If you've started looking into AI for your business, you've probably come across the term "AI consultant" and wondered what that actually means in practice. Is it someone who advises you on which tools to buy? Someone who builds things? Someone who just runs a workshop and leaves you with a slide deck?
Honestly, it can be any of those things, which is part of the problem. The term is loose enough that almost anyone can use it. So this post is about what a good AI consultant actually does, what they shouldn't do, and how to tell the difference before you sign anything.
The honest answer — what a good AI consultant actually does
A good AI consultant does four things: they audit your operations to find where automation will actually make a difference, they design a strategy for how to build it, they oversee or carry out the build itself, and they hand over something your team can actually use and maintain.
That's it. If an engagement doesn't touch all four of those things, it's probably incomplete.
Here's what each of those looks like in practice.
Discovery and process audit
This is where any decent engagement starts. Before anything gets built, you need to understand where the actual drag is in the business, which processes are eating hours, which ones have clear inputs and outputs, and which ones are genuinely automatable versus ones that just feel painful.
At AMPL, we spend the first part of every engagement mapping this out. That means talking to the people actually doing the work, not just the director who booked the call. It means looking at the tools already in place and understanding why they're not talking to each other. It means getting specific. Not "we have a problem with onboarding" but "this specific step takes three people four hours every Monday and it's entirely manual."
The audit should produce something concrete: a clear picture of where automation will deliver real ROI, and where it won't. That second part matters. A consultant who tells you everything is automatable isn't being honest.
Tool selection and build strategy
Once you know what needs solving, you need a plan for how to solve it. This is where tool selection happens, and it's where a lot of consultants go wrong. More on that in a moment.
Good tool selection is problem-first. You look at what the process requires, then find the right combination of tools to deliver it. Sometimes that's a simple workflow automation. Sometimes it needs a custom-built AI agent. Sometimes the honest answer is that an existing SaaS tool does what you need and you don't need a bespoke build at all.
The build strategy should include what gets built, in what order, how it integrates with existing systems, and what success looks like. Not in vague terms, in specific, measurable ones. "This process currently takes 6 hours per week. After the build, it should take under 30 minutes."
Implementation and integration
Depending on the engagement, the consultant either builds it or manages someone who does. The distinction matters. Some consultants are strategists who hand off to developers, others do the technical work themselves.
Neither model is wrong, but you should know which one you're buying before you start. A strategic consultant who oversees a developer can work well. A strategic consultant who doesn't have a clear path to implementation is just expensive advice.
At AMPL, we build directly using Claude Code, custom AI systems rather than patched-together no-code tools. That matters for complex operations where off-the-shelf automation hits its ceiling quickly. Integration with your existing stack, your CRM, your inbox, your project management tools, is part of the build, not an afterthought.
Handover, documentation, and ongoing support
This is the bit that separates a consultant from a contractor who disappears. A good engagement ends with your team actually understanding what was built, why it works the way it does, and what to do when something breaks.
That means documentation. It means a handover session. It means a support structure so that a month down the line, when something changes in your process, you're not stuck.
The goal isn't dependency. It's capability. A good AI consultant leaves your business in a better position than they found it, including your team's ability to operate what was built.
What an AI consultant should NOT do (red flags)
The AI consulting space is still young enough that there are no established standards. That means the range of what gets sold as "AI consulting" is enormous, and some of it is genuinely bad for clients. Here's what to watch for.
Selling you a tool they're partnered with
Some consultants have referral arrangements or reseller agreements with specific platforms. That's not inherently wrong, but it becomes a problem when their tool recommendation is shaped by commission rather than your actual needs.
The tell is usually in how the recommendation happens. If a consultant names their preferred tool before they've properly understood your operations, something's off. Good tool selection comes after the audit, not before it. If they're pushing you toward a specific platform on the first call, ask whether they have a commercial relationship with that vendor.
Starting to build before understanding your operations
This one is surprisingly common. A consultant comes in, gets excited about the technical problem, and starts building before they've properly mapped out the business context. What you end up with is usually technically impressive and operationally useless. It solves the problem they found interesting, not the one that was actually costing you money.
Any build that starts before a proper discovery phase is a risk. The business context has to come first. Always.
Promising outcomes they can't guarantee
This one matters a lot in AI specifically. Large language models are probabilistic. They're not deterministic software that does exactly the same thing every time. A consultant who promises "100% accuracy" on any AI output, or guarantees specific business results without caveats, either doesn't understand how the technology works or is telling you what you want to hear.
Honest consultants talk in terms of significant reduction, measurable improvement, and clear ranges, not guarantees. They also talk about what the system won't do well, and where human oversight is still needed. That kind of honesty is a signal of competence, not weakness.
AI consultant vs AI developer — what's the difference?
This comes up a lot. People want to know whether they need a consultant or just a developer who knows AI.
The simplest way to think about it: a developer can build what you specify. A consultant helps you figure out what to specify.
If you know exactly what you want, the process is clear, the requirements are documented, and you just need someone technical to build it, a developer might be the right hire. If you're not sure what you need, where to start, or whether AI is even the right solution for your specific situation, that's where a consultant earns their fee.
In practice, the best AI consultants in this space do both. They have enough technical depth to make sound build decisions, and enough business sense to tie those decisions back to actual outcomes. A purely strategic consultant who can't assess what's technically feasible is guessing. A purely technical developer who doesn't understand your operations will build the wrong thing perfectly.
For most businesses looking at their first serious AI build, you want someone who can do both, or at minimum, someone who can bridge the gap between your operations team and the people doing the technical work.
How to evaluate whether you need a consultant or an internal hire
This is worth thinking through honestly before you commit to either.
An internal hire makes sense when you have ongoing, high-volume AI work that justifies a full-time role. If you're a large operation that's planning to build multiple systems over multiple years, having someone in-house who understands your business deeply can be the right call.
A consultant makes more sense in a few specific situations. First, when you're at the start and don't yet know what you need. The audit and strategy work is exactly what consultants are for. Second, when the build is discrete, a specific system or set of systems with a clear scope, after which the maintenance load isn't heavy enough to justify a hire. Third, when speed matters. A good consultant can move faster than a hiring process.
The honest answer for most businesses in the 10-100 person range: start with a consultant to build the foundation, then decide whether ongoing work justifies an internal hire. Don't hire internally before you know what you're hiring for.
What to expect at each stage of an engagement
If you're thinking about working with an AI consultant, here's a rough sense of what a well-structured engagement looks like from your side.
Stage 1, the audit (weeks 1-2): Expect a lot of questions. You and your team will be asked to walk through specific processes in detail. The consultant should be mapping what's manual, what's repetitive, what has clear decision rules, and what's actually costing you. You'll get back a written assessment, not a vague report, but specific recommendations with rationale.
Stage 2, strategy and sign-off (week 2-3): Based on the audit, you'll see a proposed build plan. This should include what gets built, in what order, what it integrates with, what success looks like, and what the costs and timelines are. This is the point where you decide what to prioritise and what to park.
Stage 3, build (weeks 3-8 typically, varies by scope): The actual construction of the system. You should have visibility into progress, not necessarily into every line of code, but into whether things are on track. Expect to be involved in testing. Real-world testing with your actual data and processes is the only way to know if something works.
Stage 4, handover (final week): Documentation, training, and a clear understanding of what to do when something needs changing. If a consultant tries to skip this stage, that's a red flag. It usually means they want you to remain dependent on them for ongoing changes.
Ongoing support: A good consultant offers a support structure after handover. That might be a retainer, ad-hoc access, or a defined support window. What it shouldn't be is silence.
FAQ — People Also Ask
What does an AI consultant charge?
Rates vary significantly. Independent consultants might charge anywhere from £500 to £2,000+ per day depending on experience and specialisation. Project-based fees for a full build, audit through to handover, typically range from £5,000 to £50,000+ depending on complexity. Be wary of very low quotes on complex builds; they usually mean shortcuts somewhere.
Do I need an AI consultant or can I just use off-the-shelf tools?
If your processes are simple and your requirements are standard, off-the-shelf tools like Zapier or Make might be enough. Where consultants add value is when your operations have specific logic, multiple integrations, or enough complexity that generic tools hit their ceiling. The honest answer: try the simple tools first, and bring in a consultant when they're not enough.
What are the red flags when hiring an AI consultant?
Key warning signs: they recommend a specific tool before understanding your operations; they promise specific accuracy rates on AI outputs; they can't explain their build approach in plain English; they have no clear process for discovery before building; they're vague about what success looks like. Any consultant who can't give you a clear answer to "what will be different after this engagement?" isn't ready to take your money.
How long does an AI consulting engagement take?
A basic engagement, audit through to a working first build, typically takes 6-12 weeks depending on scope and your team's availability for input. More complex builds with multiple integrations take longer. Be sceptical of consultants who promise complex builds in very short timeframes; they're either underscoping or overpromising.
What's the difference between an AI consultant and an AI agency?
An AI consultant is usually an individual or small team with hands-on technical and strategic capability. An agency typically has more resource but more overhead, and you may find yourself working with junior staff rather than the person who sold you the engagement. Neither is better by default, it depends on the scope of what you need.
Can an AI consultant guarantee ROI?
No, and any consultant who claims they can is being misleading. What a good consultant can do is identify processes with clear, measurable time and cost implications, and build systems that demonstrably reduce that burden. The ROI is calculable and real, but it depends on your operations, your team's adoption, and how the systems are maintained over time.
If you're trying to work out whether AI is right for your business, or whether what you've been told by a consultant so far sounds right, the audit is usually the best place to start. It gives you a clear picture without committing to a full build. If you want to talk it through, book a free audit at amplconsulting.ai.
