AI Design Sprint

The sprint has not changed. The ceiling has.

Bigger bets, stronger prototypes and richer evidence for sharper decisions.

The AI Design Sprint — Mosaic Innovation overview cover

The original sprint was a shortcut to learning. AI makes the learning richer.

A Design Sprint still answers a critical business question through focused making, testing and learning.
AI raises what the sprint can produce.

Sharper framing

Broader exploration

Stronger prototypes

Better user evidence

Sharper decisions

Clearer route to action

The goal is not a shorter sprint. It is a more ambitious decision with stronger evidence behind it.

Parallel by design

Prototyping used to force a single route. AI makes it practical to carry more than one far enough to test properly.

Classic sprintAI Design Sprint

Strategic routes

One route

Multiple strategic routes

Prototypes

One prototype

Multiple prototypes

Testing

Live user testing

Synthetic pre-screen + live user testing

Synthesis

Manual synthesis

AI-accelerated synthesis

More routes where they matter.

More fidelity where it counts.

Sharper evidence for what to back.

How the sprint works

Human judgement in the room. Real users at the point of truth.

MobiliseCore sprintActivate

Pre-sprint

Mobilise

Build the evidence base and decision frame.

Day 1

Understand

Lock onto the highest-value opportunity.

Day 2

Sketch

Human first. AI stretches the strongest routes.

Day 3

Decide

Stress-test the field and choose what to build.

Days 4–5

Prototype

Build decision-grade routes in parallel.

Day 6

Test

Real users decide. AI accelerates the read.

Post-sprint

Activate

Turn evidence into recommendation and roadmap.

AI is orchestrated, not unleashed.

A sprint runs on shared context, not a room full of private prompts.

Shared context

Evidence, prompts and outputs stay visible in the same workspace.

Human-first creativity

People make the first creative leap. AI stretches and challenges it.

Real users decide

Synthetic users can pre-screen. Final evidence comes from real target users.

How we prevent AI slop

More is cheap. Better is the method.

Humans commit first. AI challenges before it creates. Every expansion ends in a cut.

What you leave with

The evidence and assets needed to decide and build momentum.

Sharper opportunity

A focused view of the strongest opportunity areas and the assumptions that still need to be tested.

Evidence Atlas
Opportunity Map
Assumptions to test

Decision-grade routes

A small number of strong routes made tangible enough for customers and stakeholders to react to.

Testable concept routes
Decision-grade prototype package
Test questions

Evidence for action

A clear recommendation backed by real-user evidence, commercial hypotheses and a practical 30/60/90 plan.

User Evidence Readout
Decision Pack
30/60/90 roadmap

Sprints that moved decisions

Cyber Security Sprint

From proposition question to investment case.

Shaped and tested a new cyber-security proposition for SMEs. The evidence helped BT take the proposition forward and supported capex approval for the next stage.

Ink Subscriptions Sprint

Turning customer barriers into testable growth experiments.

Identified the barriers holding back subscription conversion and turned them into focused experiments for improving sign-up.

Get the AI Design Sprint overview

A 26-page practical overview of the process and deliverables.