Real Coaching Adapts. Why Didnt We?
When we launched Pictor three months ago, we were proud of one thing above everything else.
Pictor is an AI Agile coach, built for anyone working in or around Agile: the Scrum Master preparing for a difficult retro, the Product Owner navigating stakeholder conflict, the Engineering Manager at the intersection of technology and people, the Agile coach facing an organisational resistance problem they can't bring to a supervisor, and the practitioner who's been doing this long enough to want depth, not basics. Not a general-purpose AI you can ask Agile questions. A coaching tool that understands the territory from the inside and adapts to wherever you are in it.
And within that, we hadn't just built another AI that answers questions. We'd built three distinct coaching approaches Expert, Mentor, Coach, and given practitioners the ability to choose. Pick the mode that fits your need. Get the kind of response you actually want.
We thought that was the insight.
It wasn't.
The mistake we didn't see until we watched
We didn't launch Pictor with a go-to-market strategy. We shared it with a handful of practitioners we knew, people who would actually use it and tell us the truth. No campaign, no press, no metrics dashboard. Just: here's what we built, tell us what happens.
Three months later, we had something more valuable than a user base. We had patterns.
The Switchers. Users who started in one mode, realised mid-conversation they needed something different, switched, then switched again. The mode selector wasn't helping them. It was interrupting them.
The Stuck. New users staring at three buttons before they'd even asked a question. Expert, Mentor, Coach. Which one do I need? I don't even know what my problem is yet. Analysis paralysis before the coaching had started.
The Invisible Feature. Users who picked a mode on day one and never changed it. Not because they only ever needed one approach, but because they forgot the feature existed, or never realised they could switch.
We'd been watching for friction. What we were actually watching was something more fundamental: a mismatch between what we'd built and what coaching actually is.
What coaching actually is
Think about the best coaching conversation you've ever had.
Did your coach open by asking: "Would you like direct answers, stories, or questions today?"
Of course not.
Real coaches read the room. They notice the question beneath the question. They start somewhere and shift when they sense something deeper. The adaptation happens invisibly and that invisibility is exactly what makes it feel like coaching rather than consultation.
We'd built something that worked like this instead:
"Before I can coach you, tell me what kind of coaching you want."
That's not coaching. That's asking you to diagnose yourself before the conversation begins. We'd put the cognitive burden in exactly the wrong place.
What we changed
Pictor 2.0 removes the mode selector entirely.
You don't choose Expert, Mentor, or Coach. You just ask your question.
Pictor reads your intent and adapts. Stuck and need immediate help? Direct, actionable guidance. Exploring a complex pattern? Perspective from documented cases. Wrestling with a decision or a team dynamic? Questions that help you think it through.
And, critically, Pictor can shift mid-conversation. Start with a practical question, and when the real issue surfaces underneath it, Pictor moves with you. No buttons. No manual switching. Just a conversation that meets you where you are.
The three coaching approaches are still there. They're the same methods. But now you don't manage them. Pictor does.
Why ChatGPT, Claude, or Gemini won't do this
You can ask ChatGPT about Agile. You can paste your retrospective notes into Claude and get a decent analysis. Gemini will give you a competent summary of the Scrum Guide.
None of them will coach you.
The difference isn't knowledge. Generic AI tools know plenty about Agile frameworks. The difference is that they're built to answer the question in front of them, accurately, helpfully, and completely without regard for who's asking or why.
They don't know if you're a first-time Scrum Master or a seasoned coach. They don't know if you've already tried the obvious approaches. They don't track whether you keep coming back to the same problem. They don't ask what's underneath.
Pictor is built around a different question entirely: not what does this person need to know, but what does this person need right now — and sometimes the answer is not an answer at all. Sometimes it's a question. Sometimes it's a challenge. Sometimes it's the observation that you've been circling this issue for three weeks and the technique you're asking about won't fix it.
Generic AI is optimised for the response. Pictor is optimised for the person.
That's not a marketing distinction. It's an architectural one. The prompts, the competency tracking, the adaptation between coaching approaches, all of it is built around the practitioner's development over time, not the individual query in the moment.
What this looks like in practice
You're a new Scrum Master, three weeks in. A senior developer dominates every stand-up — talks for five minutes, derails the flow, demoralises the team. You ask Pictor: "How do I handle this?"
A generic AI gives you a list. Five techniques for managing dominant personalities. Ground rules. Talking tokens. Self-reflection prompts. Correct. Impersonal. The same answer it would give anyone.
Pictor 2.0 does something different.
It picks up that you're asking as someone new from the context you've shared, from the way you've framed it. It notices the tension in "handle this." And it responds:
"Before we go to techniques — what's your relationship with this person like? Have you spoken to them one-on-one, or has this only come up in stand-ups?"
Then it adapts to your answer.
If you haven't tried anything yet: direct guidance on where to start.
If you've tried and it hasn't worked: patterns from others who've been here.
If you're circling the real conversation: questions that help you see what's in the way.
The difference isn't what Pictor knows. It's when it gives you the answer — and when it doesn't.
One detail worth naming
If Pictor shifts its approach mid-conversation, you'll see a quiet signal showing which approach it's using and why. Not an announcement that breaks the flow. Just a gentle indicator that builds your coaching literacy over time.
The goal is that you don't just get better at your work. You get better at understanding how coaching works — when direct guidance serves you, when perspective from another context unlocks something, when you need to sit with a question before you're ready for an answer.
The best coaching is invisible until it needs to be visible. This tries to walk that line.
What hasn't changed
Everything underneath is the same.
Pictor still tracks your competency development across coaching, facilitation, teaching, mentoring, and transformation. Every conversation still connects to your growth. Actions you commit to are still tracked. Pictor still remembers who you are and what you're working on. Your data is still yours — never sold, never used for training.
The promise hasn't changed either: Pictor isn't here to answer your questions. It's here to develop you.
The mode selector was the only thing in the way of that.
A note on how we got here
Pictor didn't launch. It started quietly, shared with a small group of practitioners who agreed to use it honestly and tell us what they found. Some used it every day. Some got stuck before asking their first question. Some never discovered that mode switching was possible.
We watched all of it. We talked to them. And we kept coming back to the same uncomfortable realisation: we'd built a feature we were proud of that was asking users to do work that the product should be doing.
Pictor 2.0 is what happens when you take that seriously.
If you're already using Pictor, the upgrade is automatic. Your next conversation will be adaptive, no learning curve, no disruption.
If you're new: just ask your question. Pictor will meet you where you are.
Questions or thoughts? We're still learning, and we're still listening. Reach us at info@yourpaths.eu