Why the Wrong Consultant Costs More Than No Consultant
A bad AI consulting engagement does not look like a dramatic failure. It looks like a slow, expensive disappointment. Months pass, money gets spent, some things get built, and at the end of it the business is not in a meaningfully better position than it was at the start. Sometimes it is in a worse one � with systems that sort-of-work requiring ongoing maintenance, team habits that are now built around a flawed implementation, and a founder who is more sceptical of AI consulting than they were before.
The cost of a poor-fit engagement is real and often underestimated when the selection decision is being made. Getting the evaluation right is not bureaucratic due diligence. It is the work that determines whether the investment produces results.
The Credentials That Matter vs. the Ones That Do Not
The AI consulting space has no standardised credentialing. Anyone can call themselves an AI consultant, and many people do. This means credentials function as signals rather than guarantees � some are meaningful and some are noise.
What matters: a clear portfolio of work with businesses at a comparable scale and operational complexity. Not case studies described in the abstract, but specific examples of what was built, what problem it addressed, and what the business looks like now relative to before. The ability to describe their process for a typical engagement in specific terms. A clear answer to the question of what you will own at the end and who will maintain it.
What matters less than it appears to: certifications from AI vendors and platforms. Degrees in AI, machine learning, or data science (unless the work involves those disciplines specifically, which small business operational consulting rarely does). The sophistication of the tools they use. The impressiveness of the client logos in their marketing materials.
A consultant who has done excellent operational AI work for businesses that look like yours is more valuable than a consultant with an impressive academic background and experience that does not transfer to your context.
Small Business vs. Enterprise Experience: Why It Is Not Interchangeable
This distinction matters more than most founders recognise when they are evaluating consultants.
Enterprise AI implementation involves large teams, long timelines, complex governance structures, significant budget, and often specialised technical infrastructure. It is slow by design and involves layers of stakeholder management that do not exist in a small business context.
Small business operational AI consulting requires moving faster, working with tighter budgets, operating with less technical infrastructure, and producing results that are visible and measurable within weeks rather than quarters. The person responsible needs to be comfortable as both the strategist and the implementer, because there is no large team to hand work to.
A consultant who has only worked in enterprise environments often brings assumptions that do not serve a small business well. They scope more extensively than necessary. They recommend tools that are built for larger teams. They move at an enterprise pace when the client needs results in eight weeks. They may produce excellent work in their native environment and struggle to calibrate for a context that operates very differently.
Ask specifically: what is the smallest business you have worked with, and what does that engagement look like compared to a larger one? The answer tells you a great deal.
What to Look for in Their Process
The single most revealing thing about a consultant is whether they start with your problem or their tools.
A consultant who opens the first conversation by describing what they typically build, the tools they prefer, or the methodologies they apply is showing you how they orient. The focus is on their capabilities. Your situation will get fitted to their framework.
A consultant who opens by asking detailed questions about how your business operates, where the friction is, what you have already tried, and what a successful outcome looks like is showing you something different. The focus is on your specific situation. The framework will be built around it.
Questions that reveal process orientation:
How do you start an engagement? What does the first two weeks look like?
How do you decide which workflows to address first?
What happens when you discover mid-engagement that the original scope was wrong?
How do you handle situations where the client team is not adopting what was built?
What does handoff look like, and what specifically will I own at the end?
The quality and specificity of the answers to these questions tells you more than any portfolio piece.
Red Flags That Should End the Conversation
These patterns are reliable indicators of a poor-fit engagement.
Vague scope with confident pricing. A fixed-price quote produced before meaningful discovery has happened is either based on guessing or on a standardised scope that will be applied to your situation regardless of fit. Neither is good.
No plan for your independence. If the conversation about what you will own at the end produces a vague or evasive answer, the engagement may be designed � intentionally or not � to create ongoing dependency rather than client capability.
Tool vendor relationships that affect recommendations. A consultant who receives referral fees or reseller revenue from specific tools has a financial interest in recommending those tools that is separate from your operational needs. Ask directly whether they have financial relationships with any vendors they might recommend.
Only enterprise case studies. If the consultant’s track record is entirely in large organisations and they are positioning themselves for small business work, the burden of proof is on them to explain why the skills transfer. Marketing materials that feature logos from large companies and conversations that reference how things work “at scale” are not promising signals for a fifteen-person service firm.
Dismissiveness about adoption and change management. Consultants who treat training and team adoption as afterthoughts � things that happen after the real work is done � consistently produce low-adoption implementations. The technical build and the human adoption work are equally important. A good consultant knows this and says so.
How to Evaluate Proposals and Pricing
When you receive a proposal, evaluate it on three dimensions.
Scope clarity. Can you tell specifically what will be built, by when, and what is explicitly not included? A good proposal names the workflows being addressed, the tools involved, the phases and timeline, and the handoff deliverables. Ambiguity in a proposal becomes a dispute during the engagement.
What is and is not included. Proposals that include implementation but not training, or training but not documentation, or build but not adoption support are scoping these components out for a reason. The engagement will cost more to complete properly than the proposal suggests, either in additional fees or in outcomes that fall short because the full process was not funded.
Alignment between your problem and their solution. The proposal should read like a response to the specific situation you described, not like a standardised document with your name inserted at the top. If you could plausibly imagine them sending the same proposal to any other small business, they probably did.
The Final Decision
After the evaluation process, the decision usually comes down to one thing: do you believe this person understands your business well enough to improve it, and do you trust them to tell you when something is not working rather than protecting the relationship?
The best AI consulting relationships involve honesty in both directions. You tell them what you are actually seeing in your operations. They tell you when your expectations are not realistic, when a workflow is not ready to automate, or when the problem you described is not the one that is actually driving the friction.
That kind of honest working relationship is not visible in a portfolio. It is visible in a direct conversation. If the first substantive conversation leaves you feeling like you were heard and the responses were honest, that is worth more than an impressive case study.
Part of the Working with an AI Consultant series.
Related reading: Questions to Ask Before Hiring an AI Consultant | What Does an AI Consultant Actually Do? | Are You Ready to Hire an AI Consultant?
