The short version: The AI consultancy space is flooded with people who watched a few YouTube videos and printed business cards. A real AI consultant shows you specific prior results, speaks plainly about limitations, and charges in line with genuine expertise. A fake one hides behind jargon, cannot name a single tool they have used in production, and vanishes when you ask for references.
I have been an AI and marketing consultant since before it was fashionable to call yourself one. I have also hired the wrong people, been burned by over-promising vendors, and sat across from "AI experts" who could not explain what a large language model does. So this is not a theoretical post. This is what I have learned the hard way.
The market right now is chaotic. Anyone can put "AI Consultant" in their LinkedIn headline. There is no certification that means anything yet. There is no governing body. That means the due diligence is entirely on you.
Why this problem is bigger than you think
In 2026, the global AI consulting market is worth tens of billions of pounds. The demand for people who can help businesses implement AI tools practically has exploded. And wherever demand spikes faster than supply, fakes fill the gap.
I am not talking about junior consultants who are still learning. That is fine, as long as they are honest about it and priced accordingly. I am talking about people charging senior consultant day rates of £1,500 to £3,000 a day while having done nothing more than prompt ChatGPT a few times and read a newsletter.
I have seen companies spend £40,000 on an "AI strategy" that was a 30-slide deck full of buzzwords and no implementation plan. I have seen marketing teams handed "AI workflows" that were just screenshots of free tools with arrows between them. This is real money going into a black hole.
The jargon test: ask them to explain it simply
The single fastest way to separate a real AI consultant from a fake one is this: ask them to explain something technical in plain English.
Try one of these:
- "Can you explain what retrieval-augmented generation is, without jargon, in two minutes?"
- "What is the actual difference between fine-tuning a model and prompting it, and when would you choose one over the other?"
- "What does hallucination mean in this context and how do you build a workflow that catches it?"
A real consultant will give you a clear, grounded answer. They might say something like: "RAG is when you give the model access to your own documents at the moment it answers, rather than baking the knowledge into the model itself. You use it when your information changes frequently, like a product catalogue or a policy database." That is useful. That is someone who has built something.
A fake consultant will start with "great question" and then spiral into words like "synergistic AI integration" and "holistic data ecosystems" without ever landing on a concrete explanation. Watch for the pivot. When someone cannot answer plainly, they pivot to complexity. It is a defence mechanism.
The portfolio problem
Here is the honest point most articles will not make: a lot of AI consultants right now cannot show you a portfolio, not because they are fake, but because their clients have NDAs that cover everything. I have been in that position myself. Some of my best work exists only in private Slack channels and internal dashboards.
So "I cannot share specifics due to client confidentiality" is not automatically a red flag. But here is what distinguishes the real consultant in that situation: they can still describe the problem type, the approach, and the measurable outcome without naming names.
"I worked with a B2B SaaS company, they were spending about 12 hours a week on customer support triage. We built an AI routing system using their existing help desk tool. After six weeks, that number was down to three hours. The main challenge was getting the categorisation logic right because their product had a lot of edge cases."
That is a real story. You can feel the friction in it. The fake version sounds like this: "I have worked with numerous enterprises to transform their AI capabilities and deliver ROI-driven outcomes." No friction, no specifics, no numbers. Run.
A story from my own inbox
About eighteen months ago, a woman got in touch with me wanting to collaborate. She called herself an AI consultant and had a polished website, good photography, and a newsletter with several thousand subscribers. She name-dropped three global brands as past clients.
I was curious, so I had a call. Within ten minutes I asked her which AI tools she had used in production for clients, meaning deployed, not just demoed. She named two tools. I asked what the outputs were used for. She said "content creation and productivity." I asked which specific workflows. She paused, then said she "preferred to keep her methodology proprietary."
I asked about the three global brand clients on her website. She admitted that one was a speaking engagement, one was a workshop she had attended as a delegate (not as a speaker), and the third was an introductory call that never became a project. None of it was fabricated, exactly, but all of it was framed to imply work that did not exist.
I politely ended the conversation. She now charges £2,000 a day according to her website. I do not know who is hiring her.
This is why I always say that vetting an AI consultant is not unlike spotting fake reviews before you buy. The signals are similar: polished surface, missing specifics, social proof that does not quite hold up when you look closely.
How to check someone's real experience: a step-by-step process
Do not rely on their website alone. Here is what I do when I am considering bringing in an external AI consultant:
Step 1: Google their name plus the tool names they claim to use
If someone says they are an expert in building AI agents with a specific platform, their name should appear somewhere in the wild. A forum post, a LinkedIn comment, a case study, a podcast interview, something. If you find nothing except their own website, that is a signal worth noting.
Step 2: Ask for a 20-minute working session, not a pitch
Tell them you want to spend 20 minutes working through a real (small) problem you have right now. Not a discovery call. Not a presentation of their services. An actual problem. How they respond tells you everything. A confident, experienced consultant will say yes without hesitation. A fake one will try to redirect you toward a proposal or a paid audit first.
Step 3: Call their references, do not email them
Email is easy to dodge. A 10-minute phone call with a reference is harder to fake. Ask the reference: "What was the hardest moment in the project and how did they handle it?" Easy, positive answers can be rehearsed. A reference who thinks for a moment and says "honestly, we hit a wall at week three with the data quality and they had to rebuild part of the approach, but they were transparent about it" is a real reference.
Step 4: Ask about a failure
Directly ask the consultant: "Tell me about an AI project that did not go as planned and what you learned from it." Real consultants have war stories. They remember the client who had unusable data. They remember the workflow that broke in production. They can laugh about it now because they learned from it. Fake consultants will say they have always delivered strong results or talk vaguely about "challenges" without any specific content.
Step 5: Check the pricing against the market rate
In the UK in 2026, experienced AI consultants with a genuine track record charge roughly £1,000 to £2,500 per day for implementation work, and £3,000 to £5,000 for senior strategic advisory. If someone is charging £5,000 a day but has no verifiable track record, that is a red flag in the other direction: overpricing without substance. If someone is charging £200 a day and calling themselves a senior AI strategist, ask hard questions too. Rate alone is not a verdict, but extreme outliers in either direction deserve scrutiny.
The difference between an AI consultant and an AI tool reseller
This one catches a lot of businesses out. There is a whole category of people who call themselves AI consultants but whose actual business model is recommending the same two or three tools to every single client, often with a referral arrangement in place.
The tell is that their "strategy" always reaches the same conclusion regardless of your situation. Whatever your problem, the answer is the same platform. A real AI consultant will sometimes tell you that you do not need a new tool at all. I have told clients that their problem was a process problem, not a technology problem, and that buying more AI software would make things worse, not better. That is not a popular thing to say, but it is often the true thing.
If you want to understand the structural difference between hiring an independent consultant versus going to a large agency that bundles AI into a retainer, I wrote about this in detail when comparing AI marketing consultants against agencies and in-house teams. The short version: an independent consultant's incentives are aligned with your outcome, not with selling you their proprietary tech stack.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Red flags in the sales process specifically
The sales process itself is one of the richest sources of data about what working with someone will be like. Watch for these:
- They ask for a large upfront payment before doing any scoping. A real consultant will do a discovery phase first. They cannot price something they do not understand yet.
- They promise specific percentage improvements before seeing your data. "We will reduce your content production time by 60%" is not a promise anyone can make without knowing your current setup.
- They refuse to put anything in writing until you sign. Legitimate consultants have no problem sending a clear proposal with deliverables, timescales, and success metrics before you commit.
- They use urgency tactics. "I only have one slot left this quarter" from someone you just met online is almost always manufactured scarcity.
- They cannot name a single limitation of AI for your use case. This is a big one. Real expertise includes knowing where the technology breaks down. If someone tells you AI will solve everything without caveat, they are selling, not consulting.
What good AI consultants talk about
When I am doing good work with a client, the conversations are not glamorous. We are talking about data quality. We are talking about where a workflow will break if a team member goes on leave. We are talking about how to monitor outputs so that errors do not compound. We are talking about change management and why the team is resistant to the new tool even though it saves them time.
I have written honestly about this in the context of my own client work, including the two times I got AI client onboarding wrong before I got it right. The second failure was embarrassing. I had automated too much, too fast, without testing the edge cases. A client fell through the gap. That kind of honesty is part of what real expertise looks like in practice.
If your prospective consultant only talks about wins and possibilities in every conversation, something is missing.
The social proof trap
Follower counts mean nothing in this context. I spent years building a large social following during my influencer days. I know exactly how disconnected audience size can be from actual expertise. Someone with 50,000 LinkedIn followers who posts daily AI hot takes may have never deployed a working AI system in their life.
This is especially relevant if you found this consultant through social media. A big following is a marketing asset, not a consulting credential. What you want to know is: what did they build, for whom, and what happened? If the only content they produce is "AI tips" and "hot takes" and "threads about prompting," and you cannot find a single piece of writing about a real project they ran, weigh that accordingly.
On that note, I have written about how most businesses get social media backwards, and the same trap applies to evaluating consultants through their social presence. Visibility is not the same as credibility.
The honest truth about the market right now
Here is where I will say something bluntly that most people writing about this topic will dance around: a significant proportion of people currently calling themselves AI consultants are not qualified to charge what they are charging. This is not because AI consulting requires years of academic training. It does not. But it requires having built things, having failed, having iterated, and having solved real problems for real clients.
The ecosystem has also attracted a lot of people who are brilliant at marketing themselves and less brilliant at the work. That is not unique to AI. It happens in SEO, in social media management, in business coaching. But in AI specifically, the gap between the pitch and the delivery can be especially costly because the technology is moving so fast that clients often cannot tell whether the work is good or not. They trusted the consultant to know. That trust is sometimes misplaced.
If you are exploring this space as someone looking to position yourself as a consultant rather than hire one, the most useful thing I can point you to is my honest guide to finding real freelance work that pays, which covers how to build credibility in a crowded market without faking it.
What to do if you have already hired the wrong one
It happens. I have been in this position as a client and as someone who has watched it happen to others. The first step is to stop paying before you get clarity on what you have received. Review the contract for deliverables and check what has been delivered against that list. If the deliverables are vague in the contract, that is useful information about the original red flag you may have missed.
Most fake consultants will walk away from a calm, documented request for accountability rather than fight it out. The ones who push back hard usually have enough confidence in their work to engage with the substance. Document everything in writing before any conversation happens. If you paid a significant amount and received nothing of value, a solicitor's letter is sometimes worth the cost as a first step.
More importantly: do not let a bad experience put you off AI consultancy altogether. The real ones are out there. They are often quieter online than the fakes, because they are busy doing the work.
Frequently asked questions
What qualifications should a real AI consultant have?
There is no single certification that validates an AI consultant in 2026. What matters is verifiable project experience: tools used in production, measurable outcomes for clients, and the ability to explain technical concepts plainly. Google's AI certifications and completion of structured courses can signal foundational knowledge, but they are not a substitute for having built and deployed working AI systems for paying clients.
How much should a good AI consultant charge?
In the UK in 2026, experienced AI consultants with a real track record typically charge between £1,000 and £2,500 per day for hands-on implementation work, and £3,000 to £5,000 per day for senior strategic advisory. US rates run roughly 20 to 30 percent higher. Anyone charging significantly above those ranges without a clear, verifiable portfolio deserves extra scrutiny. Anyone charging significantly below them for "senior" work may lack the experience the title implies.
Is it a red flag if an AI consultant cannot share client names?
Not on its own. NDAs are standard in consulting. The real question is whether they can describe past problems, approaches, and results in enough specific detail that the project feels real. Vague claims of "enterprise transformation" with no specifics are the red flag. A detailed story about a specific challenge, even without naming the client, signals genuine experience.
How do I know if an AI consultant's advice is good?
A reliable test: ask them to tell you something AI cannot do well for your specific use case. A knowledgeable consultant will have a clear, specific answer based on the limitations of current models or the realities of your data situation. Someone who responds only with possibilities and never with limitations is optimising for the sale, not for your success.
Want this done for you? See AI workflows that save small businesses hours every week.
Related reading: Earning Passive Income: The Honest Guide From Someone Who Has Done It, Lost It, and Built It Back and Make Passive Income: The Honest Guide to What Works, What Pays, and What Most People Get Wrong.
For the bigger picture, see my full guide to AI marketing.
Once you know what to look for, the next step is finding the right fit. Here is how an AI consultant for small businesses can help you put this into practice.
Budget is part of the decision too. Here is a realistic look at how much an AI consultant costs and what drives the price.
Once you can spot the fakes, the next question is fit: what an AI consultant for a small business actually does and when you need one.
