The gap between what businesses expect from AI consulting and what actually happens in the first 90 days is one of the most predictable sources of client dissatisfaction in the industry. Setting expectations clearly before an engagement starts prevents most of it.
Weeks 1-2: Discovery and Diagnosis
The first two weeks of a well-run AI consulting engagement are not about AI. They're about understanding the business. This means stakeholder interviews, process documentation review, current technology stack assessment, and data audit.
Deliverable: a clear map of where the business currently operates, what the highest-value problems are, and what constraints exist on solving them.
What clients often expect instead: a technology recommendation. This is the wrong starting point. Consultants who skip discovery and go straight to solutions are selling products, not solving problems.
Weeks 3-4: Opportunity Prioritisation
With discovery complete, opportunities are evaluated against three criteria: business impact, implementation complexity, and data readiness. The output is a prioritised list of AI opportunities with rough ROI estimates and implementation timeframes for each.
Deliverable: an AI opportunity register with the top three to five initiatives ranked by recommended sequencing.
Weeks 5-8: First Implementation
A well-sequenced engagement moves directly from prioritisation to implementation of the first initiative. This is not a pilot — it's a production deployment of the highest-confidence, highest-value opportunity identified in discovery.
This phase involves vendor evaluation (if applicable), integration scoping, build or configuration, testing, and staff training. Eight weeks is tight but achievable for a well-scoped first initiative.
Deliverable: a working AI implementation in production, with baseline metrics captured and a 30-day measurement plan in place.
Weeks 9-12: Measurement and Roadmap
The final phase of the initial engagement focuses on two things: measuring what was built, and planning what comes next.
Real performance data from the first implementation informs the assumptions used in subsequent business cases. The AI roadmap built in weeks 11-12 is grounded in actual results rather than projections.
Deliverable: a 12-month AI roadmap with sequenced initiatives, resource requirements, and business cases for the next two to three implementations.
What 90 Days Doesn't Deliver
A 90-day engagement does not transform an organisation's AI capability. It delivers one working implementation, validated assumptions, and a credible roadmap. Organisations that expect wholesale transformation in 90 days are setting themselves up for disappointment.
The value of the first 90 days is establishing a working model — a repeatable process for identifying, implementing, and measuring AI initiatives — that the business can apply independently over time.
How to Get the Most From the Engagement
Assign a dedicated internal champion with decision-making authority. Ensure data access is arranged before week one. Communicate to affected staff before the engagement starts, not after. And hold the consulting team accountable to the deliverables above — not to activity metrics like hours billed or presentations delivered.