/pricing
What does AI cost for a small business?
A small business should expect to spend $5,900 to $8,600 on AI in its first year at around 10 people, $39,000 to $76,000 at 50 people, and $88,000 to $162,000 at 200 people. Three things make up that number: software licenses at $20 to $30 per person per month, implementation at $15,000 to $40,000 per workflow, and training from $3,500. Implementation is the variable one, and it is driven by how many systems the workflow has to touch, not by how many people you employ. A 12-person company connecting four systems costs more to automate than a 200-person company connecting one.
What AI costs at 10, 50, and 200 people
| Cost | ~10 people | ~50 people | ~200 people |
|---|---|---|---|
| Software licenses (per year) | $2,400–$3,600 | $12,000–$18,000 | $48,000–$72,000 |
| Implementation | Usually later | $15,000–$40,000 | $15,000–$40,000 |
| Training | $3,500–$5,000 | $12,000–$18,000 | $25,000–$50,000+ |
| First-year total | $5,900–$8,600 | $39,000–$76,000 | $88,000–$162,000+ |
These license figures assume one paid seat per employee. Most companies start by licensing a single team rather than everyone, so real first-year license spend often lands below these numbers.
- ChatGPT Business : $20 per user per month
- Claude Team : $20 per user per month
- Microsoft 365 Copilot : $30 per user per month
License prices as of August 2026, taken from the vendors' own published pricing pages linked above and correct on that date, so check them before relying on the totals, since vendors change list prices without notice. Implementation and training figures are Ronan Keane's fixed published prices.
What actually drives the cost
Implementation cost is driven by the number of systems that have to talk to each other, not by the number of people you employ. Connecting a CRM to a quoting tool and a shared inbox is three integrations and three sets of edge cases, whether you have 12 employees or 200.
- Systems touched:every integration adds scope, and each one has its own authentication, data format, and failure modes.
- Data condition:if records are inconsistent or scattered across spreadsheets, cleanup comes before automation.
- Number of workflows:the prices on this page are per workflow. Two workflows is not one engagement priced twice, but it is not one price either.
- Compliance requirements:regulated work needs review steps, audit trails, and human approval gates, and those take build time.
What I charge
| Engagement | Price | Shape |
|---|---|---|
| Implementation: one workflow, discovery through adoption | $15,000–$40,000 | Fixed, scoped on systems touched |
| AI Pilot Rescue | $10,000–$20,000 | Fixed, 3–4 week sprint |
| AI Governance Readiness | $10,000–$25,000 | Fixed |
| Training: single department | $3,500–$5,000 | Per session |
| Training: multi-department | $12,000–$18,000 | Program |
| Training: organization-wide rollout | $25,000–$50,000+ | Program |
| Fractional Chief AI Officer | $5,000–$12,000/month | Retainer, month to month |
Every engagement above is fixed-price. You get the number before the work starts, not an hourly rate that moves. The Fractional Chief AI Officer retainer is ongoing and sits outside the first-year build costs shown earlier on this page.
Where to start at your size
| Size | Start with | Year one |
|---|---|---|
| ~10 people | Licenses plus one training session, to get the team using AI well before automating anything | $5,900–$8,600 |
| ~50 people | One workflow implemented end to end, plus multi-department training | $39,000–$76,000 |
| ~200 people | Implementation, organization-wide rollout, and an owner accountable for the number | $88,000–$162,000+ |
A ten-person company is a real client, not a company that should wait. The honest advice at that size is to spend on licenses and training first and leave automation until there is a workflow worth automating, which is cheaper advice than most consultants will give you, and it is still the advice I would give you on a call.
What this looks like for a 50-person company
A representative example, built from the published prices above. It is an illustration of how the numbers assemble, not a specific client engagement.
| Line item | Cost | Basis |
|---|---|---|
| 50 seats of ChatGPT Business | $12,000 | 50 × $20/month, the published list price |
| Quote-to-proposal workflow, built and adopted | $30,000 | Three systems connected, upper middle of the range |
| Multi-department training | $15,000 | Middle of the published range |
| First-year total | $57,000 | Inside the $39,000–$76,000 band above |
The workflow connects the CRM, the proposal generator, and a shared inbox. The training exists because the workflow is worthless if the sales team keeps doing it the old way, which is the most common reason AI projects stall.
Frequently asked questions
How much does AI cost for a small business?
A small business typically spends $5,900 to $8,600 on AI in its first year at around 10 people, rising to $39,000 to $76,000 at 50 people and $88,000 to $162,000 at 200 people. That covers three things: software licenses at $20 to $30 per person per month, implementation of one workflow at $15,000 to $40,000, and training from $3,500. Licenses are predictable. Implementation is the variable, and it scales with the number of systems the workflow touches rather than with headcount.
How much does an AI consultant cost?
Implementation engagements run $15,000 to $40,000 for one workflow taken from discovery through adoption, priced on how many systems it touches. Other engagements are fixed-price: an AI Pilot Rescue is $10,000 to $20,000 across a three to four week sprint, AI Governance Readiness is $10,000 to $25,000, and training runs $3,500 to $5,000 for a single department, $12,000 to $18,000 across multiple departments, or $25,000 to $50,000+ for an organization-wide rollout. Ongoing advisory as a fractional Chief AI Officer is $5,000 to $12,000 a month.
What does a typical AI consulting engagement look like?
A typical engagement is a single workflow taken from discovery through to adoption for a fixed $15,000 to $40,000. It starts by mapping where the time actually goes, then builds and integrates the workflow, then trains the people who have to use it, because the common failure mode for AI projects is not the build, it is adoption. Scoping to one workflow rather than a broad transformation program is what keeps the price fixed and the outcome measurable.
Not sure which of these you need?
That is a reasonable place to be. A short call is usually enough to tell you which engagement fits, or whether you should be spending on licenses and training first instead.
Book a discovery call