The Market Has a Credibility Problem
The AI consulting market has expanded faster than the supply of people who genuinely know what they are doing. Every management consultancy now has an AI practice. Every software vendor has repackaged themselves as an AI partner. Every boutique firm founded in the last two years leads with AI transformation on their homepage.
This creates a real problem for procurement officers and decision-makers trying to evaluate options. The pitches all sound similar. The case studies are carefully selected. The terminology is fluent. And the consequences of choosing wrong — a failed implementation, wasted budget, staff resistance that poisons the next attempt — are significant enough that getting the evaluation right matters.
This is a guide to doing that evaluation well. The questions below are not gotchas designed to trip up bad vendors. They are diagnostic questions designed to distinguish between consultants who understand what AI transformation actually requires and those who are primarily selling tools, hours, or confidence.
Question 1: What does your engagement look like after the technology is deployed?
This is the most important question on the list, and it is not about technology at all. AI transformation that sticks requires changes to processes, roles, and organisational habits that don't happen during implementation. They happen in the months afterward.
A consultant who is primarily a technology vendor will describe their engagement ending at deployment. A genuine transformation partner will describe what happens next: how they support adoption, how they measure whether the change is actually embedded, and how they transfer capability to your internal team rather than creating ongoing dependency.
Listen carefully to how fluently they discuss the post-deployment period. If the language gets vague, the engagement structure is probably optimised for implementation fees rather than outcomes.
Question 2: Can you describe a project that did not go as planned, and what you did about it?
Every AI implementation encounters problems. Consultants who have done significant real-world work have a specific story ready for this question — a technical integration that took three times as long as expected, a user adoption problem they didn't anticipate, a business case that had to be revised mid-project. They will be direct about what went wrong, how they responded, and what they learned.
Consultants who respond with generalities about how challenges are part of any project, or who pivot to talking about their success stories, are giving you important information about their transparency and how they will communicate when your project encounters difficulty. All projects encounter difficulty.
Question 3: What percentage of your revenue comes from software licensing or vendor referral fees?
This question needs to be asked directly. It is not rude — it is a fiduciary question about conflict of interest.
Consultants who earn commissions, referral fees, or licensing revenue from specific AI vendors are not recommending solutions objectively. They are recommending solutions that are good for their revenue model. This doesn't make every recommendation wrong, but it means the recommendations are not unconflicted, and you deserve to know that before the engagement begins.
An independent consultant who earns fees only for consulting work will tell you that clearly. A consultant who earns significant revenue from software relationships will either tell you (which is good) or be evasive (which is informative).
Question 4: How do you assess organizational readiness, and what happens if we're not ready?
AI transformation is not primarily a technology problem. It is an organizational change problem that happens to involve technology. A consultant who does not have a structured approach to assessing process maturity, data readiness, and cultural readiness for change is treating AI deployment as a technical project when it is a business transformation project.
Ask specifically: What does your readiness assessment cover? How long does it take? What are the most common gaps you find? And critically — what do you recommend when an organization isn't ready? A consultant who always finds organizations ready to proceed, or who doesn't have a clear answer about what happens when the readiness gaps are significant, is not conducting a genuine assessment.
Question 5: How do you handle data privacy and regulatory compliance requirements?
For organizations in government, healthcare, finance, or legal services, this question is non-negotiable. AI systems process data, and the regulatory environment around AI and data in Canada — PIPEDA, Quebec's Law 25, the proposed Artificial Intelligence and Data Act — is evolving rapidly.
A qualified consultant will be able to speak specifically about privacy impact assessments, data governance requirements, and how they incorporate compliance requirements into system design from the beginning, not as a retrofit. They should know the Canadian regulatory landscape specifically, not just US frameworks.
If a consultant's answer is that compliance is your legal team's problem and they focus on the technology, find a different consultant. Compliance architecture is a design question, not a legal afterthought.
Question 6: What does success look like at 6 months, 12 months, and 24 months?
This question tests whether the consultant is thinking about transformation or about project delivery. A project can be delivered on time and on budget and still fail to produce lasting organizational change.
A consultant who is oriented toward transformation will describe success in operational terms at each time horizon: what processes have changed, how staff are working differently, what capability now exists internally that didn't before, and what the measurable business impact has been. The further out they can describe this concretely, the better.
A consultant who describes success primarily in project terms — delivered on scope, on time, on budget — is telling you that their accountability ends at delivery. That is fine for some engagements. It is not fine for transformation.
Question 7: How do you build internal capability rather than long-term dependency?
This is related to the first question but worth asking separately because the answer often differs. The consulting model that maximizes revenue is one where clients remain dependent on the consultant for ongoing work. The consulting model that maximizes client outcomes is one where the consultant actively transfers capability to internal staff.
Ask specifically: What training and knowledge transfer is included in your engagements? How do you measure whether internal teams are genuinely capable of operating and evolving AI systems, versus reliant on you to do it for them? What does your relationship with clients typically look like three years after an engagement concludes?
The last question is particularly diagnostic. Consultants who are building dependency will have long ongoing retainer relationships with most of their clients. Consultants who are building capability will have clients who engage them for new projects rather than for ongoing operation of previous ones.
Question 8: What is your honest view of whether AI is the right solution for our specific problem?
A consultant who is eager to close an engagement will find AI applicable to whatever problem you bring them. A consultant who is genuinely trying to help you will sometimes tell you that a different approach would serve you better — better process design, better data management, better training — before or instead of AI implementation.
Ask this question specifically about your own situation. Describe your most pressing operational challenge and ask for an honest assessment of whether AI is the right tool for it. The quality of the answer — the willingness to say "maybe, but here is what you would need to verify first" rather than an immediate yes — tells you a great deal about whether you are talking to an advisor or a salesperson.
Putting It Together
No consultant will ace every question, and that is not the standard. The standard is honesty, specificity, and orientation toward your outcomes rather than their revenue.
The AI consulting market will sort itself out over the next few years as organizations accumulate real experience with what works and what doesn't. Until then, the organizations that evaluate carefully — asking specific questions, checking references with organizations of comparable size and complexity, and insisting on clear definitions of success — will get significantly better results than those who select on presentation quality and confidence.
The eight questions above won't eliminate risk. They will, reliably, help you distinguish between consultants who have done this work and those who are pitching it.
Remolda is an AI transformation consultancy based in Ottawa. We are happy to answer all eight of these questions — and to tell you honestly when AI is not the right starting point.