We tried really, really hard to not write another article about AI, but we just can’t help it, it’s really all-consuming at the moment, it has taken over UI Design, UX research, process flows, content generation, media generation, literally anything in regards to the digital world has been consumed, subsumed and brought into the AI blob, and we’re all just observers.

So enough with the emotional turmoil, let’s get to the fun part. It’s time for you to meet Google’s Gemini Canvas, which we’ll just call Google Canvas for short. And let me explain to you why it could possibly allow you to never put effort into UI and UX design ever again.

What is Google Canvas?

Google Canvas is an interactive, side-by-side workspace within Google’s Gemini website where you can ask Gemini to write, code, and design projects together with you (or for you) in real-time.

Basically within a website: ‘on the left’ you have a prompt box where you ask the AI to create a website/form/application and ‘on the right’ you have the resulting fully interactive interface, coded beautifully by Google Gemini’s AI.

Why this is amazing

Let’s be clear: Google Canvas is not a prototyping tool, it’s a final interface, fully coded, fully complete, you’re basically done (apart from the required backend/API connections, which you can let the developer worry about).

Normally as a UI/UX designer you would need to worry about:

• 100 edge cases
• What each state looks like (hover state, visited state, disabled state)
• How the system looks on desktop, tablet and mobile
• What is best practice
• What does great copy look like
• Which is the best component to use (radio box, checkbox, drop-down)

There’s just so many considerations, and ensuring they’re ‘best practice’ and accessible and consistent, it’s dizzying the amount we UX designers need to focus on when building systems, forms or layouts.

But now your job has changed, congratulations!

If we are to define the role of a UI or UX designer, a considerable portion of that role includes research and applying best-practice-principles, verifying the feasibility of functionality and aligning with specified frameworks or accessibility guidelines. As mentioned above.

But by using Google Canvas these requirements are removed. The code generated in Canvas -for the most part- is responsive, accessible and follows best-practice-principles. The copy is clean, concise and can tailor itself to the user persona or user group (if you tell it to).

It’s almost perfect, and in another 2 years it will be 99% perfect. Which is considerably better than what most UX designers can achieve.

No more guessing, no more critical thinking (oh wait, that’s bad)

So by using Google Canvas to create our system and form layouts, we actually don’t need to do all that critical thinking, we don’t need to worry about best-practice, we just ask Canvas to make it, it provides the coded, working system and you’re done.

So that’s great, but let’s not forget this basically means this tough mental work (that us UI and UX staff do) is gone, we’ve removed critical thinking… though does that mean we are actually making ourselves mentally sloppy and weak? Yes, yes it does. Our brain is much like a muscle, we’ll need to discover other ways to flex it.

Why this is a game changer

Previous AI large language models would create concept images or concept code, or concept prototypes. But none of them took that one extra step to produce finalised code with a beautiful design that’s accessible and responsive (yes we’re going to keep saying things need to be accessible and responsive as they are the foundations of good design). But Google Canvas just does it, in seconds.

This means not only can you concept out ideas, but you know instantly if they’re feasible from a user experience perspective and also from a development perspective. No more of those frustrating and irritating conversations with developers where they claim something isn’t possible (even when you know it is).

Why you’ll never need to work again

So why the overly optimistic title? Because it’s true, if you wish you can now ask AI to make your interfaces and you can sit back and do nothing.

Yes, you may need to tweak and align with brand guidelines and adjust to fit your requirements, but these are all little items that will become automated too in time.

So just sit with your feet up, let AI do the work for you and focus on deeper and more constructive work, or just get lazy and sloppy while still getting paid.

And if you get fired for being too lazy, just get Google Canvas to make you a new portfolio website instantly, and get ChatGPT to send out your resume and deal with agents. Fun times.

How to access and use Google Canvas

1. Log into: gemini.google.com (you may need a paid subscription)
2. In the search box click the ‘tools’ icon
3. Click the ‘Canvas’ icon
4. Type into the search box what interface you want created
5. Wait for it to code your new interface and you’re done

Happy New Year! We love all you UI and UX designers and we want 2026 to be your best year ever! So let’s help you make that happen with a selection of templates for you to use and misuse however you like.

Here are our top 5 most popular free Figma templates for January 2026:

5. Cafe Website Template

Simple but impactful, this cafe website template is super flexible and can be used for a wide variety of different cafe or restaurant styles such as Greek or Italian, steakhouse or fine dining (with a few tweaks).

Just make sure you have a great set of photos to add to this template, to make it truly sing.

> Get Template

4. Simple Website Builder

All the building blocks you need to create a simple but enticing website design in seconds. You can literally create an entire website structure and design with zero effort, just drag and drop and you’re done.

> Get Template

3. Hotel Booking App

It’s great to see a mobile app design making out top 5 list, usually people just want website design templates. But it’s understandable why this free Figma template is so popular. It’s basic but professional, refined but accessible.

Also it’s good to see a dark theme every now and then, mobile app templates are generally and white themed and very similar in look and feel.

> Get Template

2. Education Website Template

Another simple theme, but super helpful if you’re looking to build some kind of educational website design using a Figma template as the base.

It’s nice, and it’s light and it’s filled with all of the design components you would need for this unique eLearning and education niche, such as class details and booking, course browsing and lessons functionality.

> Get Template

1. Free Hero Banner Designs

Oh it’s nice to finish on something beautiful and fun. This set of 50 landing page, hero section designs is just wonderful. Most of them are of a high quality, and you’re almost guaranteed to find something useful here. There’s also a few more hero designs here and here too.

> Get Template

In Conclusion

This month is all about dark themes, simplicity and beauty. All of which we love most of all at Free Figma Templates.

Let’s start off the year with a bang! And make some truly beautiful UI designs. Use these amazing 5 templates as a base, and build upon them to making something special. Good Luck and Happy New Year!

So we here at Free Figma Templates have been heavily vibe coding lately, creating an extremely addictive and fun, cosy, card collectible game: deckcaster.com, but one of the biggest highlights of this experience has been creating AI generated music to make the game more interesting and fun.

We tried a few different free LLM AI music generators, but the king for us has been Suno.com and no we aren’t paid to say this. For a variety of reasons, but mainly because it’s just so fast (5–10 seconds), and the music generated is for the most part excellent, and we mean truly excellent.

You can add in a short description, click generate and you’re done. Though of course that creates songs that are catchy but useless for us. So what do we actually do? We get Google Gemini to make more in-depth and complicated compositions and lyrics and we simply paste that into Suno’s ‘Lyrics’ and ‘Styles’ text areas and click generate.

Our steps to creating a beautiful, relevant and high quality song in seconds for free:

1. Open Google Gemini and type: “I want you to create a new song for me, make it [insert description here] and be inspired by [insert songs, singers, bands or games here e.g. Expedition 33], I have created lore for my [game/system/song] and here it is [paste all your lore and attach any relevant images, this lore section can and should be thousands of words long].
2. Once it provides to you a set of lyrics and a ‘music style’, you simply review the lyrics and adjust/tweak them and then copy those 2 items into Suno.
3. Click generate, and 2 songs will be made for you to choose from, if you love one download it! If you don’t, simply click generate again, or go back to Gemini and get it to create a new set of lyrics or song style. Since we aren’t making our game for profit, licensing isn’t an issue, but if you are making something for profit, you will need to upgrade your Suno account.
4. Once you have your song downloaded you can edit, tweak or adjust in sound mixer software, or if you’re lazy -like us- simply compress it online using a website such as: onlineconverter.com/compress-mp3.
5. And then since we’re vibecoding in Google Antigravity (the best free Vibe Coding software currently online), we tell Antigravity where the mp3 file is located on our hard-drive and we ask it to upload the file and add a play button to the relevant website page, and it’s done! So quick, so fast and so much fun.

Now why are we making an article about all of this? Because we’ve now hit a unique moment in the future of technology and creation. And that moment is happening today, not last week but today.

• We can now create pixel perfect imagery and user interfaces in seconds with no design ability thanks to the latest versions of ChatGPT and Gemini which just came out last week
• We can create perfect, catchy and funny music in seconds with no musical ability thanks to Suno and other similar music generators.
• We can then combine these images, audio and interfaces to create a well coded (not quite perfect) website, game or system using Google Antigravity vibecoding software that also just came out a few weeks ago.

Literally in the last 2 weeks everything has changed, the quality of all types of AI generators from visuals to audio, development to design has reached a level that surpasses many professionals work, and it all happens for free and instantly. Just wow!

And to top it off, some of the audio and lyrics being generated are actually funny! AI is actually starting to get funny too. We’ve actually been laughing at the musical numbers we’ve been generating for deckcaster.com, though that’s only for a couple of songs, we’re trying to keep Deck Caster very serious and realistic, whilst also being high fantasy and Eldritch, but we’ll talk about that in another post.

In Conclusion

The time to get into AI is now, today, not last week or last month, it’s happening very fast and the train won’t keep stopping for you. Jump on, have fun and have a laugh too.

Before we begin let’s raise this early: Figma Make is currently not limiting prompts, but they will next year! So there’s still time to heavily test and play with this stunning piece of technology for very cheap (just subscribe to Figma for a month and cancel if needs be).

What is Figma Make

We always need to start with an explainer in reviews so here it is: Figma Make is a way to create websites, apps or systems using nothing but prompts. So you can say ‘make me a portfolio website written in a doctors style as I’m first year practising ‘ and it does it, or ‘make me a flappy bird game but with tanks’ and it just.. does it. It’s wild.

The team here at freefigmatemplates.com have been making websites and apps for decades and each one has taken months to build and refine, that was the standard. But in the last 4 days we’ve made a new card collectible game that is more complex than any website we’ve made, it would have taken at least 6 months.. we vibecoded it in 4 days and it’s better than anything we would have produced by far. This is wild!

If you’re a designer you may have had friends walk up to you and say something like ‘hey you’re a web guy.. I’ve got this great idea for an app. can you make it for me for free? I’ll share the profits’ it’s as tale as old as time.

Now you can make their app, in 1 minute, for free, with no effort. I mean woah! we’re really starting to lose our minds here. This is beyond future tech. This is verging on mysticism or magic.

Even though we know the technological foundations we’re still beyond impressed.

So getting back to Figma Make, what is it’s point of difference? Well we’ve tested the largest 6 AI coding programs and have found Make to be the easiest to use whilst also producing great outputs. We tested 3 prompts on each of the 6 AI LLMs and Figma Make produced the best outputs on average.

Though it’s worth noting these systems generated extremely similar outputs, some even using almost identical colour schemes.

We would love to get into the reasons for this (shared LLM backend model, shared code it had referenced) but that’s another article in and of itself.

Also Figma Make have heavily focused on removing the usual set-up barriers that come with vibecoding. Usually you would start a vibecoding project by asking the system to set up code packages and APIs, it’s overwhelming for non coders. But Figma Make have that base set up already done as it plugs in natively and instantly to a backend system for login/email/data storage, so the typical backend setup issues that crop up while vibecoding have already been resolved before you start.

The Bad

So it’s easy, fast and provides great outputs. What’s the negative? Well with any nascent tech there are a lot of negatives, and you need to choose if these outweigh the benefits.

For us they don’t, but for other more serious coders they definitely will

Figma Make’s Negatives:

1. It costs money

There are ways and means to do vibecoding for free though those options are usually limited by the number of prompts you can give. Figma costs around $20 a month but we anticipate that doubling or doubling again next year if you are vibecoding heavily and hit their prompt caps.

2. It’s for prototypes not final systems

Let’s be clear: you can make a website prototype in Figma Make in seconds but converting that to a proper hosted website is difficult and fiddly. This is clearly a massive issue, but we do predict Figma Make version 2 will include native publishing, so that issue should *fingers crossed* be resolve in time.

3. It focuses on ease-of-use over technical configurability

By pre-configuring and pre-installing code and APIs, you start your vibecoding experience quickly but you miss out on flexibility of choice. You can of course adjust afterwards, but it really isn’t designed for that.

Figma Make wants you to stay in their ecosystem.

4. It’s a pure vibecoder

Figma Make is for UI designers, clearly. They’ve specifically made it for creative people not technical people. It really pushes you hard to code through text prompts or clicking on content on the screen and commenting.

Do not expect to quickly flick between beautifully laid out, easy to edit code and a vibe code prompt box, other systems are set up for that. You just get the prompt box here.

5. It hides its reasoning

When you make a prompt the reasoning text pops up and disappears in seconds, you can expand it but we find it’s very lightweight in its descriptions.

6. It limits your uploads

Allowing just 2 uploads per prompt (just 2 images? What if I want to upload a photo gallery of 50?) It really makes the process noticeably slower than it could be.

So that’s a fairly big list of negatives, but let’s be clear here Figma Make is months not years old, a solid ‘version 2' would fix all these issues and much more. That is, if they keep iterating and the Figma team genuinely care for its success.

Which we’re dubious about. We are actually not confidant that Figma will fully support Figma Make, so be sure to hedge your bets and save your code elsewhere.

In Conclusion

Figma Make has been largely ignored but despite that it is still clearly revolutionary. It reduces months of hard and technical work (which freelancers would charge you tens of thousands for) into something you make on your lunch break, with zero technical skills required!

The future is wild and we’re diving right in, be sure to join us.

Talk about the easiest article to writer ever. There may be millions of UI/UX focused Medium accounts out there, but the best of the best really stand out noticeably.

Why? Because they’re clearly wise and have a deep and broad understanding of UI and UX, and we love to read every article.

One other important point, these are our top 5 non-paywalled Medium writers. As we feel that people who paywall 100% of their content aren’t being very community minded.

So let’s get to it, here’s the top 5 Best UX Medium accounts for 2025 and 2026.

5. Bootcamp

If you like content heavy, in-depth analysis then Bootcamp is a great place for you. Its writers seem to have huge amounts of experience, providing for really great insights. The one little negative, is they can almost be too technical, for example this article: “Building a haptic feedback vocabulary” is a bit too niche.

> View Medium Account

4. Ryan Almeida

The only individual Medium account on this list is for this fine fellow named Ryan. Why? Well mainly because he updates fairly frequently and his content resonates well with the UI community (he gets lots of claps for his articles).

That being said both Gemini and ChatGPT say his content is: “a 90–95% likelihood this text was generated by AI or AI-assisted.” So does it still have value? We’ll let you decide… p.s. our article you’re looking at right now was not created using any AI (except the images), though you can probably tell.

His content can be in-depth and also lightweight, and we also like his short and sharp way of writing with lots of bullet points and succinct copy. Though once again that’s probably just chatGPT doing it’s magic.

> View Medium Account

3. UX Collective

One of the big guys, UX Collective is one of the most popular UX accounts on Medium. It’s a little more accessible than Bootcamp and a little more random.

They also deserve props for their article writing as they lock in on viral moments and make great content out of it, such as ‘What designers can learn from Zohran Mamdani’s historical campaign’.

Our one issue, is that due to their focus on producing a lot of content, the quality of that content can be lacking sometimes, but despite that it’s still worth a follow.

> View Medium Account

2. Muzli

If you had to follow just 1 UX account we would say follow Muzli, as it’s been providing fun and engaging ‘UI focused’ weekly updates for years.

If you just want a lightweight overview of UX trends then Muzli is perfect. Need something more? I’d go somewhere else.

Muzli does an impeccable job of sourcing the latest and coolest websites or interfaces, and then compiling them into an easy to browse format. Just lovely.

> View Medium Account

1. UX Planet

Truly a 1 stop shop, UX Planet doesn’t provide as much content as the other big UX Medium accounts, but it’s content is more fun and entertaining.

They work hard to source, combine and distil content into manageable and useful articles, so as a pure UI and UX resource it’s amazing and our №1 pic for best UX Medium account.

> View Medium Account

In Conclusion

Despite AI copy starting to appear in Medium articles, leading to reduced quality, there are still great human writers out their providing in-depth knowledge that’s highly valuable. So be sure to reward them and give them lots of claps and follows (and please clap this too).

Gemini 3 (Google’s AI Large Language Model) has rolled out and it’s smashing all the AI benchmarking tests.

It’s making competing AI businesses extremely worried as they just realised they need to catch up, and fast or Google Gemini is going to leave them all behind if Google keeps iterating and improving as fast as they currently are.

That being said… we still find OpenAI is clearly better at copy generation and that’s our most used feature of AI LLMs. For some reason Gemini’s copy content just feels… off. Maybe because ChatGPT scrapped Reddit? Which is a great resource for modern communication and digital conversation, despite the fact it’s now overflowing with AI bots.

Whatever the facts, it’s still worth highlighting that ChatGPT still has some unique benefits that Google Gemini needs to reproduce.

But we digress, what is all this about us predicting Gemini’s newest feature? Well it’s about this article on Ethereal UI where we proposed that standard user interfaces will be replaced with ones that are generated by an AI LLM, used once and then deleted forever, and that’s exactly what Gemini has introduced in their latest model.

Rather than outputting basic text, images or video on request, Gemini 3 creates generative user interfaces to convey the information. They’ve actually started applying ethereal UI! That was faster than we expected.

A simple way to explain it is that generative user interfaces are mini websites or user interface screens integrated into an AI prompt window.

So if you were to prompt: “What are the best locations to travel to in Italy?” rather than a wall of text (that ChatGPT or Microsoft’s Copilot provides) you would instead get a nice little visual interface that shows photos and descriptions in structured tables. It’s just so much easier to absorb!

We here at freefigmatemplates feel that ChatGPT, Copilot and Gemini are extremely similar metaphorically to MS DOS as it’s basically a pure text input/output system with no visuals. But now Gemini 3 has us moving to Windows 95 with the application of rich user interfaces in its results.

So what’s next?

If we have only just reached the Windows 95 era of AI LLMs what is the next phase? What shall we predict next?

Well we’re not going to predict a new form of agentic AI. Yes, there’s a lot of hype around them, but we feel it’s not really what people want from AI.

So what do we really want? Well brace yourself…

Introducing Micro AI

We think (the teams here at Free Figma Templates and Free Powerpoint Decks) that AI is about to get smaller not larger, targeted not comprehensive, micro not macro.

We’re predicting mini-AI activators are going to become a thing, and we’re calling them AI sprockets (little gears that move things along). They’re little AI helpers sprinkled around your systems, websites and mobile apps.

We’re not talking about one overarching AI system that takes over your computer but the opposite: tiny, micro assistants for small targeted activities.

For example you don’t need AI to automatically fill in a form as they are usually auto-prefilled nowadays, so AI is useless but at the end of the form it may say ‘do you require additional support’, or ‘do you have a specific enquiry?’ or ‘do you need special assistance?’ then clicking a little AI sprocket can fill this in for you instantly, with your specific requirements, now that is immensely valuable.

The future is mini activities completed by micro AI assistants.

So what we’re predicting in the next wave of AI updates won’t take us from Windows 95 (basic UI interfaces) to Windows 2000 (prettier ones), we’re actually going to move from Windows 95 desktop to Samsung/Apple mobile. Simple, small, lovely applications of AI.

We think Micro AI is going to be massive and yet noone is talking about it yet. So congratulations you’re on the ground floor.

In conclusion

AI is moving significantly faster than we expected but the real world value/benefits of AI is materializing at a glacial pace.

So rather than let yourself be pushed along slowly by 1 or 2 giant AI systems, let’s build micro AI sprockets that will speed up productivity in 100s of microscopic ways, leading to massive productivity gains overall.

Hey we were right once…. let’s see if we are again.

If you’ve been involved in website, mobile app or system development or design in the last 10 years you would have heard the word Gamification, and hearing it right now you may sigh a little.

Gamification was the hero of digital design in the 2000s and 2010s. Points, badges, progress bars, daily rewards, and endless streaks promised to turn even the dullest tasks into exciting user experiences.

Businesses rushed to add game-like mechanics to apps, websites, and internal tools. And for a while, it worked, it was fun and silly and many of us were battling with our friends to get the most points on some random app or website.

But then something strange happened, it just kind of faded out of existence. All those fun little quests or mechanics slowly were removed from our apps and we kind of just forgot about it… which is weird as it was meant to be the key way for your website, app or system to become successful and engaging and popular. So what happened?

Here’s our top 5 reasons why gamification went the way of the dodo.

And maybe after we all review the lessons learned of its demise we can bring gamification back, but better… actually no let’s not do that.

1. The games weren’t addictive

When a sales and marketing man decides to recommend adding gamification to their companies’ app, they’re really not thinking about why and how this feature is important. They just think *let’s add some scoreboards and let people share their success on social media, it’ll make our system way more popular* they don’t think *hey, let’s make this thing amazingly entertaining, addictive and long lasting*.

They were so short sighted they didn’t consider what actually makes something addictive and replayable and also something that lasts. Sharing your score is fun for a week, maximum, after that it’s meaningless.

Some things that could have made them actually addictive:

• Daily/weekly challenges
• Login Streaks
• Unlockable content
• Leaderboards with tiers (not all-vs-all)
• Loot boxes (that are a little evil)
• Personalized quests and team quests

2. The rewards weren’t valuable

Rubbing your friend’s face in your high score was fun, but then what? All these points and badges after a while really felt meaningless, because they were.

A genuine reward is when you attach something real and emotional to your high score. Being the envy of your friends is alright but paper certificates, real world prizes or esteem among the community hold more value. And these badges really didn’t provide for that as they were only tied to a single app. And let’s face it, apps don’t have a very long shelf-life, especially apps or websites that used a lot of gamification such as fitness or personal improvement sites.

If the creators of these gamified elements truly wanted to highlight that these scores and badges were valued, they would have highlighted this in some form, but they never did.

If the owners of these apps/sites/systems don’t care about your score, then you definitely didn’t feel like you should.

3. Gamification was an afterthought

As gamification as an idea grew in popularity a lot of already successful apps decided to shoehorn and shove the feature into their already existing products.

The effect of this was that gamification didn’t feel like an ingrained part of the process, it was a sloppy add-on.

4. The creators weren’t gamers

Now this is a huge one. Creating games requires the creator to be a gamer (of any form not just digital gaming). Once you play a few thousand hours on a Playstation, Nintendo DS or Switch, Xbox or Amiga (yes we’re that old) you start to know innately what is an entertaining gaming experience, but a lot of the people assigned to making these gamified elements were busy developers, many whom have little interest in gaming (not all obviously, but a good percentage), they just want to do their assigned development task and move on, that’s not a recipe for fun or success!

5. Times have changed: function is now king

We are now in a world of purpose-driven-design aka no more fluff and no more fun. Which is sad but also a fact of reality. Systems are meant to provide a service instantly and easily.

If you want to buy shoes you just want to select your shoe size and click buy. That’s it. No badges, no awards, and because of that fact any system rolled out in 2025 and 2026 is boring and simple with white backgrounds and sans serif fonts. Gamification does not slide easily into that shoe.

In Conclusion

The real demise was not the concept itself but the careless way it was applied. Not everything needs gamification, it’s a niche bit of functionality with only a few use cases, and the application of it needs to be made perfectly.

Can gamification come back? To great success and love by everyone?

Actually no.

But it was fun while it lasted.

After 20 years of UX usability testing and research (omg I’m old) here’s some tips that even today I find especially useful.

You can easily find basic best practice UX testing tips from ChatGPT, but my tips are real world ones that I come across daily, so -I think- are more valuable.

Author: freefigmatemplates.com UX Design Lead

Before we begin, let’s define Usability Testing or User Testing: basically it’s a scenario where you have a tester (either online or in person) test out a prototype or wireframe, of a website, system or service and you (the interviewer/researcher) ask this tester questions before, during and after they test the prototype. It’s a standard part of UX and UI design but sadly isn’t always actioned in a system build (unbelievable but true).

UserGuiding reports that only 55% of companies conduct any user-experience (UX) testing. - Source

So let’s get right to it, here are my top 10 usability testing tips for beginners.

1. Ask the tester to think as themselves not try to represent a wider group

There’s a weird thing that happens when someone gets used for usability testing a few times; they start to think like a UX designer and not like a normal person using a system for the first time, they stop saying ‘I don’t understand this button/wording/feature’ and instead start saying ‘people wouldn’t understand this button/wording/feature ’, when you start to see this happen, you need to stop them and tell them specifically you want their personal experience not some mythical group of users they have in their head.

2. Have a script

No matter how prepared you may be, it’s a common occurrence to jump into a usability testing session late or unprepared (at least it happens to me a lot). A script allows you to switch off your brain and just work on autopilot. And once you feel comfortable then you have the option of moving away from your script. But having it as your base is always such a huge bonus safety measure.

3. Be comfortable in silence let things get awkward

Most humans can’t achieve this basic thing: silence and inaction. We need to fill silence with noise and moments of confusion with support, we can’t help ourselves when people are in awkward moments we have a huge desire to help them, but you won’t identify issues in usability testing sessions without the user experiencing a moment of awkwardness, confusion and frustration and you can’t interject with a solution, these are the moments when we (researchers) identify genuine pain points and moments of learning.

So let them stew in their confusion and be silent, simply take notes of what you see and hear. This moment right here is where insights occur.

You need to let users drown (make the wrong decisions) so you can learn how to make them swim.

4. As a researcher: do not nod and do not say ‘yep’ when watching a user test a system

I see this all the time with junior UX researchers, as the tester gets stuck and doesn’t know what to do next they ask the UX researcher hosting the test: ‘am I doing this right’ or ‘do I click this now?’ and the researcher will naturally nod or say ‘yep’. Which spoils the whole point of the test!

We need them to get stuck and question things. Why? Because when the actual system launches there won’t be someone next to them nodding when they’re stuck. So don’t help them at all, just watch and say nothing, only redirect testers when they completely go in the wrong direction and literally give up.

5. Show interest, be interested, don’t be a robot

Once again junior UX researchers often over-focus on taking notes properly and not on interacting with the tester. But testers are only truly honest with you when there’s some level of connection with you and friendliness.

So you may not be saying ‘yep’ or nodding, but you can still be a human and chat with them.

Focus more on the user than on making accurate notes. As long as you’re recording the session you really can be nonchalant about note taking and instead focus on the person and building connection.

6. Start off with a smile

I never start a session by talking through my script. I start with general ‘how are you’, ‘how’s your day’ questions. And if they have a dog or cat ask them about them too, that’s a guaranteed ice breaker.

Your first 3 minutes are for fun conversation.

Once you’ve got to have a brief chat you’ll notice the tester relax a little, then it’s time to say ‘OK I’m just going to run through a script’ and go for it.

7. Only ask open questions, no leading questions

This one seems easy but gets overlooked when you’re rushing to build out your Miro or Figjam interview question board.

And it’s rather easy to fix, just 1 hour before your interview re-look at your questions and for each one: ask yourself is this question leading or open ended?

Leading question: Do you like this form?

Open ended: What do you think about this form?

8. Instantly engage in a post interview review. And take additional notes.

If anyone else was involved in the testing session such as an additional note taker or product owner who was observing, once the interview is over and you say goodbye to the tester, straight afterwards call the additional note taker separately (or have a face to face chat) and just start talking about what you just learned.

Some amazing insights usually come from this post interview review. Be sure to write notes during this 15 minute internal post-test review. There’s gold there.

9. You don’t need to validate or dive into every comment

There are some moments during testing where users can start to mention personal details or issues, and in some instances it is OK to ignore or move past these comments quickly.

For example one user mentioned during an interview they were having issues with owing money. I had an option in this moment to talk about the comment but I really wasn’t trained to support this and it wouldn’t have helped the situation, so I just acknowledged the moment and moved onto the next question.

Asking them to delve deeper into non-prototype/emotional areas requires different training (trauma informed care is a common area of training regarding this) and if you don’t have that it’s best to either move on or ask them if they would like to pause or stop.

Don’t try to be an amateur psychologist, you’re not trained for it.

10. Have a messy work/note taking board

Miro or Figjam, Figma or Excel, whatever you use to contain your interview notes and questions, it’s ok to have this area as a total mess, I actually encourage it. Usability testing is messy and can go in extremely random directions. Don’t worry how this board may be perceived by others. Let your board get messy. Just ensure your final report or presentation is spotless.

In conclusion

Usability testing is truly amazing and useful. It forces you to think about your system in a completely user centric way and highlights all the biggest issues of your system that users will likely encounter (which you can now fix… now that you’re aware of them).

Don’t ever let your business owner or product owner skip over usability testing due to time constraints, do it on the weekend if you have to, or recruit your friends and family if there’s no budget.

It is truly vital to the success of the system and your own professional success (as poor systems can reflect poorly on you the UX researcher), so be sure to apply usability testing to every system you build and at every stage of the design life cycle.

Test, refine, test refine, test refine and you’re done. Just like that, your sytem is perfectly user centric and battle tested.

Congratulations your UX team is growing. It started as 1, just you and it’s now reached 5! And you’ve plans to double every year! That’s overwhelming and scary… what are you going to do?

Well luckily for you I’m going to give you some support. Let me be your builder, setting the foundations for your team to grow and flourish.

Who am I? well just a senior digital strategist in government helping build foundational structures for tens of millions. I’ve been doing this specific role for 6 years and would love to share some hard fought learnings. So here’s my preferred 10 step program to creating the foundations of a successful UX team, that can grow exponentially. Note: this was not written with AI so excuse the poor phrasing!

Step 1. Understand your team

Before you do anything you need to know who you’re working with, so start with questioning each UX team member, ask what their strengths and weaknesses are, in every topic from software to strategy and communication, and what they like to do, what they don’t like to do, how they like to work and where they want to go in future.

This will form the basis of all decision making for your teams future such as ‘what ability gap do I have in the team’ ‘who can I train up in that gap’, or ‘do I need to hire someone new’.

Step 2. Understand your capabilities

Now let’s understand what the team can achieve. Can they do service design, product design or just UI design? Can they support developers to build complete systems or is your team only able to produce mock-ups or digital adverts?

Type out your findings, affinity map your notes and define a capability list and highlight how powerful you are on each item in the list.

Step 3. Understand your service offering

Now this is where we need to think business strategy, based on your understanding of your team and your capabilities you now should produce a presentation that defines who your team is, what your team can do and what business services they offer. Remember ‘good at Figma’ isn’t a service offering, but ‘product branding’ is or ‘website design and build’ is another service offering. So basically this service offering preso is based on a combination of team capabilities and service capabilities.

Let’s be clear: this service offering presentation is your most important asset, as the ability to clearly define your service offering directly leads to new business and new clients.

Step 4. Define your principles and a team charter

So here’s a crazy stat: teams with a team charter -even one that’s termed ‘bad’- are 14% more productive.

So we recommend going through the collaborative process of creating a set of product principles or a team charter (either is fine), though a charter is considered a bit ‘90s nowadays.

Step 5. Build or select a design system

I’m going to assume you already have some brand or branding, so the next step is a Design System (DS). We recommend doing one of 2 things: Either create an ultra-basic design system that’s just fonts, buttons and colours (1 day build time).

Or use a pre-made design system template online and restyle to suit your needs (2 week build time), though be wary if you need lots of tailoring the effort involved in tailoring a complex DS template can be extreme.

Step 6. Define standard processes

A million questions are going to arise during the development of systems or apps that focus on consistency. So you need to centralise these decisions and put rules around them.

Some examples include:

• Do we use sentence case or camel case on form field labels?
• What tone of voice do we have in marketing material?
• Who do I escalate a specific problem with?

Step 7. Define standard terminology

Now we need a dictionary of words, terms or phrases.

This can be an excel sheet or Confluence pages or a word doc, it doesn’t matter. It simply lists out common words and their use. For example what is the final button on a form say? is it submit or enter. What term is used to describe your customers, is it ‘customer’ or ‘client’? Do you use ‘first name’ and ‘last name’ form fields? Or ‘full name’ or ‘first name’ and ‘surname’?

It’s basic stuff but you need to standardise this, as every system or service you create must be consistent. Consistency is king or things really start to fracture and fall apart.

Step 8. Create a base system template

Now that you’ve standardised processes and wording we can start to build out a design template or showcase org (Salesforce). This can be a website template or a Salesforce app template or a ServiceNow system template. Yes this is the job of a developer not UX staff, but you need to lead the project.

Basically you build out the bones of an actual coded system such that whenever a new build or design request comes in you use the template as the base and update it accordingly with new media and copy. Do not create everything from scratch every single time.

This template can include standard form layouts, standard login processes, standard case management or inquiry handling. Anything that you may use frequently. Though keep this template clean, compact and basic. Do not get fancy.

Step 9. Build guides and manuals for all the items mentioned previously

If you’re building your foundational items quickly the thing that gets forgotten is: proper documentation and guidance.

So now that you’ve launched the majority of your templates and assets, now you need to stop, pause and think about how these items are best utilised, which means creating documentation and lots of it.

That doesn’t mean make it wordy, you actually need to make guidance that’s mostly bullet points, otherwise it just won’t be read. We recommend creating pages in Confluence or adding notes within Figma, attached to each UI screen. Or just get AI to do it for you, it’s pretty amazing at documentation.

Step 10. Define a deep and broad onboarding process

So we’re basically done, we’ve built assets and frameworks, built guides and manuals. Now we need to create a process for training new employees or contractors. Don’t just drop them into the work pool and expect them to swim.

Ensure you define what assets new staff must review and use, and ensure they talk to everyone they will be involved in, force lots of communication in the first weeks. A week is sufficient for experienced designers, but that week needs to be intense.

And you’re done!

You’ve now built all the foundations for a future focused, fast paced, high quality UX team, that can jump in and start running on a project from the first week .

Now what happens after you build the foundations? Well we’ll leave that for another medium post. But basically you move from building the foundations of one house, to creating the plans to build multiple houses… we’ll leave you to ponder that one.

I’m not sure if you’ve been on the internet lately… Oh wait of course you have, we all have, we can’t get away from it.

It’s not beautifully flawed, or uniquely different. It’s just a mess, if you go on Facebook (the community for human people) it’s just click-bait articles and AI videos with AI generated comments. Head to Twitter/X and you get deeply terrible arguments and deeply terrible responses, most of which are bots pushing certain perspectives.

Then we have Tiktok, an amazing place for tailored user created content that’s slowly becoming a world of silly AI generated videos, where only the best and most believable AI videos bubble up to the top. Now when I am on Tiktok every swipe requires a 2 second mental micro-investigation to decide if it’s real before I assign it consideration for my time. It’s a really odd feeling amd a deeply frustrating one.

Now let’s get to the fun/nightmare fuel part: it’s been found that 5% of content online in 2025 is AI generated, and that’s set to become 50% in 3 years…. 50% ladies and gentleman (and that’s conservative). And with the speed that AI video, audio and image quality is progressing, 3 years means that my 2 second mental micro-investigation check to determine if something is AI is useless, I just won’t be able to tell, and I have to deal with that.

So where do we go from here? well a cold lifeless internet is the inevitability. If every comment online is now perfectly articulated by an AI LLM and every counter argument perfection in prose also generated by an AI LLM, does it really mean anything any more? If perfect arguments become standard then there’s nothing left to be online for. No reason to argue, generate or propose ideas. It’s all gone.

So what do we do instead? Go outdoors, become a more ‘real world’ person?

Dont make me laugh, we can’t do that. We’re too addicted to the dopamine hits of the internet. Nothing else compares.

So what are we left with? Active communities that are actually lifeless with no humans involved, and no true interaction just bots arguing with bots as we humans are moved to the sidelines as we’re too slow to respond, too poor in our argumentative ability.

And what about us humans? Well we’re just going to lifelessly swipe on our screens absorbing whatever AI silliness the algorithms find effective. It’s not really very exciting is it?

Though there is some shining light. Though this glimmer is only accessible to the people intelligent enough to see through the dross. The people who see the dead internet for what it is and move through it. Where they use the internet simply for data extraction not entertainment, and don’t even engage in online arguments no matter how seductively they entice you. Why won’t they argue online? Well because why argue with bots? And even video or voice calls will be corrupted, where people will create visually lifelike AI’s of themselves to chat online for them. It may even get to the stage where real life experiences multiply in value as our online persona, so long cultivated, now becomes meaningless. Why even be famous online? When you’re only creating for bots, chatting to bots or arguing with bots.

But for the unenlightened, the brainrot specialists, this world is good enough. And unfortunately I think 90% of us are just that. We love the brainrot. Sigh…