AI Answer Library

How should a company phase its first AI project?

Short answer

Four steps: pick one narrowly scoped use case, build an evaluation set, ship a minimum usable version, and only then discuss rollout. The goal of a first project is not an impressive system but an organisation that has learned how to judge one. Without an evaluation set there is no basis for judgement, and every later debate collapses into "it does not feel right". Choose a use case with real users where failure does not touch the core business — do not practise on a critical process.

Key points

  • 01The real output of a first project is judgement, not software. Afterwards the team should be able to say what "good" means — that matters far more than feature count.
  • 02Build the evaluation set before development, and have the business side write it. Adding it after the code means letting the implementer set their own exam.
  • 03Three conditions for the use case: real recurring users, boundaries clear enough to state what is out of scope, and failure that does not touch the core business. Missing any one sharply raises the odds of stalling.
  • 04Decide technology late and keep it plain. Get the whole chain working with the simplest approach, let the real problems surface, then decide whether complexity is warranted.
  • 05Launch is not the finish line; handover is. If only the vendor can change the system, it stops evolving in year two.

Each phase, with its exit criterion

The table expands those four steps a little further — "rollout" actually contains two distinct stages, a limited live trial and an operations handover, and plenty of projects die in exactly those two. What matters is not the phase list but that each phase has an explicit test for moving on; phases without exit criteria merely slice one blob of work into several and reduce no risk. Put the last column into the project plan at kickoff, and agree that an unmet criterion means slipping the date rather than skipping the gate.

PhaseGoal of the phaseDeliverableExit criterion
1. Use-case selectionFind a bounded problem that real people actually haveA one-page brief: who uses it, how often, and how it is done todayYou can state what happens if nothing is built, and a named business owner exists
2. Evaluation setTurn "is it good" into something measurableReal questions, reference answers and a scoring rubric, authored by the business sideTwo people scoring independently under the rubric reach broadly the same result
3. Minimum usable versionGet the entire chain working with the plainest possible approachA system real users can open and operate themselvesEvaluation-set performance has converged instead of swinging with every change
4. Limited live trialSurface the problems the evaluation set cannot coverUsage logs, a failure-case list and recorded user feedbackSomeone keeps using it without being told to
5. Rollout and handoverMake your own team able to maintain and evolve itAn admin console, operations documentation and a real handover sessionAn in-house engineer ships one change end to end without help

Choosing the first use case

Choosing the use case matters far more than choosing the technology, and it is where teams most often go wrong. Three failure patterns recur. First, picking something that looks valuable but nobody uses daily — an analytics assistant for executives, say. Low frequency means no feedback, and the project quietly stops at the demo stage. Second, picking a core business process: once mistakes affect production, the team adds more and more manual review out of caution until the automation benefit reaches zero. Third, leaving scope open, as in "we want it to answer anything" — such projects always have someone dissatisfied at acceptance. A workable filter requires three things simultaneously: a fixed group hits this problem every week; the current way of doing it can be described precisely; and if the output is wrong, someone downstream will notice and correct it. Use cases meeting all three are rarely impressive, but they succeed far more often — and what a first project needs most is to succeed once.

The organisational prerequisites usually decide the outcome

Three things must be in place before kickoff. First, a business owner with decision authority, not merely a technical contact. AI projects continually raise judgement calls about how a given situation should be answered, and with nobody empowered to settle them the project stalls in endless discussion. Second, clear ownership of the data: who decides which documents enter the knowledge base, which version is authoritative, and who may see what. These are technically trivial and organisationally hard, and solving them early saves the most time. Third, calibrate expectations explicitly: the system will make mistakes, version one will be worse than a human, and the value comes from absorbing a large volume of easy cases rather than replacing experts. Skip this and the first visible error becomes "this thing does not work". None of the three requires technical investment, yet their absence is the most common cause of failure — and no amount of technical strength compensates for it.

Where this applies

When this answer does not hold

  • This targets a company's first AI project. Organisations already running several AI systems should shift focus to platform consolidation, reuse and unified governance, which phases differently.
  • With no technical counterpart in-house, risk rises sharply regardless of vendor or phasing. That gap must be closed first.
  • In heavily regulated sectors — healthcare, finance, classified environments — compliance review belongs in the use-case selection phase. Deferring it to pre-launch forces architectural rework.
  • The phase order assumes reasonably stable requirements. If the business itself is changing fast, shorten each phase and iterate more often rather than stretching any single phase.

People also ask

  • Where should a company start with AI?
  • How do you run an AI pilot that does not stall out?
  • Which use case should be the first AI project?
  • What are the implementation steps for putting AI into production?
  • How do you tell whether an AI project is actually working?
Written by: YGG Technology Solutions TeamPublished: 2026-08-01Last reviewed: 2026-08-01