<?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 Adrien Marienne on Medium]]></title>
        <description><![CDATA[Stories by Adrien Marienne on Medium]]></description>
        <link>https://medium.com/@adrien.marienne_34078?source=rss-5d7aa2d56d99------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*LytH-93NPH8rn6oPoD3fGA.jpeg</url>
            <title>Stories by Adrien Marienne on Medium</title>
            <link>https://medium.com/@adrien.marienne_34078?source=rss-5d7aa2d56d99------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 15:47:31 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@adrien.marienne_34078/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[I know you’ve all seen this message before.]]></title>
            <link>https://medium.com/data-systems-internal-ai-personio/i-know-youve-all-seen-this-message-before-076082b181e7?source=rss-5d7aa2d56d99------2</link>
            <guid isPermaLink="false">https://medium.com/p/076082b181e7</guid>
            <category><![CDATA[system]]></category>
            <category><![CDATA[personio]]></category>
            <category><![CDATA[data]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Adrien Marienne]]></dc:creator>
            <pubDate>Mon, 07 Sep 2026 10:53:11 GMT</pubDate>
            <atom:updated>2026-09-07T11:18:55.186Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xc5L-fiFXER4B8GI5g9Naw.png" /></figure><blockquote><em>“Does anyone know where I can find a report for MRR? 🧵”</em></blockquote><p>I know you’ve all seen this message before.</p><p>Three years ago, at Personio, we were a scattered, decentralised handful of analysts &amp; engineers answering the same question multiple different ways.</p><p>Within minutes, this Slack thread has three different answers. “Ask the analyst sitting in Finance.” “Oh, but the analyst in GTM says to use this sheet instead.” Someone links a Tableau dashboard. Someone else says that one’s outdated — “try this Snowflake view”.</p><p>Nobody’s wrong, exactly. That’s the problem. Every org had its own analyst, its own reporting standard, its own version of the truth, and no shared foundation underneath. If you wanted a number you could trust, you needed to know exactly who to ask — and hope they hadn’t switched teams in the meantime.</p><p>It was never really a data problem. It’s an access &amp; standardization problem. And it was limiting how well Personio could use its own information.</p><p>Today, we’re Data, Systems &amp; AI — three teams, one department, one mission. Here’s how we’ve evolved.</p><h3>1️⃣ Bring Data together [ D • • • ]</h3><p>So the first move was straightforward: pull Data into one centre of excellence team, with one shared foundation underneath.</p><p>✓ A scalable Snowflake warehouse and dbt pipelines instead of five disconnected views</p><p>✓ Standardised reporting, data governance and data enablement</p><p>✓ Best practices, applied consistently, not team by team</p><p>✓ Scalable operational excellence for ways of working</p><p>✓ A team culture with trust, innovation and momentum</p><p>✓ Keeping domain knowledge through embedded PODs</p><p>That same question in #random-global today gets one answer, in the first reply, from a source everyone actually trusts.</p><h3>2️⃣ D becomes DS [ D S • • ]</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/0*EnxwqDe4vJi5eqpV.png" /></figure><p>Fast-forward to September 2024. Two new messages show up.</p><p>In #team-data: “Looks like a field got deleted from the Salesforce account object? Who in Systems should I talk to?”</p><p>In #team-cx-systems: “Who do I ask about the rETL value being pushed into Zendesk?”</p><p>A source system changed, nobody downstream is aware, and three steps later someone’s metric stops adding up with no idea why.</p><p>That was Data and Systems before they merged: valid changes being made, but a lack of communication up and downstream. The teams didn’t know each other or fully understand the dependencies.</p><p>In 2024, not only were Data &amp; Systems misaligned, but each system team was embedded in its own department’s ops function — not talking to each other efficiently, no shared best practices, no shared resources, no smooth integrations between the platforms everyone depended on.</p><p>Here’s the thing about Systems that sometimes gets overlooked: every platform Systems owns — Salesforce, HubSpot, Zendesk, Chargebee, Workday — isn’t just infrastructure. To the Data team, It’s a data source. Every one of those systems is somewhere Data has to pull trustworthy numbers from. So Systems was already, whether we called it that or not, the team producing the data Data depended on — and at the same time, equally importantly, these systems help to shape the operational processes the business runs on. These are the processes that data teams are trying to influence decisions on. That dual role is exactly why the two teams belonged together.</p><p>I now cannot imagine a world where Data and Systems don’t sit side by side. Now we see;</p><p>✓ Stronger Alignment &amp; Dependencies<br>✓ Improved Governance &amp; Quality<br>✓ Shared Architectures<br>✓ Shared ways of working &amp; Improved Team culture<br>✓ Shared Resources</p><h3>3️⃣ DS becomes [ D S A I ]</h3><p>Another year on. May 2025: “Personio wants to become an AI-first company.”</p><p>The logic here is the same shape as part 2 — just one level up. An AI product needs the same three things a trustworthy metric needs:</p><p>✓ clean data to reason over<br>✓ systems actually built to let it act<br>✓ platform underneath it that doesn’t get reinvented every time someone wants to build something new.</p><p>Split those three across separate teams and you get the old problem again: an agent that hallucinates because nobody aligned it with the source system, or an impressive-looking AI product that can’t actually do anything because the system underneath was never built to be acted on.</p><p>For the first year of rolling out AI in Personio, the D&amp;S team naturally took a lot of the drive. We ran AI surge weeks, Internal AI training sessions and an AI council committee.</p><p>Then officially in June 2026, Internal AI was set up as a team to join D&amp;S. This wasn’t a nice-to-have org chart decision — it’s the difference between AI that’s trustworthy, scalable, embedded in day-to-day operations and a vibe coded AI solution that just looks good in a demo.</p><p>And the relationship isn’t one-directional.</p><p>✓ Data is what makes AI worth trusting — clean, well-modelled information</p><p>✓ Systems is what makes AI worth having — the difference between an agent that can describe a problem and one that can actually go fix it.</p><p>But it also runs the other way: building real AI products puts a new, much higher bar on both of the others. An agent making a call in seconds exposes a stale field or a fuzzy metric definition that a human skimming a report would never notice. If you’re aiming to build a workflow-integrated agent, it needs to be wired into the Systems flow, not bolted on beside it. It’s changed the way the Systems team thinks: less about reporting, more about action.</p><h3>Three Key Takeaways</h3><p>✓ One source of truth beats local silos. If two people in the same company can pull two different numbers for the same metric, you don’t have a source of truth…</p><p>✓ Unify Data and Systems before you ship AI. If your AI roadmap and your systems roadmap are owned by people who’ve never met, you’re not ready to ship.</p><p>✓ Build for action, not just reporting. Data’s future isn’t about being replaced by AI or becoming more of an insights function — it’s about becoming operational and making changes in the day to day processes that make a real difference. That’s the thread that ties all three chapters of this story together.</p><p>I’m super excited about the journey of DSAI in Personio and excited for whats to come!</p><p>This publication is where we’ll be writing about the great work and impact happening across DSAI — the wins, the near-misses, the things we’d do differently next time.</p><p>DSAI</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/0*IyxAwCQ6fWJ1vQFr.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=076082b181e7" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-systems-internal-ai-personio/i-know-youve-all-seen-this-message-before-076082b181e7">I know you’ve all seen this message before.</a> was originally published in <a href="https://medium.com/data-systems-internal-ai-personio">Data • Systems • Internal AI @ Personio</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Test]]></title>
            <link>https://medium.com/@adrien.marienne_34078/test-496eec240311?source=rss-5d7aa2d56d99------2</link>
            <guid isPermaLink="false">https://medium.com/p/496eec240311</guid>
            <dc:creator><![CDATA[Adrien Marienne]]></dc:creator>
            <pubDate>Wed, 12 Aug 2026 09:57:23 GMT</pubDate>
            <atom:updated>2026-08-12T09:57:23.967Z</atom:updated>
            <content:encoded><![CDATA[<p>Test</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=496eec240311" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>