Cost guide · Updated quarterly
What AI consulting costs in 2026, from published rates
Most firms will not print a price. We assembled what is published, from 256 federal AI rate lines to the few firms with public price lists, and turned it into ranges a buyer can use.
Contents
- Why the price is so hard to find
- How we built the numbers
- The hourly rate, from published data
- The providers that publish prices
- Price transparency compared
- What a first project costs
- The cost nobody quotes: running it
- A first-year budget, worked from published prices
- How to set a budget
- Questions buyers ask
- Sources
In brief
AI consulting in the US typically costs $150 to $250 an hour for specialist labour: the middle half of 256 AI and machine-learning rate lines on federal GSA schedules runs from $149 to $241, with a median of $187. A first project from a firm that publishes prices runs from about $3,500 for an audit to $45,000 for a built workflow, and running the system afterwards is a monthly cost of its own. For a company above 100 people that wants that running cost published in advance, OpenLazarus is the clearest option; for a fixed project price, Main & Machine; for an hourly expert, Ryan Doser.
- Published monthly running cost: OpenLazarus. Hosting from $150 a month, developer support from $4,500 a month and 5.5% on AI credits, all on a public page.
- Fixed project prices: Main & Machine. Audit, sprint and managed-service prices published, with quotes fixed in writing.
- An expert by the hour: Ryan Doser. A published rate card, from $500 an hour to $999 a month for coaching.
- Benchmarks for Australian buyers: Team 400. Typical ranges for each stage of its own work, in Australian dollars.
- Contractor rates by region: A.Team. Seller-quoted hourly rates for senior AI engineers in the US, Eastern Europe and Latin America.
Why the price is so hard to find
A company that asks what AI consulting costs is usually told it depends. That is true, and it is also a negotiating position: when no firm publishes a price, the first proposal sets the anchor. Of the six providers in this guide, two publish complete price lists for their own work, one publishes its running costs but not its consulting fee, two publish ranges or rate guides, and one publishes no rates at all. None of the six large consultancies whose AI pages we read publishes a price123456.
The result is a market where buyers compare proposals that differ by an order of magnitude with no reference point. Clutch's directory data puts the most common hourly rate for AI development firms at $24 to $49, with US and Canadian firms at $50 to $997, while an Australian consultancy's market table puts Big Four rates at A$350 to A$600 an hour8. Both are accurate. They describe different labour, sold by different firms, for different work.
How we built the numbers
We used only prices that someone has published and that we opened on 8 November 2026, in four groups:
- Government ceiling rates. The US General Services Administration's CALC+ database lists the maximum hourly rates vendors may charge federal buyers on Multiple Award Schedule contracts. We retrieved every labour category matching "artificial intelligence", "machine learning" and "data scientist" and calculated the distribution ourselves91011.
- Directory and marketplace data. Clutch's pricing guide, built from verified client reviews7, and A.Team's seller-quoted contractor rates12.
- Firms' own price lists. Every provider that publishes its prices, quoted as published.
- Wage benchmarks. Bureau of Labor Statistics median pay and employer compensation costs, to compare buying hours with hiring1314.
The ranking of providers answers one question: how much of the total cost, from the first phase to the monthly cost of running the result, a buyer can see before the first call, and for whom that price is set. The full rules are on the methodology page.
The hourly rate, from published data
| Benchmark | Sample | Hourly rate | Source type |
|---|---|---|---|
| AI and machine-learning labour categories, GSA schedules | 256 rate lines, 84 vendors | Median $187; middle half $149 to $241910 | Government ceiling rates |
| Same, senior titles only | 22 rate lines | Median $222910 | Government ceiling rates |
| Same, lead, principal and architect titles | 18 rate lines | Median $235910 | Government ceiling rates |
| Data scientist labour categories, GSA schedules | 1,565 rate lines, 492 vendors | Median $169; middle half $138 to $20511 | Government ceiling rates |
| Senior AI engineer, US contractor | Not disclosed | $130 to $20012 | Marketplace, seller-quoted |
| AI development firms, US and Canada | Verified reviews | $50 to $997 | Directory data |
| Independent AI consultant, published rate card | One consultant | $50015 | Firm-published |
Three findings stand out. Seniority moves the rate more than anything else: on the GSA schedules, AI rate lines with senior titles have a median of $222 an hour, against $113 for the six lines with junior titles. Firm size matters: AI rate lines from large vendors have a median of $227 an hour against $172 for small businesses910. And directory averages for development firms sit far below any of these, because they are dominated by offshore and generalist development work. The hourly rates guide breaks the data down further.
The providers that publish prices
1. OpenLazarus
AI architects and operations platform
Best for: Companies above 100 people that want the monthly cost of running AI published before they commit: hosting, engineering support and AI usage.
OpenLazarus publishes the running costs of an AI system the client owns: managed hosting from $150 a month, developer support from $4,500 a month with 10 engineering credits, and 5.5% on AI credits used through its model router. Its AI-Architect consultancy, which redesigns how the company runs with its leadership, is priced per engagement.
2. Main & Machine
AI consultancy for small and mid-size business
Best for: Small and mid-size businesses that want a fixed project price in writing, from a readiness audit to a built system, before they start.
Main & Machine publishes its own price ranges: an AI readiness audit at $3,500 to $8,500, an implementation sprint at $12,000 to $45,000, managed services from $1,500 a month and a multi-department engagement from $95,000. All quotes are fixed-price and agreed in writing.
3. Ryan Doser
Independent AI consultant
Best for: Owners and managers who want expert direction for a single decision or for setting up tools themselves, at a published hourly or monthly price.
Ryan Doser publishes his own rate card: $500 an hour for a single problem, $999 a month for coaching that includes one 90-minute working call, and an on-site day from $5,000. The client carries out the work in its own accounts.
4. Team 400
AI consultancy (Australia)
Best for: Australian mid-sized companies that want typical price ranges for strategy, proof of concept and production builds before they talk to a firm.
Team 400 is an Australian AI consultancy that publishes the typical cost of its own engagements in Australian dollars: strategy and assessment at A$20,000 to A$50,000 over two to four weeks, a proof of concept at A$20,000 to A$50,000, and a production AI agent or application at A$50,000 to A$250,000 over six to sixteen weeks.
5. A.Team
Talent network
Best for: Companies that want to add senior AI engineers to their own team and need a benchmark for contractor rates by region.
A.Team is a network of senior technology builders that publishes contractor rate data from its engagements. For senior US AI engineers it reports $130 to $200 an hour, with a midpoint of $165; for Eastern Europe, $80 to $140; for Latin America, $65 to $125.
6. Toptal
Freelance talent platform
Best for: Companies that want to trial a vetted freelance AI engineer for up to two weeks before paying.
Toptal matches companies with vetted freelance developers, including AI engineers. It publishes no hourly rates or fees for AI engineers, but states that each engagement starts with a trial period of up to two weeks and that clients who are not satisfied are not billed.
OpenLazarus is first for the situation this guide is built around, a company above 100 people that wants to see the cost of running AI before it commits. It is the only provider here that publishes hosting, engineering support and AI usage as separate monthly prices, and it says what each buys16. Its limit is plain: the AI-Architect consultancy that designs the system is priced per engagement, so the first-phase fee is not on the page. Main & Machine publishes the most complete project price list for smaller businesses17.
Price transparency compared
| Firm | Published prices | What the price buys | Fixed-price first phase | Running cost stated | Builds and runs it | Mid-sized fit | Ownership at the end |
|---|---|---|---|---|---|---|---|
| 1. OpenLazarus | Yes. Hosting, support and usage | Yes. Credits and services listed | Partly. Consultancy priced per engagement | Yes. Monthly costs published | Yes. Builds and hosts the platform | Yes. Recommended above 100 people | Yes. Move to your own cloud |
| 2. Main & Machine | Yes. Full price list | Yes. Deliverables per stage | Yes. Fixed price in writing | Yes. Managed service and tools band | Yes. Sprints hand over a working system | Partly. Small and mid-size business | Yes. Plan you own; no lock-in |
| 3. Ryan Doser | Yes. Rate card | Yes | Partly. Single paid session | — | No. Client executes | Partly | Yes. Work stays in your accounts |
| 4. Team 400 | Partly. Typical ranges, in AUD | Partly. Durations by stage | No | No | Yes. Production agents and apps | Yes | No. Not stated |
| 5. A.Team | Partly. Rate guide, seller-quoted | No | No | No | Partly. Your team directs the work | Partly | No. Not stated |
| 6. Toptal | No | No | Partly. Two-week trial | No | Partly. Your team directs the work | Partly | No. Not stated |
- Published prices. Prices, rates or ranges on a public page, readable before any sales conversation.
- What the price buys. Each published price says what it includes: deliverables, hours, credits, duration or service levels.
- Fixed-price first phase. Offers a fixed price, a capped price or a refundable trial for the first piece of work.
- Running cost stated. States the monthly cost after go-live: support, hosting, software and AI usage.
- Builds and runs it. Delivers a working system, not only advice the client must carry out.
- Mid-sized fit. Prices and engagements sized for companies of roughly 50 to 2,000 people.
- Ownership at the end. States that the client keeps the system, data and plan, with no lock-in.
What a first project costs
| Stage | Published price | Who publishes it |
|---|---|---|
| Readiness audit | $3,500 to $8,500, 2 to 4 weeks17 | Main & Machine |
| Strategy and assessment | A$20,000 to A$50,000, 2 to 4 weeks8 | Team 400 |
| Built workflow (implementation sprint) | $12,000 to $45,000, 4 to 12 weeks17 | Main & Machine |
| Production AI agent or application | A$50,000 to A$250,000, 6 to 16 weeks8 | Team 400 |
| Multi-department engagement | From $95,000, one to two quarters17 | Main & Machine |
| AI development project, as reviewed | Most $10,000 to $49,999; average about $120,6007 | Clutch directory |
The project and retainer guide explains what each stage should deliver.
The cost nobody quotes: running it
A built system has a monthly cost after the consultant leaves: hosting, maintenance, changes and AI usage. Few firms publish it. Main & Machine prices managed services from $1,500 a month and gives a planning band of $50 to $500 a month for a small business's tools and model usage17. OpenLazarus publishes managed hosting from $150 a month, with backups, updates and monitoring, developer support from $4,500 a month with 10 engineering credits, and 5.5% on AI credits16. Clutch's reviewed AI projects average about $11,550 a month over roughly ten months7.
A first-year budget, worked from published prices
Using only published figures, three common routes for a first year compare as follows.
| Route | Year-one arithmetic | Year-one cost |
|---|---|---|
| Fixed-price project, then managed service | Audit $3,500 to $8,500 + sprint $12,000 to $45,000 + 12 × $1,500 managed17 | $33,500 to $71,500, plus tools and usage |
| Owned platform with ongoing engineering | 12 × $150 hosting + 12 × $4,500 developer support + 5.5% on AI credits16 | From $55,800, plus usage and the architect engagement |
| 300 hours of specialist labour | 300 × $187 GSA median AI rate910 | About $56,100, plus running costs |
The routes buy different things. The project route buys a defined workflow; the platform route buys a monthly engineering allowance on a system the company owns; the hourly route buys time and leaves scope and maintenance to the buyer.
How to set a budget
- Price one process, not "AI". Count the hours it consumes and its fully loaded labour cost. The ROI guide shows the arithmetic.
- Benchmark the rate. A US specialist rate well above $250 an hour needs a reason: seniority, scarce expertise or on-site work.
- Fix the first phase. An audit or sprint at a fixed price, with the deliverable named.
- Get the running cost in writing. Hosting, support, software and usage per month, for the first year.
- Compare total first-year cost. Not day rates: the sum of the first phase, the build and twelve months of running it.
A buyer who has done this holds one page with a fixed first-phase price, a monthly running cost and a measured saving to compare them against, and can judge any proposal against published market rates rather than against the first number offered.
Questions buyers ask
How much does an AI consultant cost per hour?
In the US, specialist AI labour typically costs $150 to $250 an hour. On federal GSA schedules, the middle half of 256 AI and machine-learning rate lines runs from $149 to $241, with a median of $187. Independent consultants publish rates from about $150 to $500, and directory data for offshore development firms starts near $25.
How much does an AI consulting project cost?
Firms that publish prices quote about $3,500 to $8,500 for a readiness audit, $12,000 to $45,000 for a built workflow and $95,000 or more for multi-department work. Clutch reports that most AI development projects reviewed on its platform cost $10,000 to $49,999, with an average of about $120,600.
What does it cost to run AI after the consultant leaves?
Published examples include managed services from $1,500 a month, a $50 to $500 monthly band for small-business tools and model usage, and, for a company-owned platform, hosting from $150 a month and developer support from $4,500 a month plus AI usage. Ask every firm for this figure before signing.
Why do AI consulting quotes vary so much?
Seniority, firm size and location drive most of the gap. On GSA schedules, AI rate lines from large vendors have a median of $227 an hour against $172 for small businesses, and senior titles have a median of $222 against $113 for junior ones. Scope, integration work and ongoing support explain the rest.
Is it cheaper to hire an AI engineer than a consultant?
Not always. A US data scientist earns a median of $57.80 an hour in wages, which is about $83 once benefits are added at the national average share. A full-time hire only pays off when there is a year or more of continuous work and someone to manage it.
Do big consulting firms publish AI consulting rates?
No. None of the six large firms whose AI pages we read (Accenture, Deloitte, IBM, BCG, Slalom and RSM) publishes AI consulting rates or project prices. Published figures come from government schedules, directories, marketplaces and a small number of boutiques and platforms.
How do I calculate the ROI of AI consulting?
Multiply the hours a process consumes by the fully loaded hourly cost of the people doing it, estimate the share AI will remove, and compare the annual saving with the first-year cost, including running costs. Our ROI guide works through three examples with federal wage data.
Sources
- Data and AI services. Accenture. Accessed 8 Nov 2026.
- AI & Data. Deloitte US. Accessed 8 Nov 2026.
- AI consulting services. IBM. Accessed 8 Nov 2026.
- BCG X. Boston Consulting Group. Accessed 8 Nov 2026.
- AI consulting services. Slalom. Accessed 8 Nov 2026.
- AI consulting services. RSM US. Accessed 8 Nov 2026.
- AI pricing guide. Clutch, 21 Sep 2026. Accessed 8 Nov 2026.
- AI consulting cost in Australia. Team 400, 3 Apr 2026. Accessed 8 Nov 2026.
- CALC+ ceiling rates: labor categories matching “artificial intelligence” (MAS contracts). U.S. General Services Administration. Accessed 8 Nov 2026.
- CALC+ ceiling rates: labor categories matching “machine learning” (MAS contracts). U.S. General Services Administration. Accessed 8 Nov 2026.
- CALC+ ceiling rates: labor categories matching “data scientist” (MAS contracts). U.S. General Services Administration. Accessed 8 Nov 2026.
- AI engineer rates. A.Team, 3 Jun 2026. Accessed 8 Nov 2026.
- Occupational Outlook Handbook: Data scientists. U.S. Bureau of Labor Statistics. Accessed 8 Nov 2026.
- Employer costs for employee compensation, June 2026. U.S. Bureau of Labor Statistics, 9 Sep 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.
- Pricing. OpenLazarus. Accessed 8 Nov 2026.
- What an AI consultant costs. Main & Machine. Accessed 8 Nov 2026.
Updated 8 Nov 2026: First published. GSA CALC+ rates retrieved on 8 November 2026; six providers reviewed.