Your own website can clearly say who you are. Independent sources play a different role: they can corroborate what you claim from outside your own domain.

Someone who only gets a brief sees a summary of reality. Someone fully absorbed into the internal workings no longer notices some of the assumptions. Good external work takes enough context to understand what's going on, and enough independence to keep questioning it.

AI systems increasingly carry out tasks themselves: updating a status, getting a message ready, preparing a change for production. Whether the system can do that is usually answered quickly. Whether it may decide about it is a different question, and one that is rarely asked on its own.

Good communication can expose what an organisation itself doesn’t yet have a clear answer to. Sometimes the text needs to be better. Sometimes the way things work needs to be better first.

AI is starting to remember, update context and act on what it thinks it knows. That makes the status of information part of the workflow.

Multiple brands, audiences and channels don't automatically add up to stronger communication. The real question is who's speaking, why, where — and who makes sure all those choices still add up to one line.

A simple language switcher turned into a small real-world test of agentic working: ChatGPT helped set the rules of the game, Claude built, Kimi checked in an actual browser — and the decision stayed exactly where it belonged.

A withheld payment, an assertive email, the urge to reply today. Underneath many commercial disputes lies something else first: scattered information, claims that start to sound like facts through repetition, and photos that prove less than they seem to. Whoever first sorts out what is established, what is claimed and what is still missing keeps more room to respond deliberately afterwards.

Being findable used to mean: knowing which keywords you rank for and who outranks you. That's still true, but it's become incomplete. Google is building AI deeper into Search, and customers are increasingly asking their question directly to ChatGPT and other systems too. The relevant question then becomes: who gets named, for which commercial question, next to whom — and based on which sources?

A concrete customer request quickly feels like proof that there's an opportunity. Sometimes it is. Often the requested feature turns out to be a symptom of a different problem, the solution already exists, or the concept works technically but not operationally or economically. For me, product discovery therefore doesn't start with building — it starts with trying to knock an idea down.

A positioning exercise or brandbook can set a direction. But a brand only becomes truly visible in the hundreds of small decisions that follow: which photo you use, which line you cut, who says something, on which channel, and when you're better off not publishing at all. For me, brand consistency is therefore not a one-off exercise. It's daily editorial work — and, ultimately, execution.

A recurring administrative task looks like an obvious candidate for automation. Until it turns out that the real workings consist of exceptions, historical agreements, numbering, and knowledge that was never formally recorded anywhere. For me, automating therefore doesn't start with software, but with the question of which decisions are actually being made today — and which of those can never be allowed to go wrong.

What two fibre lines, Tailscale and Screens 5 changed about the device in front of me

Many assignments only begin once someone has already named what seems to be needed. Simply Dom often becomes useful earlier: when there is a signal, but the real business question is not yet clear enough.

A benchmark can show that you differ from others. That makes it useful as a signal, but not yet an answer to how much you should spend yourself.

AI makes production cheaper, but not the consequences of a wrong choice. The real value shifts from generating to observing, checking, selecting and cutting.

Producing content means looking closely. Whoever does that thoroughly runs into things that need checking or discussing before they're published.