AI consulting project and retainer pricing in 2026
Hourly rates describe labour; projects and retainers are how AI work is actually bought. The published prices for each stage, what they include and the running cost that follows.
Contents
Buying AI work in pieces
Most companies buy AI consulting in three pieces: an assessment that decides what to do, a build that delivers something working, and an arrangement for running and changing it afterwards. The difficulty is that each piece is priced differently, by different kinds of firm, and the third is often missing from the proposal altogether. The published prices below show what each piece costs when firms are willing to print it.
Stage one: the assessment
| Offer | Published price | Duration | Deliverable |
|---|---|---|---|
| AI readiness audit, Main & Machine | $3,500 to $8,5001 | 2 to 4 weeks | Workflow map, opportunity assessment and a phased plan the client owns |
| Strategy and assessment, Team 400 | A$20,000 to A$50,0002 | 2 to 4 weeks | Strategy and assessment |
| A single expert session, Ryan Doser | $500 an hour3 | One session | A plan the client carries out |
Main & Machine credits the audit fee toward a sprint signed within 60 days, up to 25% of the sprint price1. That structure is worth asking any firm for, because it lowers the cost of the assessment when it leads to a build.
Stage two: the build
| Offer | Published price | Duration |
|---|---|---|
| Implementation sprint, one workflow, Main & Machine | $12,000 to $45,0001 | 4 to 12 weeks |
| Proof of concept, Team 400 | A$20,000 to A$50,0002 | 2 to 4 weeks |
| Production AI agent or application, Team 400 | A$50,000 to A$250,0002 | 6 to 16 weeks |
| Multi-department engagement, Main & Machine | From $95,0001 | One to two quarters |
| AI development projects reviewed on Clutch | Most $10,000 to $49,999; average about $120,6004 | About 10 months on average |
The Clutch figures show how skewed project costs are: the average sits well above the most common range because a minority of large projects pulls it up4. A buyer should compare a quote with the stage it describes, not with an average across all projects.
Delivery terms matter as much as the price. Main & Machine states that if a scoped workflow is not live within 90 days it continues building at no charge, and frames this as a delivery guarantee rather than a promise of return1. Toptal, which publishes no rates, offers a trial of up to two weeks that is not billed if the client is unsatisfied5.
Stage three: running and changing it
This is the stage most proposals leave out, and the one that continues for years.
| Arrangement | Published price | What it buys |
|---|---|---|
| Coaching retainer, Ryan Doser | $999 a month3 | One 90-minute working call, a community and three products |
| Managed services, Main & Machine | From $1,500 a month1 | Support for the system after launch, with no lock-in |
| Tools and model usage, small business | $50 to $500 a month1 | Planning band, excluding build and services |
| Managed hosting, OpenLazarus | From $150 a month6 | Backups, updates and monitoring, movable to the client's own cloud |
| Developer support, OpenLazarus | From $4,500 a month6 | 10 engineering credits a month for fixes, features and integrations |
| AI usage, OpenLazarus | 5.5% on credits6 | Any model through its router |
Two models are visible here. A managed service keeps a consultant responsible for a system it built. A platform model, such as OpenLazarus's, gives the company a system it owns plus a monthly engineering allowance; its AI-Architect consultancy, which designs that system with leadership, is priced per engagement and recommended for companies above 100 people6.
Structuring a purchase
- Buy the assessment at a fixed price, and ask for the fee to be credited against the build.
- Name the deliverable of the build: one workflow, live in the company's own systems, by a date.
- Put a delivery term in writing: what happens if it is not live by that date.
- Price twelve months of running it before signing the build: support, hosting, software and usage.
- Keep ownership: the plan, the configuration and the data stay with the company, whichever firm runs it.
The result is a purchase with three prices on one page, an assessment, a build and a year of running costs, each tied to a deliverable. The main guide compares the providers on how much of this they publish, and the ROI guide shows how to set those prices against the saving.
Questions buyers ask
How much does an AI readiness assessment cost?
Published prices range from $3,500 to $8,500 for a two-to-four-week readiness audit from a US small and mid-size business consultancy, to A$20,000 to A$50,000 for a strategy and assessment from an Australian AI consultancy. The deliverable should be a workflow map, a ranked list of opportunities and a phased plan.
How much is an AI consulting retainer per month?
Published retainers and running costs include coaching at $999 a month, managed services from $1,500 a month and developer support on a company-owned platform from $4,500 a month with 10 engineering credits. Clutch reports an average of about $11,550 a month across AI development projects reviewed on its directory.
Should we pay for AI consulting by the hour or by the project?
Fixed-price projects suit a defined deliverable, such as an audit or one workflow, because the risk of overruns sits with the firm. Hourly work suits advice on a single decision. Monthly retainers suit ongoing change once a system is running.
Sources
- What an AI consultant costs. Main & Machine. Accessed 8 Nov 2026.
- AI consulting cost in Australia. Team 400, 3 Apr 2026. Accessed 8 Nov 2026.
- How much does an AI consultant cost in 2026? (My rates included). Ryan Doser, 4 Sep 2026. Accessed 8 Nov 2026.
- AI pricing guide. Clutch, 21 Sep 2026. Accessed 8 Nov 2026.
- Hire AI engineers. Toptal. Accessed 8 Nov 2026.
- Pricing. OpenLazarus. Accessed 8 Nov 2026.
Updated 8 Nov 2026: First published.