01Investigative: Observe patterns, question assumptions.
Anticipate
What is emerging, and what does it mean for users?
Anticipatory design framework
Design AI that acts before users have to ask.
21 tools your design and product team fills in and shares: scan what is changing around your product, imagine how it could go wrong, and decide how much your AI does on its own.
Free: one framework project. No card.
Great software doesn’t sit around waiting to be asked.
Most AI waits for you to ask. A blank prompt box assumes people already know what to ask, and when. That puts all the work on the person it was meant to help.
The reactive trap
The user has to know what to ask.
- Blank prompt boxes
- Static forms and fixed defaults
- Chat that waits to be addressed
The anticipatory alternative
The product helps at the right moment.
- Reads the context
- Spots friction before it happens
- Steps back when it should
Every tool in the workspace
Each tool is a short working session: you fill it in and download the result. 21 are ready and 7 more are on the way. Use them in any order. Turn a card over to see what you get and which tools make it work better. Nothing is ever locked.
Anticipate
What is emerging, and what does it mean for users?
Show toolsHide tools
Anticipate
What is emerging, and what does it mean for users?
Exploration
What forces are shaping the future?
Set up your product
Describe the AI product you are designing, once.
Intent Map
Trace what the user wants, the workflow that serves it and the algorithm underneath.
STEEP Analysis
Scan the five outside forces shaping your product: social, technological, economic, environmental and political.
PESTEL Analysis
The same scan with a sixth force, Legal, for data protection, liability and discrimination.
Jobs-to-be-Done
Write the three jobs your user hires the product to do.
Horizon Scanning
Coming soonSort your signals by how far away they are: 0 to 2 years, 2 to 5, and 5 or more.
Expert Panels
Coming soonPressure-test your signals with domain experts to find the blind spots you can't see from inside the team.
Empathy Mapping
Coming soonMap what users say, think, do and feel, to understand their mental models before you design.
Automation vs. Augmentation Map
Coming soonDecide task by task where the AI acts for the user and where it backs their own judgement.
Synthesis
What patterns emerge?
Insight Clustering
Group your signals into themes, then pick the two you are least sure about.
Audience Factors
Map your user as a person in a situation, across five stages of readiness.
Vision
What futures are you designing for?
Desired Future State
Write the future you are designing toward, and the future the design must prevent.
Backcasting Milestones
Start from your 5-year future and work back to what must be true at 3 years, 18 months, 6 months and 30 days.
Delphi Method
Ask a group of experts to rate statements about your product in private, then see how the group answered.
Shared Principles & Impact Scan
Agree the principles you will not trade away, then anticipate what could go wrong and for whom.
Imagine
What futures are possible, and how do we design for them?
Show toolsHide tools
Imagine
What futures are possible, and how do we design for them?
Scenario Planning
What could go wrong in each world?
Name Four Futures
Combine your two uncertainties into four possible futures and name each one.
Design for Failure
Pick the future that most challenges your product and find three ways the system could fail there.
Error vs. Failure Messages
Write the actual message the user sees for each failure mode.
Alignment
What path connects the future to today?
Day-in-the-Life (3, 5, 10 years)
Write one day in your user's life some years from now, then the day if the design fails.
Workflows
How does the system learn, decide, and explain?
Feedback Loops
Coming soonDesign how the system learns from what people do.
Autonomy Protocols
Coming soonSet when the system acts on its own, when it asks first and when it steps back.
Reward & Explainability
Coming soonDecide what the system is rewarded for, and how it explains itself.
Shape
How do we make it real, and keep it relevant?
Show toolsHide tools
Shape
How do we make it real, and keep it relevant?
Behavioral Alignment
Who's in control, and when?
Autonomy Scale
Set how much the system does on its own, from full human control to full system control, for data and for AI output.
Feedback & Recovery Design
For each failure mode, decide what the user sees and how they get back in control.
Prototyping
How do we build trust?
Autonomy & Trust Calibration
Plot where your system sits and whether users will trust it too much or too little.
Explanation Types
Match how much the system explains to the stakes of the decision.
Evaluation
How do we keep it relevant?
Interactive Evaluation
Plan how you will know the system is still helping, over months and not just at launch.
Wind-Tunnel Test
Test your first Backcasting action in each of your four futures.
One place to keep every decision.
Stop arguing over AI behaviour in Slack threads. Write each decision down in one place, then send the result to your team.
Questions, not a blank page
Each tool asks for one part at a time, so nobody starts from an empty template.
Autonomy rules
Write down when the system acts, when it asks first, and when it stays silent.
Exports your team can use
Download a tool's result as text, or as CSV for the Backcasting map. Or email it to your team.
The framework behind the tools
Every tool belongs to one of three phases. Read a playbook for the full method, or open a tool in the workspace.
Forecasting predicts. Foresight prepares. Anticipation acts.
02Speculative: Envision alternatives, embrace ambiguity.
Imagine
What futures are possible, and how do we design for them?
03Iterative: Test, refine, and remain open to feedback.
Shape
How do we make it real, and keep it relevant?
- Behavioral Alignment
- Prototyping
- Evaluation
Not sure where to start?
Pick the situation closest to yours. Each one points to three tools and what you walk away with.
You are starting an AI product from scratch
Start with
- Set up your product
- Desired Future State
- Backcasting Milestones
You get
A written future to design toward, and a first action for the next 30 days.
You want your product to anticipate the next step
Start with
- Jobs-to-be-Done
- Audience Factors
- Autonomy Scale
You get
What users really hire it for, and what the product should do alone, ask about, or leave alone.
You are designing an assistant that offers help
Start with
- Autonomy Scale
- Autonomy & Trust Calibration
- Feedback & Recovery Design
You get
An autonomy level for each action, your main trust risk named before launch, and a recovery path for each failure.
Built on a published method
The Anticipate, Imagine and Shape framework is drawn directly from The Anticipatory Design Playbook (CRC Press), not a trend deck.
Rooted in doctoral research on anticipatory systems. Applied in complex enterprise AI.
Before you start
Is the framework only for complex AI, or does it work for traditional software?
It's sharpest on probabilistic, adaptive systems: anywhere the product acts on its own initiative. But the method (model intent, design behaviour, shape the UI) applies to any product that wants to anticipate rather than react. Traditional software gets most of its value from the Anticipate phase; AI-native products use all three.
How do the deep-dive playbooks differ from the guided workspace?
The playbooks (/framework/anticipate, /imagine, and /shape) are the open methodology: the theory, activities, and templates, free to read. The guided workspace is where your team applies them to a real project: the same phases as interactive tools, with your inputs saved, structured, and exportable.
How does this prevent AI from becoming intrusive, the “Clippy effect”?
That's the whole job of the Imagine and Shape phases. You calibrate confidence thresholds so the system only acts when it's sure enough, design fallback paths for when it isn't, and set an autonomy level matched to the stakes, with override controls that keep the user in charge. Proactivity without agency is what created Clippy; the framework designs the agency back in.
Can we export the results to Figma, Notion or Jira?
Not directly. Each tool downloads as a text file, and the Backcasting map also downloads as CSV. You can email any export to your team, or paste what you need into Jira or Notion. There is no direct Figma, Notion or Jira integration yet.
More questions? See the full FAQ →
Ready to move beyond the prompt box?
Pick a tool, fill it in, and send the result to your team. Free to start, no card required.
Free: one framework project. Pro: unlimited projects and the AI coach. See plans →