AI Consulting Firms: Substance vs. Hype
Choosing an AI consulting firm in 2026 is harder than choosing almost any other kind of professional partner, because the market is young, crowded, and inconsistent. Thousands of firms now advertise AI transformation expertise, and a large share of them pivoted into it recently, from general IT, ERP implementation, or process consulting, as late as 2023. Their credentials are often real. Their track record specifically in delivering enterprise AI value is often thin or absent. The standard tools for evaluating a consultant, analyst rankings, case-study portfolios, certifications, were built for mature service categories, and AI transformation is not one yet.
That gap between how a firm presents and what it can deliver is the entire problem, and it is expensive to get wrong. So the useful skill for a buyer is not finding the firm with the best deck. It is developing a reliable way to tell substance from hype before committing budget, and that comes down to knowing which signals actually predict delivery and which are just good presentation.
The failure rate is the backdrop
Start with the sobering context, because it explains why this matters so much. Most enterprise AI efforts are not delivering what was promised at the proposal stage. NTT Data research found that 70 to 85% of AI deployment efforts fail to meet their desired return on investment, which echoes the broader finding that the large majority of enterprise AI initiatives produce no measurable business return.
Read that number carefully. It means that if you pick an AI consulting partner at random from the ones pitching you, the base rate says the engagement probably underdelivers. That is not a reason to avoid help; it is a reason to evaluate it far more rigorously than the market’s marketing invites you to. The firms in the minority that deliver are distinguishable from the majority that do not, but not by the things most buyers look at.
The red flags
A few signals reliably indicate a firm that is selling hype, and they are visible early if you know to watch for them.
Leading with tools before understanding your business. A firm that proposes AI solutions before it has understood your workflows, your data, and your goals is optimising for speed of sale, not for your outcome. When the tool comes first, the impact comes last, if at all.
Acting as an order-taker rather than an advisor. If the firm builds whatever you ask without pushing back on whether it is the right thing to build, you are getting hands, not judgment, and judgment is the thing worth paying a consultant for.
Overpromising ROI. Guaranteed, unrealistic return figures with nothing behind them signal either a misunderstanding of how hard AI delivery is, or a willingness to say anything to close.
AI-washing. Buzzword-heavy pitches, “AI-powered everything”, with no clear, technically specific answers underneath are the clearest tell. A firm that cannot explain, plainly, when a given AI approach is and is not appropriate does not understand it well enough to deliver it.
No real in-house capability. A firm reselling others’ work with no depth of its own inherits none of the accountability that makes delivery reliable.
The green flags
The positive signals are harder to fake, which is exactly why they are more predictive.
Domain mastery plus specific AI fluency. The firms that deliver pair genuine understanding of your business problem with concrete, technically sound AI knowledge, not vague transformation language. They can tell you specifically why a particular approach fits your situation.
Integration as a core competency. Real AI value comes from integrating with your actual workflows and proprietary data, how your sales data moves, how your operations run, not from layering a tool on top. Firms that treat integration as central understand where value actually comes from.
Honesty about a failed engagement. This is the most revealing test. A firm that can describe, specifically, an engagement that underdelivered and what it changed as a result is showing you a learning orientation that predicts delivery. A firm that pivots instantly to another success story, or says “we stand behind every engagement” with no specifics, is telling you more than it intends.
Measurable outcomes and ownership. The firms worth hiring talk in terms of business outcomes they will own, not prototypes they will build. As the honest voices in the industry put it: AI is not the story; business outcomes are.
The two questions that cut through
You can compress most of this evaluation into two questions, and the quality of the answers tells you most of what you need to know.
The first: “Describe an engagement that underperformed relative to expectations. What happened, and what changed as a result?” A substantive firm answers specifically and without defensiveness. A hyped one deflects. The willingness to discuss a real failure is one of the most reliable indicators of delivery maturity, because firms that treat each engagement as a learning event, rather than a transaction, are the ones that sustain value.
The second: “What would you refuse to build for us, and why?” A firm with judgment has clear answers, things that are not worth doing, approaches that do not fit your situation, AI that would not survive contact with your reality. A firm selling hype wants to build whatever you will pay for, and has no “no” in it. The presence of a considered refusal is a sign you are talking to an advisor rather than an order-taker.
Substance is a model, not a pitch
Notice that none of the green flags are about the pitch. They are about how the firm is built: whether it owns outcomes or ships deliverables, whether it integrates with your business or layers tools on top, whether it treats engagements as learning or as transactions. Substance is a business model, and hype is a marketing style, which is why you cannot tell them apart from the website. You tell them apart from how the firm behaves under specific questions.
It is the distinction underneath the broader repricing of consulting in the AI age: the durable firms own outcomes and proprietary capability, while the ones selling hours of undifferentiated AI expertise get found out. It comes down to whether you are renting an answer or hiring someone to own the result. A substantive AI partner leaves you with a working system and measurable outcomes; a hyped one leaves you with a prototype and a story.
Thane Alaric is built on the substance side of that line by design, owning the outcome and leaving an owned system rather than selling AI-transformation as a pitch. But the point of this piece is not who to hire; it is how to evaluate anyone you are considering. Pressure-testing a specific firm against these questions goes faster in conversation than on paper. Book a call if you want a second read on one you are already talking to.
Frequently Asked Questions
How do I choose an AI consulting firm?
Evaluate how the firm is built, not how it pitches. Look for domain mastery paired with specific AI fluency, integration with your actual workflows and data as a core competency, honesty about a past failed engagement, and a focus on measurable outcomes it will own rather than prototypes it will build. The signals that predict delivery are behavioural, not the deck or the certifications.
Why is it hard to evaluate AI consulting firms?
Because the market is young and inconsistent. Thousands of firms now claim AI expertise, many having pivoted from general IT or ERP work as recently as 2023, so their credentials are real but their AI delivery track records are often thin. The standard evaluation tools, analyst rankings, case studies, certifications, were built for mature service categories and do not reliably separate substance from presentation here.
What are the red flags in an AI consulting firm?
Leading with tools before understanding your business, acting as an order-taker rather than an advisor, guaranteeing unrealistic ROI, buzzword-heavy “AI-washing” with no technically specific answers, and no real in-house capability. Each signals a firm optimising for the sale rather than your outcome. They are visible early if you ask specific questions rather than accepting the pitch.
Do most AI consulting engagements succeed?
No. NTT Data research found that 70 to 85% of AI deployment efforts fail to meet their desired ROI, consistent with broader findings that most enterprise AI initiatives produce no measurable return. That base rate is not a reason to avoid help; it is a reason to evaluate potential partners far more rigorously than the market’s marketing encourages, since the minority that deliver are distinguishable from the majority that do not.
What questions should I ask an AI consulting firm?
Two cut through most of the noise. First: describe an engagement that underperformed, what happened, and what changed as a result. A substantive firm answers specifically; a hyped one deflects. Second: what would you refuse to build for us, and why? A firm with judgment has clear answers; one selling hype wants to build whatever you will pay for and has no considered “no.”
What separates a substantive AI firm from a hyped one?
How it is built, not how it presents. Substantive firms own outcomes, integrate with your business and data, and treat engagements as learning events; hyped firms ship deliverables, layer tools on top, and treat engagements as transactions. Substance is a business model and hype is a marketing style, which is why you can only tell them apart from behaviour under specific questioning, not from the website.
About the Author
Daniel Förster is the Managing Partner of Thane Alaric. For over a decade he has built and run companies, served over 2,000 founders, and worked as the embedded operator (COO, CFO and CMO in function) inside founder-led businesses. He now leads Thane Alaric, where companies become AI-native the right way: own the engine, rent the models, keep the judgment that’s yours.
