-dependency-backup-check?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=founder_operating_systems&amp;utm_content=backup_handoff_counterpart">guide to checking a backup’s demonstrated scope</a> covers competence, capacity and fallback arrangements. Once those have been checked, inspect how the remaining colleagues actually use the cover. A shared note is a starting point; the next request shows whether people understood it.</p><h3>Ask the sender to use the temporary arrangement</h3><p>The manager can ask the course lead what confirmation they needed. Perhaps the room change was already authorised and the sender used an old email thread. Perhaps the request requires a new decision. Those explanations call for different repairs.</p><p>For an already confirmed change, the sender should address the backup directly through the agreed work channel and identify the confirmation. For a disputed room allocation, they should bring the choice to the operations lead. Copying everybody does not tell the backup which of those requests they received.</p><p>Keep acknowledgement separate from acceptance. The backup might reply: “I’ve seen the request. Both courses currently need the same room, so I’ve asked the operations lead to decide. The timetable still shows the current allocation.”</p><p>That reply names the unresolved choice. It neither promises the change nor treats silence as approval. Cooperly’s <a href="https://cooperly.ai/insights/employee-not-responding-to-emails?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=founder_operating_systems&amp;utm_content=backup_handoff_counterpart">guide to agreeing the response a request needs</a> separates seeing a message from answering or approving it.</p><p>If the sender continues privately messaging the absent expert, the manager needs to address that behaviour with the sender. Do not make the backup responsible for persuading every colleague to honour an arrangement the manager announced.</p><h3>Follow the result to the person who needs it</h3><p>Now suppose the operations lead approves the room change. The backup updates the timetable by the start of Wednesday’s course. The file is accurate. Yet the teaching assistant needed the confirmed room on Tuesday afternoon to prepare the materials.</p><p>The backup met the event time and missed the useful handoff. Check whether anyone told them about Tuesday’s preparation. If the cover agreement only said “before the course”, the manager needs to repair that expectation before treating it as a broken promise.</p><p>The <a href="https://cooperly.ai/insights/employee-keeps-missing-deadlines?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=founder_operating_systems&amp;utm_content=backup_handoff_counterpart">Cooperly guide to usable delivery commitments</a> distinguishes the employee’s output, any review and the recipient’s use time. Bring the receiving colleague into the cover arrangement wherever that distinction changes what must happen.</p><p>In the example, the operations lead and teaching assistant need to agree when the room decision must arrive for preparation to proceed. If that time cannot be met, they must settle a feasible alternative and tell the backup what is now agreed. A forecast that the timetable will be late leaves this choice unresolved.</p><h3>Try one request through the whole route</h3><p>Before the absence, use one representative request to check the connections. Tell everyone involved what you are testing. Keep it to a bounded, ordinary task; use a practice copy and synthetic details if real consequences cannot be contained. Leave financial transactions, production changes and specialist work to their proper controls.</p><p>This is a suggested working check, not a validated resilience test. Add four short lines to the existing cover note:</p><ul><li><strong>Sender:</strong> where the request goes, what it asks for and what confirms it falls within cover.</li><li><strong>Backup:</strong> the response they can give, including the unresolved decision they must identify and when they will raise a missing decision.</li><li><strong>Decision owner:</strong> who accepts a change outside that scope, when their response is needed and how the current instruction reaches the people affected.</li><li><strong>Recipient:</strong> what usable result they need and when their next action depends on it.</li></ul><p>For the timetable, let the course lead send the practice room change. Have the backup distinguish an approved update from a disputed allocation. Ask the operations lead to respond through the agreed route, then have the teaching assistant check whether the resulting timetable would support preparation.</p><figure><img alt="Illustration: ‘Timetable updated. Can they prepare?’ You connects through the backup updating the timetable to the teaching assistant who needs the confirmed room." src="https://cdn-images-1.medium.com/max/1024/1*zLrF4Q9TGhqwKBL14RS6Ew.png" /><figcaption>The manager’s cover check continues past the backup’s update to the colleague preparing the course.</figcaption></figure><p>Observe the help that makes it work. If the expert has to forward the request, locate the approver or explain the recipient’s deadline, record that specific intervention. The finished timetable alone would conceal it.</p><h3>Repair the connection you actually observed</h3><p>A request sent to the wrong person needs a routing correction with its sender. A request that exceeds cover needs an available decision owner. An accurate result received too late needs a different handoff agreement. If the backup cannot produce the agreed result despite the promised inputs and support, return to the relevant teaching and practice.</p><p>Keep the conclusion as narrow as the evidence. One complete practice request supports this arrangement under the conditions tried. It does not show that the backup can absorb every new course, room dispute or change in workload.</p><p>If the decision owner cannot be available during leave, resolve that before announcing cover. Name another authorised person, narrow what will be handled or agree what waits with the affected colleagues. The expert’s holiday inbox should not be the unspoken fallback.</p><p>When the expert returns, review where requests went and which decisions remained pending. Explicitly end or renew the temporary arrangement with the people involved, restore the backup’s displaced work and ask the authorised owner to review temporary access.</p><p>Cooperly is the Human Operating System for founder-led, AI-native companies. Its working profiles and team context can help leaders prepare conversations about these dependencies. The manager still has to agree the cover and check the actual work.</p><p>Before the next absence, choose a request that colleagues would normally send to the expert. Use <a href="https://cooperly.ai/insights/key-person-dependency-backup-check?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=founder_operating_systems&amp;utm_content=backup_handoff_counterpart">Cooperly’s backup coverage guide</a> to check the substitute’s scope, then follow that request all the way to the person waiting for the result.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1eeb63924ee5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI Roleplay for Managers: Are You Rehearsing the Right Conversation?]]></title>
            <link>https://medium.com/@edkh/ai-roleplay-for-managers-are-you-rehearsing-the-right-conversation-e28d0a7963cf?source=rss-d7b2805cd5b0------2</link>
            <guid isPermaLink="false">https://medium.com/p/e28d0a7963cf</guid>
            <category><![CDATA[management]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[communication]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Ed Khristus]]></dc:creator>
            <pubDate>Sat, 17 Oct 2026 08:11:10 GMT</pubDate>
            <atom:updated>2026-10-17T08:11:10.218Z</atom:updated>
            <content:encoded><![CDATA[<h4>Before practising difficult feedback, check who can actually change the instruction that caused the problem.</h4><p>Before you ask AI to play the employee in a difficult conversation, write down what you need that employee to change. Then check whether they have the authority to change it.</p><p>That small check can send your preparation in a different direction. A team lead may be rehearsing a conversation about delivery when two leaders have given the employee requirements that cannot both be met. The employee can explain the problem. They cannot withdraw either leader’s instruction.</p><p>If you use AI roleplay for managers to prepare, first decide which conversation needs practice. If the unresolved choice belongs to another leader, start there. A more fluent feedback script will not settle that choice.</p><figure><img alt="Two small luminous constellations send separate fine curves towards one quieter point, but the curves stop apart on either side of it." src="https://cdn-images-1.medium.com/max/1024/1*1XNP5ZcfCevOU5E0WvooUw.png" /><figcaption>Two instructions reach towards one employee; the leaders have yet to make them usable together.</figcaption></figure><h3>Find the instruction underneath the feedback</h3><p>Consider a hypothetical studio preparing an open day. Its coordinator must send visitors their guide on Tuesday. The commercial lead has asked for the complete guide to go out then, so visitors can plan their arrival.</p><p>The operations lead has also told the coordinator to hold the guide until the final walkthrough route is confirmed. That choice will happen on Friday morning. The instructions concern ordinary visitor information in this example; no required safety or specialist check is being waived.</p><p>The coordinator raises the mismatch. The commercial lead replies that they need to “take more ownership”. On Monday afternoon, the contradiction is still unresolved and the lead is preparing feedback about initiative. An AI rehearsal could help them polish the accusation.</p><p>Look at the actual requests first. Sending the complete guide on Tuesday and holding that guide until Friday cannot both happen. Asking the coordinator to be more proactive has not changed either requirement. Nor would a confident recommendation from the practice chat.</p><p>Cooperly’s <a href="https://cooperly.ai/insights/conflicting-instructions-two-managers?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=ai_roleplay_right_conversation">guide to resolving incompatible manager instructions</a> puts responsibility for the revised brief with the people who can amend it. Here, that means the two leads need a conversation before either treats the delivery problem as a failure of initiative.</p><h3>Choose the first person who can alter the requirement</h3><p>The commercial lead still has something useful to ask the coordinator: “Can we look at the two requests you received and what each of us understood was due?” That is fact-finding. It should leave room for a message the lead missed or a misunderstanding about what “complete” meant.</p><p>Suppose the messages confirm the contradiction. The first negotiation is now with the operations lead. It might begin: “We have asked for a complete guide on Tuesday and told the coordinator to hold it until Friday. Visitors need arrival details earlier. What information can we release now, and what needs a separate update?”</p><p>The leads could agree to send the confirmed arrival information on Tuesday and provide the walkthrough details on Friday. They must check that this arrangement serves the visitors and fits the coordinator’s workload. Splitting a document can create another piece of work; somebody has to account for it.</p><p>They could instead move the circulation date or bring the route decision forward, if either option is feasible. The right choice depends on the studio’s actual constraints. The employee should not have to guess which lead will object less.</p><p>If neither lead can authorise the change, take the specific conflict to the person who can. Give the coordinator a usable interim instruction: which confirmed material to prepare, which circulation must wait and when they will receive an answer. An indefinite request to “work it out” leaves them carrying the same problem.</p><h3>Use a recipient check before opening the practice chat</h3><p>For an ordinary coordination problem, try completing these four lines before rehearsing. This is a suggested preparation aid, not a validated coaching method.</p><ul><li><strong>The expectation I want to discuss:</strong> the actual deliverable or behaviour, stated precisely enough to check.</li><li><strong>The person who can change the disputed condition:</strong> the employee, another leader or a decision owner elsewhere.</li><li><strong>My first real question:</strong> what I need to learn or settle with that person.</li><li><strong>What the employee can do meanwhile:</strong> the work that can proceed and the answer they are waiting for.</li></ul><p>For the studio lead, the second line points to their colleague. The first rehearsal therefore concerns a peer conversation about the guide. It does not begin with persuading the coordinator to accept responsibility for an impossible instruction.</p><figure><img alt="A bright fine line connects two upper luminous constellations, and one coherent path continues from their connection to a quieter point below." src="https://cdn-images-1.medium.com/max/1024/1*nBvSCy85ohTSJ27-AD87uw.png" /><figcaption>Connect the leaders who can change the requirements first; then carry one workable instruction to the employee.</figcaption></figure><p>A founder supporting this lead can use the same check. Ask which instruction the lead expects the employee to follow and whether the lead can actually make it workable. Cooperly’s <a href="https://cooperly.ai/insights/managing-managers-one-to-one?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=ai_roleplay_right_conversation">guide to supporting a team lead’s judgement</a> distinguishes coaching from a constraint the founder owns. If only you can settle the disputed requirement, another round of rehearsal is an incomplete response to their request for help.</p><p>You can still coach how the lead noticed and raised the conflict. One case does not establish their overall management ability. If a decision needs more information, name the missing input and give the lead a time to expect your answer.</p><h3>Rehearse the peer conversation with its uncertainty intact</h3><p>Once the recipient is clear, use an approved AI tool and a synthetic version of the case. Keep private employee material out of the request. You can practise the structure of this negotiation with invented names and ordinary requirements.</p><p>A request to adapt: “Play a fictional fellow manager. I need to resolve two incompatible instructions for a visitor guide. Ask me what each instruction protects. Offer a plausible constraint I have not considered, clearly labelled as fictional. Wait for my response, then quote any words I used that closed off correction or demanded that you give way.”</p><p>Represent the two requirements accurately in the synthetic case, using invented names and nonsensitive details permitted by your team’s rules. Respond in your own words. A fictional constraint might expose a question worth asking. It does not establish that the operations lead has that constraint. Keep it out of the factual account.</p><p>Cooperly’s <a href="https://cooperly.ai/insights/ai-roleplay-for-managers?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=ai_roleplay_right_conversation">conversation-rehearsal guide</a> treats generated replies as practice material and directs attention to the manager’s questions and handling of correction. For this peer discussion, inspect whether your opening names both requirements fairly. Could your colleague correct your account? Did you ask what must remain, or assume their requirement was dispensable?</p><p>A friendly simulated agreement cannot confirm the revised brief. Only the authorised people in the real conversation can do that. Take the questions to them and let their answer change the proposal.</p><h3>Return to the employee with the change made explicit</h3><p>Suppose the studio leads agree the two-part guide. The operations lead explicitly withdraws the instruction to hold all circulation; the commercial lead withdraws the demand for final walkthrough details on Tuesday. They confirm who checks each part and account for the additional update.</p><p>The coordinator can now explain what they will send and when. The manager should check the actual delivery and whether either leader revives the replaced requirement. Any separate concern about the coordinator’s work needs its own evidence. Resolving the contradiction neither proves poor performance nor excuses an unrelated missed commitment.</p><p>After the real conversation, the authorised people need to update the appropriate record with the current instruction and what it replaces. Cooperly, the Human Operating System for founder-led, AI-native companies, makes selected, permissioned, read-only team context available to approved AI tools for later preparation. Access depends on user, team and integration permissions. The connection cannot write that agreement for you or resolve the leaders’ dispute.</p><p>If your managers prepare conversations across these boundaries, explore <a href="https://cooperly.ai/ai-context-layer?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=ai_roleplay_right_conversation">Cooperly’s AI Context Layer and book a demo</a> to see how relevant team context can reach approved tools.</p><p>Before your next rehearsal, name the person who can change the condition you are about to discuss. Prepare your opening for that person.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e28d0a7963cf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Who Gets Credit for Making AI-Generated Work Usable?]]></title>
            <link>https://medium.com/@edkh/who-gets-credit-for-making-ai-generated-work-usable-a125088d8dc4?source=rss-d7b2805cd5b0------2</link>
            <guid isPermaLink="false">https://medium.com/p/a125088d8dc4</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[management]]></category>
            <category><![CDATA[workplace]]></category>
            <dc:creator><![CDATA[Ed Khristus]]></dc:creator>
            <pubDate>Fri, 16 Oct 2026 08:09:35 GMT</pubDate>
            <atom:updated>2026-10-16T08:10:08.620Z</atom:updated>
            <content:encoded><![CDATA[<h4>Recognise the human judgement behind a faster draft, then decide who should carry the checking next time.</h4><p>Imagine a founder praising the new internal FAQ in the team update: “Leah has shown us how much faster we can move with AI.” Tom, who corrected the answers before anyone could use them, reads the message too.</p><p>Leah did useful work. So did Tom. The update has made one contribution visible and left the other inside a finished document.</p><p>When you recognise AI-enabled productivity, ask who made the output usable. Check the actual contributions and correct an incomplete account where it mattered. Then make a separate decision about the next assignment. Naming the person who repaired the work does not settle whether they should keep repairing it.</p><figure><img alt="A white three-sided rectangular frame is completed by a grey right edge, while a small white marker points only to the white portion." src="https://cdn-images-1.medium.com/max/1024/1*o1-pGMRrDQU5whgyNj63jg.png" /><figcaption>The finished result depends on both contributions; a success account can still point to only one.</figcaption></figure><h3>Read the changes behind the success update</h3><p>In this hypothetical software company, Leah uses an approved AI tool to organise existing notes into answers for customer-facing colleagues. She selects relevant questions, checks several answers and prepares a readable draft.</p><p>Tom, the product specialist asked to check it, finds an answer that describes an export option the software does not offer. Another omits a limitation that changes when colleagues should recommend the feature. He checks the current product documentation, rewrites those answers and explains the restriction to Leah. The FAQ becomes usable after those changes.</p><p>The founder sees the final page and Leah’s delivery message. “Leah prepared the FAQ” is accurate. “Leah solved the enablement problem with AI” gives an incomplete account if it implies that the material arrived ready to use.</p><p>Look at what changed between the draft and the version people relied on. Ask both colleagues which decisions they made and what evidence they checked. Do not infer the whole contribution from the file owner, its word count or the last edit. Some work may have happened in a conversation; ask about it before concluding that the record is complete.</p><p>Cooperly’s <a href="https://cooperly.ai/insights/employee-taking-credit-for-others-work?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=ai_output_contribution_credit">guide to checking and correcting credit claims</a> separates supported contributions from disputed history. The same discipline applies when the success being described involves AI. It does not require deciding that anyone intended to mislead.</p><h3>Describe the contribution precisely enough to recognise it</h3><p>“Thanks to Tom for helping” may leave the original account almost intact. Helping could mean fixing a typo or supplying the judgement that prevented a misleading answer from reaching customers. Say which work you can support.</p><p>For the hypothetical FAQ, a corrected update could read: “Leah selected the questions and prepared the draft. Tom checked the product claims and rewrote the answers on export availability and limitations. Leah incorporated the changes into the final page.” Check that wording with the people involved.</p><p>This gives Leah credit for her contribution too. Using AI does not erase the decisions she made. Nor should a reviewer inherit sole ownership merely because they found a consequential error. Describe the work each person actually did.</p><p>Correct the account in the message or record where people relied on it. If the omission was in your team update, repair your update. A private thank-you to Tom cannot change what the team heard in your update.</p><p>If an assignment, review or reward decision relied on the incomplete account, take the correction to the person authorised to review that decision. Do not promise a changed outcome. Where a material contribution remains disputed, preserve that uncertainty while correcting the parts you can establish.</p><h3>Decide whether checking is part of the job or an unplanned rescue</h3><p>Accurate credit can coexist with an unreasonable allocation. Tom might receive public praise and still lose the time reserved for work he was meant to lead.</p><p>Ask what his role requires. Product review may be a central responsibility with recognised value. A colleague’s request for expertise can also be entirely appropriate. The question is whether this check was planned and feasible, or whether a finished-looking draft concealed substantial work that someone else had to absorb.</p><p>Cooperly’s <a href="https://cooperly.ai/insights/ai-workslop-review-burden?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=ai_output_contribution_credit">AI-workslop guide</a> distinguishes useful early collaboration from a handoff that leaves the recipient to reconstruct essential thinking. Keep that distinction when discussing the FAQ. Tom’s knowledge may be needed to interpret a limitation; Leah may be able to check basic feature availability herself.</p><p>Before the next FAQ update, agree which checks Leah owns before sending it and which specific question needs Tom’s expertise. Give him time for that review. Suppose it conflicts with a release note he already owns. The manager agrees that the internal FAQ can wait until after the release note, and tells the colleagues expecting it. That choice depends on the work and its consequences; it is not a standard priority rule. If a draft needs much more repair than agreed, he should be able to return the identified gaps to its owner or ask the manager to change the scope. Required checks still happen before the material is used.</p><p>A general instruction to “be more collaborative” leaves the allocation unresolved. So does requiring Tom to approve everything while keeping all his other commitments.</p><h3>Ask two different questions about the same piece of work</h3><p>Use these two questions with the people involved. They are a suggested discussion aid, not a validated attribution method.</p><ul><li><strong>What account of the finished result can we support?</strong> Name the output, the substantive contribution each person made, any unresolved point and the message or decision that needs correcting.</li><li><strong>What are we asking each person to carry next time?</strong> Name the author’s checks, the specialist contribution still required and the competing commitment the manager will change if that work will not fit.</li></ul><p>Keep the answer beside the relevant work. You do not need percentages of credit, a ranking of helpful colleagues or a transcript of private conversations. A short, checked account is enough when it explains the contribution and the next decision.</p><figure><img alt="A large white three-sided rectangular outline joins a grey closing edge; separate stems connect the white and grey portions to two distinct outlined markers below." src="https://cdn-images-1.medium.com/max/1024/1*Zuq_JdnuK4YY4_yVKBV_Zw.png" /><figcaption>Trace each substantive contribution to the finished output before deciding who should provide it next time.</figcaption></figure><p>Cooperly’s <a href="https://cooperly.ai/insights/non-promotable-tasks-fair-allocation?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=ai_output_contribution_credit">guide to allocating necessary support work</a> asks managers to consider that work alongside development opportunities. Review is not automatically low-promotability work. Check the criteria for Tom’s actual role and whether the contribution appears in the decisions that use those criteria.</p><h3>Look at the next output and the work it displaced</h3><p>Suppose Leah completes the agreed feature checks on the next update, and Tom reviews only the remaining interpretation question. Inspect the result together. Did the narrower review prove sufficient? Did another colleague quietly finish the repair? Was Tom’s planned work able to proceed?</p><p>If the specialist contribution remains necessary, recognise it and allocate it deliberately. If the same avoidable errors return, address those specific gaps with the author. If the volume exceeds the team’s capacity, reduce or reschedule work. Giving the repair a name has not created more hours.</p><p>For Cooperly, the Human Operating System for founder-led, AI-native companies, this is a reason to keep working expectations available to the people preparing the next assignment. Reconstructing who checks what in every meeting adds management work as output grows.</p><p>Approved AI tools can use selected, permissioned, read-only Cooperly context to help prepare a brief or conversation. That context cannot establish who created an output or decide who deserves a reward. People check the evidence and make those decisions.</p><p>If your team uses AI to prepare work across these handoffs, explore <a href="https://cooperly.ai/ai-context-layer?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=ai_people_context&amp;utm_content=ai_output_contribution_credit">Cooperly’s AI Context Layer and book a demo</a> to see how approved tools can receive relevant team context.</p><p>Open your last AI success update and check whether the person who made the work usable appears in it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a125088d8dc4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[When Two Managers Disagree About an Employee’s Performance]]></title>
            <link>https://medium.com/@edkh/when-two-managers-disagree-about-an-employees-performance-621ba9a5343a?source=rss-d7b2805cd5b0------2</link>
            <guid isPermaLink="false">https://medium.com/p/621ba9a5343a</guid>
            <category><![CDATA[workplace]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[management]]></category>
            <category><![CDATA[performance-management]]></category>
            <dc:creator><![CDATA[Ed Khristus]]></dc:creator>
            <pubDate>Thu, 15 Oct 2026 08:14:58 GMT</pubDate>
            <atom:updated>2026-10-15T08:16:13.987Z</atom:updated>
            <content:encoded><![CDATA[<h4>Before reconciling their ratings, check whether they saw different work or gave different meanings to the same event.</h4><p>One manager calls an employee dependable. Another says they keep letting the team down. You own the performance review, and both managers want their assessment reflected in it.</p><p>Asking them to settle on a rating may get you an answer before you know what the disagreement means. They might have seen different work. They might be judging different responsibilities. Or they might be retelling the same incident as if each account supplied another example.</p><p>Start with the work episode behind each claim. Establish what happened, what was expected at the time and what each person actually observed. Then decide which conclusions the evidence supports. Agreement between managers is useful only when you can explain what they have agreed about.</p><figure><img alt="Two overlapping white and grey elliptical outlines contain one white disc in their shared space on a dark violet field." src="https://cdn-images-1.medium.com/max/1024/1*U4lo5ufOMUidsoj269cPLQ.png" /><figcaption>Two views of one work episode can support different findings; they do not create two episodes.</figcaption></figure><h3>Find the event hiding inside each assessment</h3><p>Consider a hypothetical community-events organiser in a small software company. Their line manager says they have become unreliable: the preparation pack for a customer event arrived late. A partnerships lead describes them as exceptionally responsive: they rescued a partner demonstration during the same week.</p><p>The founder hears incompatible verdicts. The managers are describing different assignments, connected by a priority change. The dated work tells a more specific story. The partnerships lead asked the organiser to take over the demonstration, and the line manager agreed that the preparation pack could move. The event owner was never told its new date.</p><p>Both managers saw something real. The demonstration was delivered. The pack reached its recipient later than originally expected. Those facts alone do not settle whether the organiser missed an active commitment, whether a manager failed to communicate the change, or whether both happened.</p><p>Begin with a question that forces the account back onto the work: “Which assignment are you referring to, and what happened on that occasion?” Ask it of the manager offering praise too. “Always reliable” needs a basis just as much as “keeps letting us down”.</p><p>Write the accounts beside the episode they describe. Keep distinct assignments separate. If a second account points to the same preparation pack, attach it there instead of adding another missed deadline to the employee’s history.</p><h3>Keep corroboration separate from repetition</h3><p>Suppose a project lead also says the organiser has a deadline problem. When asked for an example, they mention the late event pack. They heard about it from the line manager; they did not receive the pack or see the agreement that changed its date.</p><p>You now have another reported concern, but no additional example of late work. The route by which a judgement travelled through the company is relevant to understanding the feedback. It is not a second performance incident.</p><p>A second firsthand account can still add evidence. The event owner can establish when the pack arrived and explain what its lateness prevented. Their account may corroborate timing and reveal an effect the line manager did not see. Preserve that contribution. Avoid converting the number of people discussing an event into the frequency with which the employee repeated it.</p><p>Cooperly’s <a href="https://cooperly.ai/insights/new-manager-performance-review?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=performance_evidence&amp;utm_content=two_managers_same_episode">guide to assessing work you did not observe</a> distinguishes reported interpretations from dated work and direct observation. The same distinction helps when several current managers supply review input. A confident account remains an account to check.</p><p>This check cuts both ways. Three people praising one successful demonstration do not establish consistent performance across the quarter either. Record what that success shows, then look for the other work relevant to the review.</p><h3>Check what the employee was asked to protect</h3><p>In a real review, check that account before relying on it. Which commitment was the organiser expected to prioritise when the demonstration request arrived? Did the line manager actually authorise the delay? Was that within their authority? Did the organiser still have an agreed responsibility to notify the event owner?</p><p>Check the appropriate work records and ask the people involved. Keep an unresolved account attributed: “The organiser says the pack date moved; the line manager has not yet confirmed that.” Writing “the date moved” would settle a fact you have not established.</p><p>Cooperly’s <a href="https://cooperly.ai/insights/conflicting-instructions-two-managers?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=performance_evidence&amp;utm_content=two_managers_same_episode">guide to reconciling two managers’ instructions</a> makes a related point: a criterion withdrawn for an assignment should not survive as an invisible test at review. First establish what was changed and by whom.</p><p>Once the authorised change is confirmed, the deadline criticism needs narrowing. That leaves a separate question about notifying the event owner. If that responsibility was explicitly retained by the organiser, assess whether they carried it out. If the manager undertook to notify the owner, keep that failure with the manager.</p><p>A new agreement for the next event cannot supply evidence about the old one. Record it as the next expectation. Do not use a freshly clarified responsibility to prove that the employee previously failed a requirement nobody can substantiate.</p><h3>Prepare one note the employee can correct</h3><p>For a disputed episode, try the following preparation note. This preparation aid has not been validated and does not calculate a rating. Use only the fields needed for the material issue.</p><ul><li><strong>Episode:</strong> Name the assignment and relevant dates. Link the appropriate work record rather than collecting unrelated personal material.</li><li><strong>Expectation:</strong> State the criterion in force, any authorised change, and what remained the employee’s responsibility.</li><li><strong>Accounts:</strong> Identify who observed which part, who supplied an interpretation and who repeated someone else’s account.</li><li><strong>Supported finding:</strong> Say what the available evidence establishes, including a relevant strength or a specific shortfall.</li><li><strong>Open question:</strong> Preserve disputed facts, name who will check them and agree when the review owner will return with the answer.</li></ul><p>For the event pack, the note might establish that its date changed with the authorised manager’s agreement and that the event owner did not receive an update. The responsibility for that update may still need checking. Record the successful demonstration separately; it provides context for the priority change and evidence about a different assignment.</p><p>Invite the employee to correct the relevant facts before the conclusion hardens. They can identify an overlooked instruction or explain their contribution; they should not have to negotiate peace between their managers. Keep appraisal material in the company’s appropriate restricted records.</p><figure><img alt="Two overlapping elliptical outlines share one white disc; a separate grey square to the right encloses a white diamond." src="https://cdn-images-1.medium.com/max/1024/1*ueltsYPcwIZJ9vQmHrgfkQ.png" /><figcaption>Attach accounts of the same event to one episode. Keep evidence about a different piece of work separate.</figcaption></figure><h3>Let different findings survive the discussion</h3><p>Suppose the organiser’s invitation list also contains avoidable errors. Both managers agree that checking the list remained their responsibility, and the relevant work supports the concern. Resolving the deadline dispute does not remove that accuracy issue.</p><p>Cooperly’s <a href="https://cooperly.ai/insights/employee-performance-problem-unclear-responsibility?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=performance_evidence&amp;utm_content=two_managers_same_episode">guide to separating work conditions from capability</a> keeps both explanations open. Management can have mishandled a priority change while an employee still needs to improve a particular part of the work.</p><p>The review can therefore recognise the delivered demonstration, explain the changed pack deadline, retain the unresolved notification question and address the supported list errors. Those findings do not need to collapse into one adjective.</p><p>Follow the company’s actual review criteria and approval process when translating findings into an assessment. Do not improvise an average of conflicting ratings or promise a revised score. Where a material gap remains, take it to the authorised review owner and agree the permitted next step.</p><h3>Stop making every reviewer reconstruct the dispute</h3><p>Keep the checked episode note with the relevant review evidence. When the next authorised reviewer reads “several people raised this”, the note should show whether they described several assignments or passed on one account.</p><p>For Cooperly, the Human Operating System for founder-led, AI-native companies, the practical aim is fewer meetings spent rebuilding which promise applied.</p><p>Before your next calibration discussion, take the two claims that seem hardest to reconcile and ask each manager for the work behind them. Use <a href="https://cooperly.ai/insights/new-manager-performance-review?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=performance_evidence&amp;utm_content=two_managers_same_episode">Cooperly’s review-evidence guide</a> to check the basis. If both claims lead back to the same episode, put that episode on the table once and keep each supported finding attached to it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=621ba9a5343a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Ready for Promotion, but Too Useful in the Job They Already Have]]></title>
            <link>https://medium.com/@edkh/ready-for-promotion-but-too-useful-in-the-job-they-already-have-831ba3893939?source=rss-d7b2805cd5b0------2</link>
            <guid isPermaLink="false">https://medium.com/p/831ba3893939</guid>
            <category><![CDATA[career-development]]></category>
            <category><![CDATA[workplace]]></category>
            <category><![CDATA[management]]></category>
            <category><![CDATA[leadership]]></category>
            <dc:creator><![CDATA[Ed Khristus]]></dc:creator>
            <pubDate>Wed, 14 Oct 2026 08:19:32 GMT</pubDate>
            <atom:updated>2026-10-14T08:19:32.532Z</atom:updated>
            <content:encoded><![CDATA[<h4>When old responsibilities block a real career move, give the coverage problem an owner and check that new work actually starts.</h4><p>“We can move you as soon as someone else can handle this.”</p><p>That may describe a real constraint. The employee is now waiting on a career condition they may have little power to change. Who can hire the replacement, reduce the service or decide which requests will wait?</p><p>If someone is ready for promotion and the next role exists, being hard to replace belongs in the company’s operating plan. Keep it separate from their readiness. Then check two results: the old work reaches a workable owner, and the employee can begin the responsibility their new role was meant to give them.</p><figure><img alt="A tall white capsule remains in a small grey U-shaped holder with side tabs, while a larger elevated white holder stands empty." src="https://cdn-images-1.medium.com/max/1024/1*UU8cwuKk2i8Cy7zHdUBjSw.png" /><figcaption>The next role remains attached to the old work until the company changes who carries it.</figcaption></figure><h3>Name the dependency before making another career promise</h3><p>Imagine a designer in a small product company who has been offered a principal design role. The company needs her to shape a cross-product design approach and resolve difficult design choices with project leads. She has demonstrated the relevant judgement and wants the move. This is a hypothetical example.</p><p>She also maintains the asset library, prepares export packs and answers routine requests from colleagues who have learned that she knows where everything lives. The founder agrees the new role, then asks her to keep those duties “until things settle”.</p><p>A month later, the design approach has barely started. The old requests still have deadlines. The new work needs sustained attention and conversations that she keeps postponing. The company has added the principal role to a week still filled by the old job.</p><p>Start with the recurring obligations that prevent the move, rather than a judgement that the designer struggles to let go. Take a recent request. Who asked, what did they need, where did it go and who finished it? Check whether the employee took work back or whether everybody kept sending it to her.</p><p>An old request route needs changing. A missing recipient needs a leader’s decision. A person repeatedly taking back work after a workable transfer needs a separate conversation.</p><h3>Keep established readiness out of the coverage dispute</h3><p>Cooperly’s <a href="https://cooperly.ai/insights/employee-ready-for-promotion-no-role?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=career_transition_operating_systems&amp;utm_content=promotion_old_role_release">guide to readiness when the company cannot offer a role</a> separates the employee’s capability from the organisation’s constraint. Here, the role is available; covering the old work is the unresolved part.</p><p>Make that distinction explicit: “The principal role is agreed. We have not yet arranged coverage for these recurring requests. I own that decision, and I will return with an arrangement on the date we agree.” Promise the action you can take, not a transition date you cannot support.</p><p>Fresh evidence can change a readiness assessment. The absence of a replacement supplies no such evidence by itself. Asking for another stretch project merely to fill the waiting period can give the employee more responsibility while leaving the original block intact.</p><p>If the role itself is still unapproved, say so. If the employee wants a different kind of growth, discuss that first. This approach begins with a real, wanted career move; it cannot create a role or settle its terms.</p><h3>Decide what the business will stop asking of them</h3><p>The designer’s old duties need separate decisions. Keeping approved assets findable may remain essential. Some export requests could use an existing self-service format. Another request might wait until a planned review. The founder should check each option with the people who rely on the output.</p><p>Do not classify all maintenance as work with no career value. Its relevance depends on the role. Operating records may be central to an operations specialist’s contribution; the same activity may displace the work a principal designer is now expected to lead.</p><p>The <a href="https://cooperly.ai/insights/non-promotable-tasks-fair-allocation?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=career_transition_operating_systems&amp;utm_content=promotion_old_role_release">Cooperly guide to necessary helping and career-relevant work</a> makes that distinction useful for allocation. Recognition matters, but praise cannot resolve two commitments competing for the same hours.</p><p>Check the proposed recipient’s existing work, skills and approved access. Agree what changes in their week as well. Sending every routine duty to the next dependable colleague can reproduce the problem one desk away.</p><p>The founder owns the conversation with requesters whose service or deadline changes. Leaving the designer to negotiate every exception keeps part of the old coordination job attached to her.</p><h3>Pair one released obligation with one new contribution</h3><p>For the next work cycle, use a short note that both the employee and receiving owner can correct. This is a suggested working aid, not a validated promotion method. It connects one old obligation with the specific new work its transfer should enable.</p><ul><li><strong>Old obligation:</strong> What recurring output or decision is leaving the employee, and who depends on it?</li><li><strong>Receiving arrangement:</strong> Who accepts it, what time and authority do they have, and where do exceptions go?</li><li><strong>Remaining overlap:</strong> What teaching or support will the employee provide, within which limit, and when does that agreement end?</li><li><strong>New contribution:</strong> What work in the agreed role can now begin, with which decisions, time and support?</li><li><strong>Leader’s next decision:</strong> If coverage or the new assignment remains blocked, who changes the arrangement and when will they return?</li></ul><p>For this hypothetical work cycle, another designer accepts the scheduled export packs using approved templates. Their manager moves a competing internal task and receives requests outside those templates. The principal designer gives one handover session and reviews the first pack; further support returns to the manager for a fresh agreement.</p><p>The principal designer’s first new contribution is a proposal for the shared design approach. She chooses which existing patterns to compare and recommends conventions; the project leads still agree changes affecting their products. The manager protects a morning for the comparison and arranges the discussion with those leads, routing library requests elsewhere during preparation. These are illustrative boundaries, not a standard schedule for another team.</p><figure><img alt="A grey capsule occupies a small grey holder beside task tabs; a separate white capsule sits with room inside a larger white holder." src="https://cdn-images-1.medium.com/max/1024/1*iaczYSfxuGNpouVkr_pSnQ.png" /><figcaption>Check both sides of the move: another owner receives the old obligation, and released capacity reaches the agreed new contribution.</figcaption></figure><h3>Inspect where requests went, then whether the new work happened</h3><p>At the agreed review, look at the export packs and the requests around them. Did the receiving owner complete the covered work? Did questions go to the current decision owner? How much help did the promoted designer actually provide?</p><p>A file completed under her constant direction is evidence of continuing support. It does not establish a completed transfer. Cooperly’s <a href="https://cooperly.ai/insights/manager-to-individual-contributor-handover?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=career_transition_operating_systems&amp;utm_content=promotion_old_role_release">role-handover guide</a> examines this problem in a manager-to-specialist move: useful expertise can remain available while the old responsibility still needs a definite owner.</p><p>Then inspect the other side. Did the principal designer have the authority and time to prepare the proposal? Were the project leads available? Suppose the export packs transferred successfully, but unrelated library requests filled the released time. Coverage improved for one duty; the agreed new contribution still could not start.</p><p>That finding calls for another allocation decision. It does not justify marking the transition complete, or treating the missing proposal as evidence that the designer cannot work at the new level.</p><h3>Give failed coverage back to the leader who can resolve it</h3><p>If the recipient lacks access, arrange it through the authorised owner. If the work takes more time than expected, change scope or competing commitments. If an exception needs expertise nobody else has, agree who handles that narrow exception and account for the employee’s time.</p><p>The employee can document the routine and teach the receiving owner; the leader decides what staffing, service or competing work changes. They should not be left waiting indefinitely for a replacement they cannot appoint. A bounded overlap can be reasonable when the leader names the remaining work, its cost and the next decision. “Until we find someone” provides none of those boundaries.</p><p>Cooperly’s Human Operating System connects working profiles and team context for founder-led, AI-native companies. A manager can use that context to prepare how to discuss support and responsibility during a role change. The company still owns coverage, authority and the promotion decision.</p><p>Before repeating the promise to move a reliable employee, name the old obligation that holds them in place and the new contribution it is preventing. Use <a href="https://cooperly.ai/insights/non-promotable-tasks-fair-allocation?utm_source=medium&amp;utm_medium=organic&amp;utm_campaign=career_transition_operating_systems&amp;utm_content=promotion_old_role_release">Cooperly’s allocation guide</a> to work through the redistribution. Put the unresolved coverage decision beside the leader who can actually change it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=831ba3893939" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>