AI is turning the knowledge consultants used to sell into a free commodity. On our latest episode of Bricks, Bucks and Bytes, Buro Happold CTO Alain Waha argued the profession’s surviving edge comes down to three things a client cannot download: trust, insurance, and compounded experience from projects no textbook ever captured. If your value is a generic answer, he says, you are already out of business. If your value is what your firm has done 20 times, you might be more valuable than ever.
Martin asked the question most consultants would rather avoid. The entire consulting model was built on an information gap: the adviser knows a lot, the client knows a little, and the fee lives in the space between. What happens when the client can query the same knowledge, instantly, for the price of a chatbot subscription?
It is a fair challenge to put to Alain Waha. As CTO of Buro Happold, a 3,500 person global engineering firm, he sells expertise for a living, and he has spent the past year telling this show that AI is the biggest shift the built environment has seen since steel and electricity.
His answer did not defend the old model. He buried it, then explained what gets built on top.
The Old Arbitrage Is Dead
Consulting was built on knowing more than the client
Alain’s starting point is blunt. Historically, he argued, a consultancy was an arbitrage business: it accumulated knowledge its customers did not have and rented it back to them. That arbitrage only works while the customer stays on the wrong side of the information gap.
Large language models close the gap for anything generic. A developer can now ask a frontier model how a district heating scheme works, what a typical facade buildup looks like, or how other projects phased their groundworks, and get a decent answer in seconds. The consultant whose product was that decent answer has a pricing problem that no amount of relationship management will fix.
If commodity knowledge is generic and available for free, then everybody that gave you a generic answer is out of business. Which is great, because it means we’re gonna get better answers.Alain Waha, CTO of Buro Happold, on Bricks, Bucks and Bytes
Note the second half of that quote. Alain is not mourning the generic answer business. In his view, its collapse forces the industry to compete on something better, and clients end up better served.
What Clients Still Pay For
Trust, insurance, and the judgment call
So what survives? Alain’s test case was pointed: imagine an AI produces the mechanical and electrical design for your building, or the fire engineering strategy. Would you build it? Would you put people inside it? For him, the honest answer today is no, not without a human who has critically reviewed the output and is prepared to stand behind it. “When we say human in the loop,” he said, “what we mean is trust.”
Behind trust sits a very unglamorous second layer: insurance. Professional advice in construction is backed by professional indemnity cover, the insurance that pays out if the advice turns out to be wrong. Silicon carries no such policy, and as Alain put it, it hallucinates. A client can hire that accountable human in-house or buy them through a consultant, but somebody insurable has to own the judgment.
- Trust. A qualified human critically reviews the machine’s output and puts their name on it. The review, not the raw answer, is the product.
- Insurance. The advice is backed by professional indemnity cover, so there is recourse when something goes wrong. An AI cannot be sued, insured or struck off.
- Experience. Compounded, firm-specific knowledge from having done this exact thing many times before, which never made it into any training data.
Sharper thinking on AI in AEC, weekly
The strategy shifts, the deals and the operators redefining the built environment, without the hype.
Join 3000+ Readers
The Alpha Is What You Have Done 20 Times
Experience that never made it into a textbook
The third layer is where Alain got specific. Picture a firm that has delivered 20 arenas. It knows the delivery issues that actually bite, the acoustic outcomes that different designs produced, how turnstile layouts change matchday revenue, and how fast the building can be turned around overnight between one band and the next. None of that operational knowledge is generically available, because it was never written down anywhere a model could read it. That, he argued, is why the client appoints that firm, and it goes well beyond simply having a human with an insurance policy.
He borrowed a word from finance to describe it: alpha. Investors already ask fund managers what their edge is over the market. Clients, Alain believes, will start asking consultants the same question. Commodity knowledge sits at everyone’s fingertips, so the firms that win are the ones that capture their hard-won project lessons, encode them, and bring them to every new job on top of the free baseline. Owen noted on the episode that Palantir has been pushing a version of the same message in its positioning around sovereign, proprietary AI: the value concentrates in the knowledge only you hold.
The uncomfortable flip side is that firms which let experience walk out the door with every retirement are giving away their only durable asset. Alain has made this case on the show before, and the arrival of free generic knowledge just raises the stakes on it.
Commodity Knowledge vs Compounded Knowledge
The two assets, side by side
| Commodity knowledge | Compounded firm knowledge | |
|---|---|---|
| Where it lives | Textbooks, standards, the open web, model training data | Project archives, senior heads, lessons-learned reviews |
| Cost to access | Close to zero, a chatbot subscription | Years of delivered projects and deliberate capture |
| Who can use it | Everyone, including your client | Only the firm that earned it |
| What it is worth | Falling toward zero as models improve | Rising, because it is the remaining differentiator |
| Example | How arena acoustics work in principle | How 20 delivered arenas actually performed, and why |
Frequently Asked Questions
Knowledge arbitrage is charging for the gap between what you know and what your client knows. Consulting leaned on it for decades. Large language models collapse that gap for anything generic, because the client can now query the same baseline knowledge directly, which is the shift Alain Waha described on the episode. (Source)
Not in Alain’s telling. It replaces the part of consulting that was only ever a generic answer. What remains, and arguably grows in value, is accountable judgment: a human who critically reviews machine output, stands behind it professionally, and brings experience the model has never seen.
Professional indemnity insurance backs advice with financial recourse if it proves wrong, which is a condition of doing serious work in the built environment. Alain’s point is that silicon carries no PI and it hallucinates, so an insurable human must own the judgment before anyone builds or operates from an AI-produced design.
Alpha is the finance term for returns above the market baseline. Applied to consulting, it is the proprietary, compounded knowledge a firm brings on top of the free commodity layer, like the operational lessons from 20 delivered arenas. He expects clients to start asking firms directly what their alpha is. (Source)
Treat knowledge capture as a strategic priority rather than an admin task: encode project lessons before people retire, build systems that bring that knowledge to every new job, and price on the outcomes it produces. Alain has argued on previous episodes that firms which compound knowledge will pull away from those that let it leak. (Source)
Alain Waha is Chief Technology Officer at Buro Happold, the global engineering consultancy, where he leads technology strategy and AI transformation across roughly 3,500 people. He joined construction after careers in chip design and automotive, and holds a monthly slot on Bricks, Bucks and Bytes. (Source)
Related Articles
Buro Happold’s CTO on Why AEC Should Stop Chasing Esperanto
Silicon Is the New Steel, Our Full Episode With Alain Waha
Why a16z, Palantir, BlackRock and SoftBank Are All Buying Construction
