The Intent Shaping stack · 46 cards
Intent Shaping: meeting users inside their own decision
Intent shaping is the craft of noticing what a user is trying to do and quietly making that easier.
Not pushing them somewhere they were not going. Reading the signals, hesitation, repetition, backtracking, and responding with the right nudge at the right moment.
Done well it feels like the app understands you. Done badly it feels like being sold to, which is why restraint is half of this stack.
Why do users wander, stall, or leave without deciding?
Because wanting something and knowing how to get it are different states.
A user comparing the same two options for the third time has intent. A user re-running a search with new words has intent. A user hovering at a price has intent.
Most apps watch all of this happen and respond with nothing.
The opportunity is not more persuasion. It is answering the question the behaviour is already asking.
Common intent shaping problems, fixed
The same failures come up again and again when founders describe broken intent shaping. Each one below has its own guide: how to spot it in your funnel, why it happens, and the fix.
How do you guide users without being pushy?
Serve the visible goal, never a hidden one.
Someone stuck between two options can be offered a comparison. Someone re-filtering can be offered the refinement they seem to want. The response earns trust because it manifestly helps.
The same mechanics pointed at the app's goal instead, upsells disguised as help, urgency without cause, teach users to ignore every future nudge. The trust cost outlasts the conversion.
How does personalisation shape intent?
By narrowing an overwhelming space to a navigable one.
A feed tuned to demonstrated taste, defaults set from real behaviour, a next step suggested from where someone actually is: personalisation at its best removes the paralysis of infinite options.
It needs honesty about its inputs. Recommendations built from three signals should feel like suggestions, not verdicts. Overclaiming certainty is how personalisation turns creepy.
When is the right moment to make an offer or suggestion?
When behaviour signals the decision is already live.
Hesitation at a threshold, a repeated return to the same item, a completed milestone that naturally opens a next step: these are open windows.
A suggestion inside the window feels like service. The same suggestion sprayed at everyone on a timer feels like noise, and the difference is measurable in both conversion and uninstalls.
The building blocks inside the Intent Shaping stack
- Intent Mirroring — Behaviour is a question: answer it in the moment.
- Discovery — Users find what is possible while doing what they came for.
- Personalisation — The overwhelming becomes navigable, honestly.
- Spark Curiosity — A partial reveal invites the next step.
- JTBD Copywriting — Every suggestion speaks the user's goal, not yours.
There is no universal answer. Make it yours
None of the above is a rulebook, and the playbook does not pretend to be one.
Every building block ships with Make It Yours prompts: guided questions that shape the tactic around your product, your users, and your stage.
Work through them yourself, or hand them to your AI agent as context. The thinking should happen about your app, not the average app.
- Where in your app do users visibly hesitate, and what information is missing at that spot?
- Which repeated behaviours could honestly reshape what a user sees next?
- Take your most-shown suggestion: would users thank you if they saw exactly why it appeared?
How the best apps do it
Pinterest: Every save reshapes the feed. Intent is read from action, and the product visibly gets more yours.
Amazon Music: The considered pause before a recommendation makes the suggestion feel chosen for you rather than pulled from a shelf.
Airbnb: Filters, maps, and saved lists let intent sharpen itself. The app organises the decision without making it for you.
Usually works
- Respond to hesitation with the missing information
- Let visible behaviour tune what appears next
- Frame suggestions in the user's goal language
- Leave every nudge easy to ignore
Usually backfires
- Dress upsells as guidance
- Fire suggestions on timers instead of signals
- Overclaim certainty from thin data
- Make the nudge louder when it gets ignored
Questions founders ask
What is intent mirroring?
Reading a behavioural signal, hesitation, repetition, backtracking, and responding to the need it implies.
It is the difference between an app that plays a script and one that appears to listen. The card covers the signals worth watching and the responses that respect them.
Is shaping intent manipulative?
It depends which intent wins.
Helping a user finish a decision they started is service. Engineering a decision they never had is manipulation. The playbook's test: would the user thank you if they could see exactly how the nudge worked?
What data do you need for this to work?
Less than you might think.
Most of the signals are in-session behaviour: taps, repeats, dwell, backtracks. You can respond to those without any profile at all, which is also the privacy-friendly place to start.
Ask a question
Thinking…