Sound familiar?
“We're under pressure to add AI, and we've already built a few things. Now every customer, competitor and stakeholder has a different idea of what we should do next. I don't want my engineers spending six months on AI features that don't really matter.”
Most teams have already answered “should we use AI?” The harder question is what AI should do next, now that you have some. That is the question the sprint answers.
The AI Product Decision Sprint
Three weeks from AI ambiguity to a product decision.
We diagnose the product, rank the AI opportunities, run the build or don't-build decisions with your leaders, set the experience direction, and hand over a roadmap your team can run. You won't get a pile of Figma files, research reports or brainstorms. You get clarity before you commit engineering.
01
Diagnose
Week 1
We look at the product as it stands: customer and user problems, the AI you have already shipped, the competition, and the real limits on Product, Design and Engineering. Then we judge each opportunity on value and feasibility.
AI product diagnosis
AI opportunity map
Competitive positioning insights
Value you get: one honest picture of where AI could create value in your product, and where it won't.
02
Decide
Week 2
We turn the opportunity map into decisions. We prioritise against value and feasibility, then run a leadership workshop so Product, Design and Engineering agree on what gets built, and just as clearly, what doesn't.
Prioritised AI opportunities
Build / don't-build recommendations
Leadership alignment workshop
Value you get: decisions the whole leadership team has made together, not a list of suggestions.
03
Direct
Week 3
We give the decisions a shape. An AI product thesis, an experience direction, and one prototype that makes the key bet concrete. Then we sequence it into a 90-day roadmap with measures of success.
AI product thesis
Experience direction
Evidence prototype / concept
90-day AI roadmap
Executive decision deck
Value you get: a direction and a 90-day roadmap your team can start on the following Monday.
Built for founders, CEOs, CPOs, VPs and heads of product at B2B SaaS companies from Seed to Series C that have an established product and are already investing in AI.
The shift
AI ambiguity → AI product clarity.
Before
Lots of AI ideas
Competitive pressure
Customer requests pulling in every direction
Stakeholders with different agendas
Unclear priorities
Product, Design and Engineering disagreeing
No coherent AI product thesis
Engineering committed without enough confidence
AI features judged mainly by adoption
After
A clear AI product thesis
A prioritised opportunity map
Explicit build / don't-build decisions
A clear experience direction
Shared leadership understanding
Product, Design and Engineering aligned
A 90-day roadmap
Defined measures of success
Selected work
Six case studies, six kinds of complex
Design → HTML
Enterprise eProcurement
Corporate buying spread across four catalogue types and five kinds of user, with an approval chain nobody could see until it stalled. We wrote the vision and the user groups first, mapped the procure-to-pay ecosystem, then cut the IA from it — spend limit and balance riding in the header, the 26-approver path drawn before submission. Specified to pixel level, then built as front-end pages.
Before
After — one surface
3 tools → 1
0–1 Product
A net-new supervisor desktop, built from three tools never designed for the job — monitoring, metrics and messaging in one modular layout, with smooth adoption across migrating enterprises.
AI across four horizons
H1
Context, not intelligence
H2
Retrieval meets the agent
H3
Consolidated under one Assistant
H4
The AI-native rethink
Transcripts → retrieval → the Assistant → a first-principles rethink
Transcripts → AI-native
AI Horizons
Five years of AI across four horizons — each wave re-diagnosed rather than re-skinned: agent attention protected, retrieval designed around the moment it mattered, a flood of features consolidated under one Assistant, and the old shape questioned from first principles.
POSITIONING
Assistant bolted on the side
Native, single-brain architecture
STRUCTURE
Broken flow, vibe-coded finish
Repaired, on top of settled IA
CATEGORY FIT
Crowded, undifferentiated
Positioned in empty wedge
STORY
Unclear to VCs and the team
One fundable, buildable direction
Side feature → core
AI Native
An 'AI-native' CRM that wasn't — turning a label into a real architecture, and a pitch problem into a fundable, defensible product concept.
Catalogue → shelf
Online Grocery
India's online grocery leader needed people to find more than they came for. We studied how a real store lights, zones and stacks its shelves, then rebuilt browsing on that — products resting on drawn shelves and gourmet aisles you walk, instead of a grid you query.
Log rows → topology
Security Operations
Analysts triaged hundreds of open threats out of log tables, pattern-finding by hand under pressure. We cut the first thirty seconds to three questions — how bad, where, who — and made them lenses over one canvas, so threat volume reads as shape before it reads as a number. The POC opened a larger programme.
About
Product judgement, not just screens.
Twenty-five years in UX and enterprise product strategy, and more than seven of them spent designing AI products and features. We bring research, metrics and product delivery into the room, and get executives aligned around a single direction.
We work with founders and product leaders at B2B SaaS companies from Seed to Series C, where the product is established, the customers are real and the AI investment is already under way.
7+ yrs
In AI product work
25 yrs
In UX
25+
Products shipped
Engagements
Start with the sprint. Go deeper where it counts.
Main offer · 3 weeks
AI Product Decision Sprint
For B2B SaaS companies from Seed to Series C that already have a product and are exploring AI. You move from AI opportunities to prioritised decisions, a product direction and a 90-day roadmap.
What you're really buying is decision confidence: knowing what to build, what not to build, and why.
Book a call
Also available
AI Feature Value Audit
For teams already running 3–10 AI features
“We've launched AI, but we don't know which features are actually creating value.”
A clear view of what's working and what isn't, and why. What to improve, what to retire, and where to invest next. Includes a feature inventory, adoption and journey analysis, a metrics framework and a prioritisation matrix.
Also available
Product Experience Reframe
For enterprise SaaS that has become fragmented or complex
“Our product has grown faster than the experience.”
A coherent future-state experience and a transformation roadmap: experience diagnosis, journey map, principles, target-state experience, a concept prototype and a prioritised roadmap for your leadership team.
Free checklist
The AI Build / Don't-Build Checklist
Before you lock your 2027 roadmap: ten checks to run on every AI idea before it gets a date. Five tests, about ten minutes per idea, with your product and engineering leads in the room.
Open the checklist
No sign-up. Print it or save it as a PDF.
Contact
Planning your next AI roadmap? Let's talk before you build.
Book a 30-minute call below. We'll talk through where your AI work stands and whether a sprint fits.
How to reschedule your meeting
- Open your email inbox and find the booking confirmation message from Calendly.
- Scroll to the bottom of the email or your calendar event.
- Click Reschedule to select a new date and time instantly.
Prefer to write first?