Uniting payors, providers, and pharmacies for seamless care.
53M+
Members supported
100%
Compliance Rate
- Strategy
- Web
- App
A product discovery workshop turns an idea into a specification, wireframes, an architecture and a fixed build price in two to four weeks. An AI strategic discovery workshop answers a different question first: where AI can move a number in your business, which use case to prove first, and whether your data can support it. Both run the same way at Ailoitte: the architect and designer who would later build the product are in the room, a mutual NDA comes first, and everything they produce is yours whoever builds it.
Tell us the idea or the problem in a 30 minute call. We will say which track fits, whether you need discovery at all, and send a fixed price for the workshop within a week. No preparation is needed and nothing is owed to find out.
Two tracks, one method
Product discovery or AI strategy, same team and rigor.
Everything is yours
Specification, designs and plan assigned to you on payment.
Ends in a decision
A costed plan, a fixed build quote, or a reason not to build.
A short, fixed-price engagement that turns an idea into a plan you can budget, staff and build against. The output is a decision, not a document nobody reads.
Product discovery when you know what you want to build and need it specified, designed and priced. AI strategic discovery when you know AI matters but not where to start.
Rapid Concept Discovery runs two weeks. Comprehensive Discovery runs four and adds full visual design and an interactive prototype. The AI track runs about four weeks around two working sessions.
On the product track, eight documents from functional specification to project plan. On the AI track, a ranked use-case shortlist, a data assessment, a proof of concept recommendation and a roadmap.
About $4,700 for Rapid and about $8,000 for Comprehensive, fixed before we start. If you then build with us, part of the fee is credited against the build; the amount depends on the project and is written into the proposal.
You do, on payment. Take it to your own team or another vendor if you like. If you build with us, the fixed build quote is drawn straight from the discovery output.
Most software projects do not fail in the build. They fail in the weeks before it, when nobody wrote down what success meant, which systems the product had to talk to, or whether the problem was worth solving at all. PMI found that 47 percent of unsuccessful projects miss their goals because of inaccurate requirements (PMI Pulse of the Profession). NASA’s cost data, cited by Deloitte, puts a requirements error at one unit to fix while it is still a requirement and 21 to 78 units once it reaches integration and testing (Deloitte, citing NASA). McKinsey and Oxford studied 5,400 large IT projects and found the average ran 45 percent over budget and delivered 56 percent less value than promised (McKinsey).
AI makes the picture sharper, not softer. RAND interviewed 65 senior data scientists and engineers and put the AI project failure rate near 80 percent, double the rate for ordinary IT, with the first root cause being that leaders misunderstood which problem needed solving (RAND, August 2024). Gartner expected 30 percent of generative AI projects to be abandoned after proof of concept by the end of 2025, naming poor data quality, weak risk controls, rising costs and unclear business value (Gartner, July 2024). MIT’s NANDA group went further in 2025: 95 percent of enterprise generative AI pilots produced no measurable return on $30 to 40 billion of spend (MIT NANDA, 2025). None of those failures were engineering failures. They were discovery failures.
That is also why we judge our own discovery by one test: did it change what you are going to build, or only describe it more thoroughly? Twelve interviews, three personas and a competitor matrix that left every decision where it was is expensive theater. Two assumptions tested and one disproved, an integration confirmed as feasible, a third of the scope cut: that is progress, because you now know something you did not know. If a session is not going to change a decision, we would rather not run it and hand the time back. The same discipline carries into the build, where it becomes a fixed price that holds.
Send us what you already have. If the scope is clear enough to quote, we will say so and skip straight to a build price.
They answer different questions and run the same way. Many clients run the AI track first and then a product track for whatever it identifies as worth building. If you are not sure which you need, the first call settles it.
Answers: what should we build, and how?
Answers: can AI move a number here, and which use case first?
Discovery that skips the first half produces a beautifully specified answer to a question nobody asked. Where teams go wrong is arriving with a solution already decided and asking us to write it up. We will do that if you insist, but we will first spend a session testing whether the problem behind it is the one worth your budget.
Is this problem worth solving?
Exit condition: a problem worth building for, stated in one sentence.
Is this solution worth building?
Exit condition: a costed, sequenced plan you can commit to.
Two questions decide whether a finding should change your plan. Is it real or anecdotal? Is it relevant or lip service? Every finding in your discovery report carries one of these four labels, so you can see which conclusions are load-bearing and which are still hypotheses. One principle follows from it: discovery investment should match the stakes of the decision, not chase certainty. A two-week choice does not need two months of research, which is why there are two package sizes rather than one.
Observed behavior from someone who owns the problem and would pay to fix it. This changes your plan immediately.
Genuine behavior, but from someone who is not your buyer. Interesting, and not actionable.
The right person said the right thing once. A hypothesis to test, not a fact to build on.
"That sounds useful, I would definitely use it." Says nothing. Stated intent is not evidence.
Each one is structured so a development team, ours or anyone else’s, can start sprint one without a second round of questions. Comprehensive Discovery adds Figma user flows, journey mapping and an interactive prototype you can test with real users before committing to a build. All of it assigns to you on payment.
Stakeholder sessions turned into goals, constraints, users and a measurable definition of done. Written as something you can test, not a feature list.
The FSD: every function written down and agreed before anyone builds. The foundation for sprint planning and for the fixed build price.
Screen by screen structure first, then visual design once the flow is settled. Rapid stops at low fidelity; Comprehensive carries through to a full design system.
How it will be built, hosted and scaled, written down and reviewed. Data flow, frameworks and the compliance controls the architecture has to carry.
Scope decomposed into estimable, buildable, testable units, ranked so the backlog can be reordered without breaking the plan.
A contract for every system boundary, verified against what your systems actually expose rather than what the brochure says.
The SDKs and services worth buying rather than building, with integration feasibility, plus hosting options and their running costs.
Sprint plan, milestones, team shape and a costed timeline, ending in a fixed price for the build you can accept, take elsewhere or decline.
A rapid concept discovery workshop gets you from idea to a fixed build quote in two weeks.
The AI track runs as two working sessions of about four hours each, one to two weeks apart. In the first we ask and you talk: strategic goals, who your customer is, what a working day looks like, which tasks drain the most time, what data you collect and where it lives. The gap is the point. It gives us time to develop the opportunities properly and gives you time to think rather than react in the room. In the second session we come back with around five specific project proposals, what each would involve, how long each takes and roughly what it costs, and we assign priorities together. Where AI does not make sense in an area, we say so in that session.
Where AI can move a number you already track, and whether your processes, people and data are ready for it.
A ranked shortlist scored on value, feasibility and effort. Usually around five concrete proposals, not a strategy deck.
What data you hold, its quality, the gaps, and what it can realistically support. The single most common reason AI projects stall.
Models, deployment shape, private or hosted, the integration points, and the guardrails that go around them.
Where the human stays in the loop, exactly where approval gates sit, and what the user sees when the model is unsure.
Which single use case to prove first, the threshold that counts as proof, and the time box for proving it.
Data privacy, model risk, audit trail and the regulations in scope, from HIPAA and GDPR to the EU AI Act, written into the plan rather than retrofitted.
Sequenced and costed, with governance and success metrics attached to each phase, and a fixed price for the first phase if you want one.
Book the AI track. You leave with five ranked proposals and one to prove first, or a clear reason to wait.
Rapid for a defined idea. Comprehensive for anything with real integration, compliance or multi-role complexity. The AI track when the question is not how to build it but whether and where AI pays. Discovery is priced separately from the build on purpose: a recommendation you have paid for independently is worth more than one bundled into a sales process, and it means we can tell you not to build.
| On | Rapid Concept Discovery | Comprehensive Discovery | AI Strategic Discovery |
|---|---|---|---|
| Duration and effort | Two weeks. | Four weeks. | About four weeks around two four-hour sessions. |
| Best for | A defined idea that needs validating, specifying and pricing fast. | Anything with real integration, compliance or multi-role complexity, or where design has to be tested before the build. | Leaders who know AI matters and need a ranked, costed list of where it would pay. |
| What you get | FSD, low fidelity wireframes, solution architecture, third party and infrastructure analysis, estimate and fixed build quote. | Everything in Rapid, plus high fidelity design, Figma user flows and journey maps, an interactive prototype, API definitions, compliance scope and a sprint-level costed roadmap. | AI readiness, ranked use cases, data assessment, AI architecture, workflow maps, PoC recommendation, governance guardrails and a roadmap. |
| Who is in the room from us | Solution architect, business analyst, product designer. | Solution architect, business analyst, product designer, and a domain specialist where compliance applies. | Solution architect, AI strategist, data scientist, product designer. |
| Ends with | A fixed build quote and a first-release date. | A tested prototype, a fixed build quote and a first-release date. | One use case to prove first, with a fixed price for the proof of concept. |
| Price | About $4,700, fixed. Part credited if you build with us. | About $8,000, fixed. Part credited if you build with us. | Fixed price in the proposal, within a week of the call. |
Comprehensive is the most chosen package. The extra two weeks go on things that are expensive to get wrong later: high fidelity design across every screen, an interactive prototype tested with real users, API contracts verified against live systems and a compliance scope written into the architecture. If you are not sure, start Rapid. We will tell you by the end of the first week if the complexity calls for Comprehensive, and the work done so far carries straight into it.
An agile discovery workshop defines enough to be accountable for a plan without freezing a specification. Scope is ranked rather than fixed, architecture is chosen for change, and the output is a prioritized backlog you can reorder. What we do not do is start building before anyone has articulated the outcome.
Turn up and talk about your business. If you want to arrive better prepared, bring whatever already exists: a rough brief, existing designs, API documentation or a list of the systems involved. None of it is a precondition.
Whoever can approve scope trade-offs attends the sessions. That single factor decides whether discovery finishes on time more than anything on our side. Four to eight people is the working range; beyond that, sessions become presentations.
Sessions are recorded with your consent, every exercise lives on a shared board you keep access to afterwards, and a written summary of decisions, open questions and owners goes out within a day, not at the end.
The architect and designer in your sessions are the ones who would deliver the build, not a pre-sales team who hand over notes. There is no account manager between you and them.
Rapid compresses these into two weeks, Comprehensive gives them four, and the AI track runs them around two working sessions. Every stage ends with something written that you keep. Delivery after discovery runs on Clarity, our delivery method.
A 30 minute call picks the track. Then stakeholder sessions on goals, constraints, users and success criteria, with the mutual NDA signed before any of it.
Integrations verified against what your systems actually expose. Data assessed for the AI track. Compliance scope confirmed. Risks logged, each with an owner.
Journeys mapped, wireframes drawn, architecture chosen and documented for review. On the AI track, use cases are scored and the architecture for the leading one is sketched.
Functional specification, use cases, epics and user stories, API definitions. Comprehensive adds the interactive Figma prototype to put in front of real users.
Project plan, sprint breakdown and a fixed build quote, or a proof of concept recommendation with its price. Then your decision, with no obligation attached.
A design discovery workshop lives or dies on who attends. The people on our side are the ones who would deliver the build. The people on yours are the ones who can decide.
The first card is a published discovery phase. The rest are builds whose scope, architecture and compliance posture were settled before sprint one. See all case studies.
Signed references, each linked to the person and the work.
Rapid Concept Discovery runs two weeks and costs about $4,700. Comprehensive Discovery runs four weeks and costs about $8,000. The AI strategic discovery workshop is priced in the proposal within a week of the first call. Every price is fixed before the workshop starts and does not move unless you change the scope. If you go on to build with us, a part of the discovery fee is credited against the build. How much depends on the project, and the proposal states the figure before you sign.
Against the build it precedes, discovery is small. A $50,000 MVP that avoids one unverified integration or one feature nobody needed has paid for Rapid many times over, and a fixed build price is only possible because a vendor can commit to a number against a defined scope, not against an idea. Discovery is priced separately from the build so the recommendation stays independent, including the recommendation not to build.
If you proceed with us, the build is quoted as a fixed price per feature drawn directly from the discovery output, usually $20,000 to $50,000 for an MVP in 6 to 10 weeks. If you take the documents to another team, they are complete enough to start from.
Each external system we have to verify adds investigation time. Systems we cannot inspect before quoting carry a risk allowance.
HIPAA, GDPR, SOC 2 evidence or sector rules are assessed inside the workshop so the architecture carries them. They add sessions, not surprises.
A product with one user type is specified faster than a marketplace with four. Comprehensive exists for the second kind.
Low fidelity wireframes validate flow. An interactive prototype tested with users is what the extra weeks of Comprehensive mostly buy.
On the AI track, the data assessment is the variable. Clean, accessible data shortens it; data nobody has looked at lengthens it.
Most workshops run remotely in your working hours. On-site days for the opening sessions are available and priced at cost.
Confidentiality is not an upsell. All three of these come with every workshop and cost nothing.
A product discovery workshop is a short, structured engagement that turns an idea into a costed, buildable plan before development starts. It clarifies the goal, maps users and workflows, settles the architecture, breaks the scope into estimable units and produces documentation and designs you own. The output is a decision: what to build, in what order, for what budget.
An AI strategic discovery workshop answers whether AI can move a business metric, which use case to attempt first and whether your data supports it. It runs as two working sessions one to two weeks apart and ends with around five ranked proposals, a data assessment, a proof of concept recommendation and a roadmap. Where AI does not make sense in an area, we say so.
Product discovery answers what to build and how. AI strategic discovery answers whether and where AI pays. Both run the same way, with the same kind of team, and produce the same kind of costed roadmap. Many clients run the AI track first, then a product track for whatever it identifies as worth building.
Two weeks for Rapid Concept Discovery, four weeks for Comprehensive Discovery, and about four weeks for the AI track. Rapid suits a defined idea that needs validating and pricing. Comprehensive suits anything with real integration, compliance or multi-role complexity, and includes full visual design and an interactive prototype.
Our two-week product track for a defined idea, ending in a functional specification, low fidelity wireframes, a solution architecture, a third party and infrastructure analysis, and a fixed build quote. It is the fastest route from an idea to a number you can take to a board or an investor.
About $4,700 for Rapid Concept Discovery and about $8,000 for Comprehensive Discovery, both fixed before the workshop begins; the AI track is priced in the proposal within a week of the call. If you build with us afterwards, part of the fee is credited against the build, with the amount set per project and stated in the proposal. The workshop is priced separately from any build so the recommendation stays independent.
Yes, entirely. Everything assigns to you on payment and you are free to take it to another partner or your own team. Discovery that only works if you then hire us is not discovery, it is a sales process.
Agile. We define enough to be accountable for a plan, not a frozen specification. Scope is ranked rather than fixed, architecture is chosen for change, and the output is a prioritized backlog you can reorder. What we do not do is start building before anyone has written down the outcome.
A design discovery workshop puts user research, journey mapping and interface prototyping at the center, and it is what Comprehensive Discovery is when adoption risk is higher than technical risk. You leave with tested wireframes and an interactive Figma prototype rather than only a document. Our UI and UX design team runs it.
By judging every session against one test: will this change a decision, or only describe it more thoroughly? Findings are labelled by evidence strength, so you can see which are decision-grade and which are still hypotheses. If a session is not going to change a decision, we would rather not run it and hand back the time.
Whoever can make decisions, which is the single biggest factor in whether discovery finishes on time. Typically a product owner, someone who knows the current process end to end, a technical contact if you have existing systems, and on the AI track whoever owns the data. Four to eight people is the working range.
No. Turn up and talk about your business. If you want to arrive better prepared, bring whatever already exists: a rough brief, existing designs, API documentation or a list of the systems involved. None of it is a precondition, and its absence is not a problem.
Yes, and it should wherever they apply. HIPAA, GDPR, SOC 2 evidence and sector rules are assessed during the workshop so the architecture accounts for them, rather than being retrofitted after a security review blocks a launch. See our HIPAA and security and compliance pages.
That is a legitimate and valuable outcome, and it happens. Finding out in two weeks that the market signal is not there, the integration is not feasible or a platform already solves the problem is a far better result than finding out in nine months.
A fixed-price proposal for the build or the proof of concept, with scope, milestones and acceptance criteria drawn directly from the discovery output. Accept it, take the documentation elsewhere, or do nothing. There is no obligation attached to the workshop fee. Most product tracks continue into MVP development.
Send it over. A solution architect replies within 12 hours.
A 30 minute call establishes whether you need Rapid, Comprehensive, the AI track, or no discovery at all. Mutual NDA first, then a fixed price for the workshop within a week. If we think you can skip it, we will say so on the call.
Loading available times…
Information Security
Management System
Quality Management
System
You have a Vision, we are here to help you Achieve it!
Your idea is 100% protected by our Non-Disclosure Agreement.
You have a Vision, we are here to help you Achieve it!
Your idea is 100% protected by our Non-Disclosure Agreement.