Session content as input
Generated from the conversation, never a library.
Anchoring coaching insights into daily life.
Insights from a coaching session rarely survive the return to everyday life. anqr is a coach-supported, AI-assisted workflow that turns them into small activities anchored to the real situations clients are about to face.
The intention behaviour gap.
Behavioural science has a name for it. Intentions formed inside a calm conversation collide with the pressures of daily life. Context, not willpower, is what triggers action.
A persistent weakness runs through coaching practice. The insight that emerges in a session does not reliably carry over into everyday behaviour.
Survey · n = 30
What fades first is emotional, not cognitive.
When clients re-enter daily life, the plan is not what disappears first. The felt sense of breakthrough is.
Built from the session, for the situation.
anqr abbreviates anchor. An activity works when it is drawn from the actual session, tied to a real situation the client is about to meet and delivered inside the coaching relationship rather than around it.
Existing tools offer generic libraries or AI with no coach. anqr keeps both the session and the coach in the loop.
Seven principles.
Context-specific anchoring
Tied to a real upcoming situation, not an abstract goal.
Practitioner curation
The coach shapes and bounds what gets generated.
Client-centred language
The client's own words, the top condition for safety.
Support for agency
Adapt, reinterpret or skip. Never prescribed.
Low-friction integration
Light input, immediate output, zero admin.
Behavioural relevance over engagement
Success is follow-through and eventually not needing the tool at all.
One loop. Two people.
The coach configures once. Then the client carries the loop and every cycle feeds the next session. Six steps, repeating.
- 1
The session
Coach and client meet. An insight surfaces.
- 2
The coach configures
Tone, scope and boundaries, set once for this client.
- 3
The client captures it
The takeaway in their own words and an upcoming situation it touches.
- 4
anqr translates
The summary and situation become three small, doable activities.
- 5
The client tries one
They act in the real moment and can note what worked or not.
- 6
It returns to the room
The next session picks it up, configuration is refined and the loop begins again.
- Together
- Coach
- Client
- anqr
Each session lowers the anchor a little further.
Every cycle adds a point to the rope and the anchor descends. Insights become grounded in daily life rather than stacked as a flat list.
3 sessions
New
Session
Session
Tap to lower the anchor
How anqr took shape.
One thesis, five chapters. Each one shaped the next.
-
Introduction
A gap worth designing for
Coaching grew because it sits between self-help and clinical care, with less stigma and fewer barriers. But its weakness is transfer. An insight formed in a calm room rarely resurfaces under daily pressure and existing tools are either generic or leave the coach out entirely.
So the project asked the field why follow-through fails.
-
Research field
What turns an insight into an action
Nine practitioners placed the mechanism of change in the relationship, not the intervention. A survey of thirty added that the felt sense fades first and that the client's own language is what makes generated content feel safe. The learning theories explained how to bridge the gap.
- Implementation intentionsAn if-then plan tied to a cue makes an intention far more likely to happen (Gollwitzer).
- Habit formationBehaviour repeated in a stable context becomes automatic, triggered by the cue itself (Neal).
- Tiny habitsNew behaviour sticks when it is small, anchored to a routine and reinforced at once (Fogg).
- Experiential learningLearning only completes through active experimentation, which cannot happen inside the session (Kolb).
- Self-determinationBehaviour persists when it is self-directed rather than imposed (Ryan & Deci).
-
Concept
What nothing yet solves
Measured against contextual relevance, session-specificity, client autonomy and a practitioner in the loop, no platform meets all four. Practitioner tools like Ylah and Quenza keep the coach but assign from libraries. AI coaches like 1440 and ChatGPT feel personal but drop the practitioner. Behaviour apps like Feeling Great do neither.
Before building the full product, an early experiment tested the AI directly: could it generate activities from a session summary? Without careful instructions it tended to give generic advice and invent details the client never mentioned. That finding shaped the whole build — the coach's upfront configuration acts as a set of guardrails, and the AI only works inside them.
The answer was a three-step minimum product.
-
Project development
Three iterations, one contradiction
Testing reshaped the system. The combined flow was too long, so coach and client split apart. Coaches who demanded oversight in the survey preferred configuring once in practice and all four clients rated situation-anchored activities above generic ones. The interface warmed from a form into an invitation.
Iteration 2 · a green form 
Iteration 4 · an invitation Voice input came last and three tabs became one journey screen.
-
Reflection
Proven and honestly open
Proven within scope: anchored activities feel immediately relevant and the client's own words make them feel personal. Still open: whether upfront configuration sustains coach trust, whether hallucination stays rare without review and whether any of this closes the gap over time. Those need cycles a thesis does not have.
Which is what the live prototype below is for.
Three findings that mattered.
Success for anqr means eventual disengagement. Clients learn to anchor insights on their own and the tool becomes secondary.
Bounded AI
It translates. It never reads emotional states or decides what a client needs.
Clear boundaries
Non-clinical coaching only. Trauma and crisis stay out of scope.
Light on compute
A few inference calls per cycle, not an always-on companion.
One full cycle, in motion.
anqr
Turn your sessions
into activities
Client opens the live visitor build. Coach opens the configuration screen.
anqr
Bachelor Project Interaction Design 2026.