AI consultant vs AI implementation: which does your firm need?
An AI strategy consultant tells you where AI fits and what to do first. AI implementation builds it into your work, rolls it out and keeps it running. Larger firms facing risky decisions benefit from strategy first. Most small firms already know where the hours go; for them, the roadmap is the easy part and implementation is the work.
| AI strategy consultant | AI implementation (build only) | AI implementation on a Forge monthly plan | |
|---|---|---|---|
| Typical cost, US | $75-150 an hour freelance; $125-250 boutique; $250-500 enterprise firm | $5,000-15,000 for a single-workflow first version | $2,000 / $3,500 / $7,500 a month, first build included |
| Typical cost, UK | Median AI consultant contract day rate £564 | Usually quoted per project | £1,600 / £2,800 / £6,000 a month |
| Speed to start | Days to weeks; a discovery phase often takes a few weeks | Weeks of scoping, then the build | Two-week ops audit, first build live in 2-6 weeks |
| Who maintains it | Nothing to maintain - you get a plan | You, after hand-off | We do, as part of the plan |
| Best for | Larger firms with many options, regulated work, or a board to persuade | One clear AI workflow with an owner afterwards | Small firms that know where the hours go and need it working |
| Worst for | Small firms that already know their problem | Firms with no one to look after it | Firms that need a board-level AI strategy |
US ranges are typical, compiled from published rate guides (Layer3 Labs, September 2026). The UK figure is the median AI consultant contract day rate for the six months to October 2026 (IT Jobs Watch, October 2026). Our prices are from our pricing page.
Two different jobs with similar names
"AI consultant" can mean almost anything in 2026. Here is the split that matters.
AI strategy answers: where could AI help us, what should we do first, what should we buy, what are the risks, and what will it cost? The output is a plan, often called a roadmap. Good strategy work includes interviews, a look at your data and tools, and a prioritised list of opportunities.
AI implementation answers: how do we make this work in our business, on Monday? The output is a working tool. Documents read and sorted, first drafts of reports ready to review, notes turned into actions, apps kept in step. Then the part that is easy to forget: rolling it out to the team, training people, and keeping it running when the AI models and your apps change.
Many firms offer both. The question is which one you need to pay for first.
Why the roadmap is rarely the hard part
For a people business - a business that sells its people's expertise and time, such as an agency, accountant, law firm or consultancy - the list of where AI could help is usually short and obvious. Ask the team what they copy, chase, rekey and rebuild by hand, and you have most of the roadmap in an afternoon.
The hard parts come after:
Connecting AI to the apps you already use,
so it reads the right information and puts its output where people work.
Getting it right often enough to trust.
AI is confidently wrong sometimes. Every tool needs a person checking the output before it reaches a client. AI drafts, people decide.
Rolling it out.
A tool nobody uses gives no hours back. People need to see it, try it and trust it - and hear clearly that it is there to make their work better, not to make them redundant.
Keeping it running.
Models change, prices change, apps change. Somebody has to watch it.
That is why a beautiful roadmap can sit in a drawer. The plan was never the bottleneck.
What AI gets wrong, and how to plan for it
Whichever route you choose, go in with clear eyes about what AI is like at work. It is very good at first drafts, summaries, sorting and pulling details out of documents. It is also confidently wrong sometimes: a figure misread, a name swapped, a sentence that sounds right and is not. That does not make it useless. It means every tool should be built with a person checking the output before it matters.
In practice that looks like a review step before anything reaches a client, a clear way for people to correct the AI, and a habit of checking a sample of results each week. It also means choosing the right model for each job. Prices vary a lot between models, so a simple sorting task should not run on the most expensive one.
A strategy consultant should flag these risks in the plan. An implementation partner should build the checks in from the first version. If you are comparing proposals, ask each one: "Where does a person check the AI's work, and what happens when it is wrong?"
Who should hire an AI strategy consultant
We want to be fair here, because strategy work is genuinely valuable in the right firm. Hire a strategy consultant first if:
You are a larger firm with many teams,
and choosing where to start is itself a hard decision.
You work in a regulated area
legal, financial, health - and need careful thinking about data, confidentiality and risk before anything is built.
You need to persuade a board or partners,
and a credible outside view will move the decision.
You are choosing a platform
that will shape the next several years, such as a firm-wide AI assistant or a major vendor contract.
Your leadership disagrees about AI,
and you need a neutral person to get everyone to one plan.
In those cases, a few weeks of strategy before building saves money and arguments. A discovery or audit phase typically costs $2,000-5,000 (Layer3 Labs, September 2026); larger firm-wide strategy engagements cost more. In the UK, the median AI consultant contract day rate is £564 (IT Jobs Watch, October 2026). Our AI consultant cost guide has more detail.
Who should skip straight to implementation
You are a 10-60 person firm
and the owner or ops lead can name the three jobs that eat the most time.
The opportunities are about hours,
not about launching an AI product.
You have no one to build or look after it
in-house.
You have read enough about AI.
You want one thing working, then the next.
For these firms, a short audit that ends in priced builds is the right amount of strategy. Anything longer delays the hours back.
Pick an AI strategy consultant if...
- The decisions are big, risky or political.
- You need a document that a board or partners will act on.
- You have, or will hire, people to implement the plan afterwards.
Pick a one-off AI implementation if...
- You have one clear workflow in mind, like sorting incoming documents or drafting a weekly report.
- Someone on your team can own it after the build.
- You accept that changes and upkeep will be priced as they come.
A single-workflow first version typically costs $5,000-15,000, and $15,000-25,000 if it has to talk to an older system (Layer3 Labs, September 2026).
Pick implementation on a monthly plan if...
- You want several pieces of AI help built over time, not just one.
- You want them rolled out properly, with the team trained, so people actually use them.
- You want someone watching them and fixing them when models or apps change.
That is our model, in four steps: scope, build, roll out, improve. A two-week ops audit ($2,500 / £2,000) is our strategy phase: we interview your team, map your apps, count the hours lost and scope your first three builds with prices. It is credited to your first month. Monthly plans are $2,000, $3,500 or $7,500 (£1,600 / £2,800 / £6,000), with your first build included and a 3-month minimum. AI usage is billed at cost, with no mark-up. The software runs on your own accounts, so you own it. More on our AI implementation consultant page.
Our recommendation: how we'd decide for a 20-person firm
- Spend one hour listing the manual work. Ask each team lead: what do your people copy, chase, rekey or rebuild each week? If you get a clear list, you have most of a roadmap already.
- Check for the strategy triggers. Regulated data, a board to persuade, a platform choice, or a leadership team that disagrees. If any of those apply, hire a strategy consultant for a short, fixed engagement first.
- If none apply, buy a short audit, not a long strategy. Two weeks, ending in priced builds. That can be ours or anyone's; the point is that it ends in something you can build.
- Insist on a person checking the output. Whoever builds it, AI output that reaches a client should pass a human first.
- Plan the rollout before the build. Who will use it, who trains them, and how you will know it is being used.
If a firm needs a board-ready AI strategy, we will say so and suggest they hire a strategist before talking to us.
Worked example: automated client reports, strategy vs implementation
Take one real job: automated client reports, a live page per client showing where every piece of work is, so nobody writes the Friday update. Our custom software cost guide puts the typical market cost of a dashboard like this at $8,000-18,000 in the US or £6,000-14,000 in the UK, taking 2-6 weeks (typical, compiled from published rate guides). At Fame, account managers retired the Friday status email, and clients check the link before they message.
With an AI strategy consultant
A discovery phase in the $2,000-5,000 range would likely identify client reporting as a priority, recommend reading progress straight from your project tool, and perhaps suggest AI to draft the summary paragraph for a person to check. Useful - and you still need someone to build it.
With a one-off implementation
A build in the $8,000-18,000 (£6,000-14,000) range, in 2-6 weeks. Then you own the upkeep when your project tool or the AI model changes.
With Forge
The ops audit ($2,500 / £2,000) covers the strategy step for this firm size and is credited to month one. The reports are the first build, included. On the $2,000 / £1,600 plan, the 3-month minimum is $6,000 / £4,800 in total, including rollout to your account managers, and we look after it after that. If you only ever need this one report and have someone to maintain it, a one-off build may cost less over time.
Across all of Fame's tools, the team got back 5,460 hours as of 14 October 2026.
Questions, answered plainly
What is an AI implementation consultant?
Someone who builds AI into your existing work - connecting it to your apps, setting up human checks, rolling it out to your team and keeping it running. An AI strategy consultant, by contrast, produces a plan of where AI fits and what to do first.
What is an AI strategy?
A plan for where AI will help your business, what to do first, what to buy or build, the risks, and the costs. For a small firm it can be short; for a large or regulated one it deserves real time.
What is an AI implementation?
Putting an AI tool to work in a real process: connected to the apps you use, with a person checking the output, rolled out to the team and kept running as models and apps change.
How much do AI implementation consultants make?
Typical US rates run $75-150 an hour for freelancers and $125-250 for boutique firms, with single-workflow builds at $5,000-15,000. In the UK, the median AI consultant contract day rate is £564.
Should a small business hire an AI strategy consultant or go straight to implementation?
Most 10-60 person firms can go straight to a short audit and implementation, because the list of manual work is usually clear. Hire a strategist first if you are regulated, need to persuade a board, or face a big platform choice.