Paul Boag

Last call for the Agency Academy October discount. It ends tomorrow, Saturday, October 31. Anyone who joins before then pays £59 a month instead of £79, or £49 a month billed annually (plus VAT). If you've been sitting on the fence, there's no lock-in. You can cancel from your account at any time and you won't be charged again, so the risk is a single month. The next call is Thursday, November 5, at 3:30pm UK time, and if you can't make it live you can send your question in advance and watch the recording later. Freelancers are as welcome as agency owners, and so are people still deciding whether to go freelance at all. The problems are the same, only at a different scale. agencyacademy.boagworld.com/…

A lot of organizations have a shadow AI problem at the moment, whether they know about it or not. When the in-house AI tool won't let people do their jobs, plenty of them don't wait for the steering group. They open a personal AI account on their phone or home laptop and get on with it. We saw exactly the same thing with shadow IT during the digital transformation years, when people signed up for their own file sharing and document tools because the official ones couldn't do what they needed. The trouble is that this is far more dangerous than the problem the restrictions were meant to prevent. Company data ends up in tools the organization has no control over, with no oversight and no record of what's been shared, so in trying to lock everything down you end up pushing the risk somewhere you can't see it. The instinctive reaction is to clamp down with stricter policies, blocked websites and warnings about disciplinary action, but that rarely works. Either it drives the behavior further underground, or people just give up, and we never see the AI adoption that senior management say they want. I'd treat shadow AI as a useful symptom instead, because it shows you exactly where the official tools are falling short. Talk to staff about what they're using personal accounts for and what they feel they can't do with the AI they've been given, then use that to shape what gets rolled out next. Yes, that means living with a bit of chaos in the early days. But you can always lock things down later, once the real needs are clearer, and the risk of moving too slowly is greater than the risk of moving too fast. More in my new piece for Smashing Magazine: smashingmagazine.com/2026/10… #ShadowAI #AIGovernance #AIAdoption

Quick reminder. The early-bird price for Better User Experiences with AI ends today. It's £195 + VAT until the end of today, then £250 + VAT. Four live sessions on 10, 11, 17 and 18 November, 4:00 to 5:30pm UK (11am ET). All recorded, so you won't miss anything if a meeting gets in the way. → better-ux-with-ai.boagworld.… #AI #UX #Workshop

The early-bird price for Better User Experiences with AI ends this Friday, 30 October. £195 + VAT until then, £250 + VAT after. I'm not going to pretend there's a dramatic countdown clock involved. It's just a date, and I'd rather you knew about it before it passed than after. Four live sessions across 10, 11, 17 and 18 November, 4:00–5:30pm UK. Briefing AI so it gives you something useful. Turning customer insight into decisions. Designing AI experiences people actually trust. Helping your organization get on board with AI rather than having it imposed on them. All recorded, so a clashing meeting costs you nothing. If booking keeps sliding down your to-do list — a feeling I know far too well — this is a good week to get it done. better-ux-with-ai.boagworld.…

Some of the stories I hear about AI rollouts at the moment make me want to go and lie down in a dark room. Staff get told to embrace AI, and then a few weeks later somebody gets reprimanded for using too many tokens. Or people get lectured so much about what they should and shouldn't do with AI that the safest option is to carry on doing exactly what they've always done, which, to be fair, is a perfectly rational response. I saw the same pattern during digital transformation. Senior management would demand change while still expecting the same KPIs to be met and still pointing out every mistake in annual reviews, and then they were confused when nothing happened. There was no space to fail, because the pressure to deliver never let up. Learning a new tool like AI involves making mistakes by its very nature. So if management want staff to experiment, they need to make it feel safe, and that means adjusting KPIs, changing how people are reviewed, and relaxing some of the governance while people learn. I wrote about this, along with a few other lessons I've dragged back out of the digital transformation years, in my new Smashing Magazine article. smashingmagazine.com/2026/10… #AIAdoption #ChangeManagement #Leadership

You can find out whether your landing page works before you build it, in about an hour, for nothing. Show people the layout for 8 seconds, take it away, then ask what the page offered and what they remember being on it. That is a snap test, and it is brutal in the most useful way, because visitors often abandon a page within the first 10 to 20 seconds anyway. If your message has not landed by then, you rarely get a second chance. Follow it with a first click test, where you ask people where they would click to take action and then look at the heatmap. If the clicks are scattered, your call to action is not as obvious as the four people who designed it believe. Predictive eye tracking tools are worth a look too, but be clear about what they tell you, which is where attention is likely to land on a typical page, not whether anyone finds your offer compelling. How to run these tests is lesson 12 of my free email course. Sign up for the whole thing here. → landing-page-course.boagworl… #UX #UserResearch #ConversionOptimization

Most SaaS homepages don’t fail because they’re bad. They fail because they’re almost good. My latest First Impressions review is up. This time I look at Aaardvark, an accessibility tool built by Natalie. For anyone new to the series, people submit their websites and I review them on camera, seeing them for the first time. Part usability test, part expert review. Aaardvark gets the basics right. You know what it does within seconds and the page follows a sensible structure. But a few things are holding it back: 🔘 Two calls to action that look identical. Both say “free.” One creates an account, the other runs a scan. Pick a primary and make the secondary look secondary. 📍 The best content is on the wrong page. Social proof and objection handling are sitting on the pricing page. They belong next to the sign-up buttons, because that’s where people hesitate. 🖼️ Clip art instead of screenshots. When you hide your interface, people start wondering why. 🎯 No clear audience. It turns out to be aimed at agencies and freelancers, but you only find that out halfway down the page. None of these are big fixes. Together they could make a real difference to sign-ups. Thanks to Natalie for letting me prod her homepage in public. Want me to review your site or app? Links below 👇 Reply: 🎥 Watch the review: youtube.com/watch?v=to5E2HnK… 🙋 Submit your site or app for free, or book a paid review if you want it guaranteed or kept private: review.boagworld.com

I've got a new article up on Smashing Magazine, and it got a little ranty in places. The way most organizations are rolling out AI looks more like a farce than a strategy to me. If you put it in a film, reviewers would call it unbelievable, because surely nobody would buy AI licenses for the whole company, run one round of generic training and consider the job done. And yet here we are. What makes it more frustrating is that we've been here before. Back in 2014 I wrote a book for Smashing called Digital Adaptation, about the mistakes organizations were making with digital transformation, so I went back through it to see how those lessons apply to AI. The similarities are striking, and slightly depressing. The biggest mistake, then and now, is treating it as a technology challenge. The technology is the tip of a very large iceberg, and underneath it sits a cultural shift with enormous implications for staff, who this time around also fear for their jobs. So the article covers why staff would want to adopt AI in the first place, the time and support they need, making it safe to fail, handling governance with care, and rolling out the tech in a way that doesn't stifle everything. smashingmagazine.com/2026/10… #AIAdoption #ChangeManagement #DigitalTransformation

Hey all. 👋 Just a reminder that I am running a deal on one-to-one coaching until the end of this month. 4 hours of coaching for only £660+VAT rather than its normal price of £860+VAT. Message me if you are interested. boagworld.com/boagworks/work…

Most business advice for agencies assumes you want to scale fast and sell up. If that's you, Agency Academy probably isn't the place for you, and that's fine. It's for freelancers and small agency owners who want steady work and a life outside the laptop. We talk about protecting your evenings and taking a proper holiday without the business falling over, and about what AI means for how we work, minus the hype. Most of the best advice comes from members who've recently solved the exact problem you're stuck on, whether that's a client who rewrites the brief halfway through or a pipeline that's suddenly gone quiet. They share it on the Thursday call or in Slack, where a question asked at 11pm usually has an answer by morning. Until the end of October, anyone who joins pays £59 a month instead of £79, or £49 a month billed annually (plus VAT). There's no lock-in. agencyacademy.boagworld.com/…

Most change programs fail for reasons that have nothing to do with the quality of the idea. They fail because somebody got surprised in a meeting. If you are trying to shift how your organization works, it helps to stop thinking about stakeholders as one lump and start thinking about three groups. Supporters, who need something useful to do before their enthusiasm quietly evaporates. Skeptics, who mostly want their questions taken seriously rather than deflected. And blockers, who usually have a real concern about status, workload or risk that nobody has bothered to ask them about. Get the objections out early and in private. An objection raised over coffee is a conversation. The same objection raised in a group meeting becomes a position that person now has to defend for the rest of the project, largely out of pride. Then pitch to the individual, not the room. What convinces the finance director is not what convinces the head of engineering, and trying to write one deck that does both produces something that convinces nobody. The rest is culture hacking. Small, sustained, faintly boring changes nobody objects to, repeated until the new way is simply how things are done and everyone has forgotten it was your idea. There is an email on the political side of all this in my free Modern UX Practitioner course, and 13 others on the work itself. uxcourse.boagworld.com/ #DesignLeadership #ChangeManagement #UX

I've put my thoughts on AI adoption into a YouTube video, mostly because I couldn't stop ranting about it to anyone who'd listen. I keep watching organizations roll out AI and make exactly the same mistakes I wrote about in Digital Adaptation back in 2014. They appoint a Chief AI Officer, set up a central committee, and lock the tools down so tightly nobody can do anything useful with them. Meanwhile their staff are wondering whether all this enthusiasm is about replacing them, which, given the headlines, is a fair question. In the video I talk about why your staff are right to be nervous and what you can actually do to bring them with you. youtu.be/GzCLyMRKRLY

Most user research gets presented once and then forgotten in a shared drive. Somebody spent weeks talking to real people. There was a nice deck. Then months later a decision gets made in a meeting and nobody in the room remembers a word of it. Session two of my Better User Experiences with AI workshop is about changing that: - Building a research repository that anyone in your organization can ask questions of. - Creating personas that focus on what users need to do, drawn from real research. - Giving those personas a voice, so colleagues can test ideas against them in the meeting where the decision is being made. Four live sessions from 10 November, all recorded. Early-bird pricing ends this Friday, 30 October. → better-ux-with-ai.boagworld.… #UserResearch #Personas #AI

Every time a website or app underperforms, someone suggests a redesign. New layout. New buttons. A fresh coat of paint. But in my experience, the interface is rarely the whole problem. The real blockers usually sit behind it. Things like: ❌ Five departments fighting over space on the homepage ❌ Legal adding three paragraphs of small print to the checkout ❌ Simple fixes stuck in a development backlog for 18 months ❌ A CMS that can’t do what the design needs ❌ No budget for research, so the loudest opinion wins ❌ Success measured by launch dates rather than whether anything improved None of these are design problems. But every one of them shows up in the interface. So I’m curious. What’s the real problem getting in the way of a better experience where you work? Share your stories in the comments. The more specific the better. I suspect a lot of us are fighting the same battles, and it helps to know you’re not the only one. I'm hoping to use some of them as examples to create helper guides and videos. (Feel free to change the names to protect the guilty.)

If you freelance or run a small agency, a community full of people who do what you do can sound like a community full of competitors. In Agency Academy it rarely works out that way. Someone posts in Slack that they need a developer for a project, or asks whether anyone does SEO, and the work goes to a person they already know and trust from the weekly calls. Members end up as collaborators and referral partners far more often than rivals, and they're the people you call when a project is too big to take on alone. That all sits alongside a group call every Thursday at 3:30pm UK time, which is recorded and summarized if you can't make it live. If you've been meaning to join, there's a discount until the end of October. Anyone who signs up before then pays £59 a month instead of £79, or £49 a month billed annually (plus VAT). There's no lock-in and you can cancel any time. agencyacademy.boagworld.com/…

I am bored with my proper work and so am hanging out in Google Meet, chatting with anybody who fancies joining me. Ask me questions, moan about AI, ask me to review your site. Whatever you fancy, come and join... meet.google.com/sdo-eksv-yre

A founder showed me the dashboard he'd built himself last month. Pulling in data from every corner of the business, across several countries, and tracking the numbers his board actually cared about, done in a few days after years of watching an expensive reporting tool sit there unused. He was rightly chuffed with himself, and I made all the right impressed noises, quietly filing it next to my own recent triumph of finally working out what half the buttons on my phone actually do. Then I asked how the rest of the business felt about it, and he went quiet, because he'd had his own personal lightbulb moment with AI and solved his own biggest headache, and somewhere along the way had assumed that meant the company was on the same journey. It wasn't. His finance team hadn't had that moment. Neither had the people running operations on the other side of the world, or the staff who actually talk to customers, because nobody had gone round and found out what their version of his dashboard problem even was. That's the gap I keep running into wherever a leader gets an early win with AI. The eureka happens to one person doing one job, and the assumption is that it just spreads outward on its own. It doesn't work that way. Somebody still has to go department to department, find what's actually eating people's time, and show them what disappears once AI takes a swing at it. Skip that step and you end up with a genuinely brilliant dashboard that nobody outside the boardroom has ever opened. If you've had your own AI win recently, it's worth asking who else in the building is still waiting for theirs. #AI #Leadership #DigitalTransformation

Every AI rollout I've watched stumble, starts in the same room. Someone in charge gathers the sceptics together and tries, patiently and reasonably, to talk them into caring. It hardly ever works, and I've started to think that's because it targets exactly the wrong people first. I heard about one attempt where a consultant sat the most experienced, most doubtful staff down and walked them through the basics of AI as though it were a fire safety briefing. It landed about as well as you'd expect. Everyone nodded, ticked the box, and went straight back to doing things exactly as before. Garbage in, garbage out, and no amount of enthusiasm from the person at the front changes that. What actually shifts a culture is going after the keen ones first, the two or three people already quietly using AI to make their own week easier without telling anyone. Put them in a room together, let them compare notes and show off a bit, and that little group starts generating its own gravity. Once enough of the willing are visibly getting somewhere, the holdouts stop feeling brave for resisting and start feeling like the last ones to the party. Nobody argued them into it. They just didn't fancy being left behind. Going after the reluctant first feels like the responsible, thorough thing to do. It's usually the slowest possible way to get anywhere. #AI #ChangeManagement #WorkplaceCulture

I haven't been to a conference in years, and I think it has cost me. What I lost were the hallway conversations. The ones where you mention a problem over coffee and someone says "we had exactly that last year, here's what worked." If you work in-house, especially as a team of one, those conversations are hard to come by. There's nobody down the corridor who has dealt with this before, and your manager has their own fires to put out. With AI changing the job, we need them more than ever. That's the gap UX Insiders is meant to fill. A weekly call and a private Slack with other in-house UX people, working through the stuff that sits around the design work. A founder place costs £588 a year + VAT. That's less than a ticket to most UX conferences, and instead of two days of talks you get help every week of the year. → uxinsiders.boagworld.com/?ut… #UX #UXCommunity #UXLeadership

Every AI rollout I've watched stumble, starts in the same room. Someone in charge gathers the sceptics together and tries, patiently and reasonably, to talk them into caring. It hardly ever works, and I've started to think that's because it targets exactly the wrong people first. I heard about one attempt where a consultant sat the most experienced, most doubtful staff down and walked them through the basics of AI as though it were a fire safety briefing. It landed about as well as you'd expect. Everyone nodded, ticked the box, and went straight back to doing things exactly as before. Garbage in, garbage out, and no amount of enthusiasm from the person at the front changes that. What actually shifts a culture is going after the keen ones first, the two or three people already quietly using AI to make their own week easier without telling anyone. Put them in a room together, let them compare notes and show off a bit, and that little group starts generating its own gravity. Once enough of the willing are visibly getting somewhere, the holdouts stop feeling brave for resisting and start feeling like the last ones to the party. Nobody argued them into it. They just didn't fancy being left behind. Going after the reluctant first feels like the responsible, thorough thing to do. It's usually the slowest possible way to get anywhere. #AI #ChangeManagement #WorkplaceCulture