An AI advisory pilot needs an agreed fee for scoped work even when its technical or commercial result is uncertain. This fictional consulting team earns no software subscriptions or model-ownership revenue. At the selected sold hours and fees, its retained payroll and other costs exceed advisory contribution, so a successful demonstration cannot by itself make the practice profitable.
Price the decision, test and handover actually commissioned, then evaluate the firm’s retained cost separately. Do not use future product revenue or promised client savings to close the advisory budget.
What does the client buy from the AI practice?
The case is a small independent business AI consultancy in a moderate-cost US hybrid office, serving national clients during January 2027–December 2029. Strategy, data-readiness review, workflow pilots, governance and training divide one consulting-hour pool. Autonomous production operation, foundation-model training and assured regulated approval are excluded.
Three salaried consultants provide 480 available hours monthly. At 45% utilization the full-ramp mix sells 216 hours, leaving 264 delivery hours outside sold work. Half-time owner management and half-time operations support are paid prorated budgets without extra delivery capacity. External specialists and project compute are purchases, not an uncounted salaried team.
Which fees support the retained delivery calendar?
The table allocates the 216 sold hours to five distinct advisory scopes. A workflow pilot’s fee is payment for the agreed work and deliverable, not proof that the tested approach will become a production service. Training and governance are separate allocations rather than automatic extra revenue on the same engagement.
Set the business question, baseline, permitted data, evaluation, failure conditions and sponsor acceptance before estimating the pilot. Identify which preparation, revision and handover work belongs in the quoted scope. A demonstration that requires unpriced clean-up can consume the shared team’s unsold calendar and its cash.
| Advisory scope | Sold hours | Fee/hour | Earned fees |
|---|---|---|---|
| AI strategy | 54 | $240.00 | $12,960.00 |
| Data readiness | 43.2 | $200.00 | $8,640.00 |
| Workflow pilots | 54 | $220.00 | $11,880.00 |
| AI governance | 32.4 | $230.00 | $7,452.00 |
| Training | 32.4 | $180.00 | $5,832.00 |
What must be scoped outside retained salary?
Project specialists and compute are included in direct purchases; recurring business software and IT support sit in operating expenses. Keep an assignment’s purchased resources, access period, expected usage and review responsibility separate from the general office budget. No vendor rate or current product capability is established by this scenario.
A pilot should end with the observed result, limitations and a decision about any further work. Production support, security assurance or specialist-domain conclusions need an explicitly agreed resource and responsibility plan. Do not count a hypothetical subscription, savings share or license sale before that business exists.
Do the selected advisory fees cover the firm?
March revenue is $46,764.00. Direct purchases of $2,338.20 and operating expenses of $53,617.00 leave EBITDA of -$9,191.20. The loss includes paid owner management and occurs before depreciation, interest and income tax. First-year net income is -$157,196.96, with losses continuing in the later years.
The 30-day collection assumption adds a timing need to that operating deficit. Proposed uncommitted equity is $500,000.00, with no debt modeled; startup uses are $42,000.00. Minimum forecast cash is $12,063.79. The reserve makes the saved cash path possible, conditional on funding, but does not prove the current advisory offer pays for its team.
Which pilot evidence should guide the next commitment?
Follow a bounded paid pilot through accepted delivery and collection. Measure all consultant preparation, implementation, evaluation, rework, sponsor review and handover, together with specialist and compute purchases. Keep the result separate from the client’s eventual adoption or claimed savings.
Use that evidence to revise the scope, accepted fee, sold workload or retained-cost plan. If the pilot does not justify production use, deliver that conclusion honestly while assessing whether the commissioned work itself earned adequate contribution. Expand only after collected engagements support both the paid calendar and the operating cost base.
Make the next decision with your own evidence
Complete one paid pilot through collection, reconcile all delivery/review effort and project purchases, and revise the advisory fee or retained-cost plan before funding a larger team.
Explore published business plan examples ↗PracticalLoop Advisory is a fictional prelaunch scenario prepared with AI assistance. Inputs are illustrative; forecast results come from the matching saved financial model. Funding is proposed. The diagrams explain the case and do not depict actual premises, established demand or promised earnings. Replace the assumptions with evidence before committing.
Read our methodology