AI Puts Consultancies Between a Rock and a Hard Place

To evolve into a maxim, an observation must recur so often that it leaves a coherent and lasting impression. It’s obviously more difficult to learn from an exceptional event than from a succession of similar occurrences that have common characteristics and consistent outcomes.

When we use a maxim, what we’re saying is that we perceive a truth and a rule of conduct in the statement. For example, consider the maxim: "Underpromise and overdeliver." That's unquestionably sound advice, and not only for salespeople. It’s a maxim that represents wise counsel for nearly every business initiative imaginable; I would contend that it has widespread applicability in life more generally.

Here's the problem, though: What if circumstances conspire against underpromising? What if you face a situation where your organization, or market forces, or your customers will not countenance restraint or tempered expectations? Unfortunately, these situations do occur, and we have evidence of one example in a recent article published in the Wall Street Journal.

The article is titled, “How the AI Boom is Leaving Consultants Behind.” That title is ambiguous — you can interpret it in more than one way — but the text that follows is clear as the purest spring water. Looking at the title, you might think it’s another article hyping AI’s presumed ability to put humans — consultants, in this instance — out of work. But no, that’s not what this article is about.

Instead, this article deals with a market conundrum that bedevils consultants.

Consultancies are pressured to sell expertise in AI, a technological capability that — you might have noticed — is all the rage in executive boardrooms and enterprise IT departments. The overwhelming demand for AI mastery, however, is not the salient problem. When you’re a consultant, or a vendor, demand is usually not a problem. Demand becomes a problem only when you are not in a position to meet it satisfactorily. Meeting demand doesn’t merely involve providing customers with an adequate supply (a quantity) of whatever is in demand; it also involves providing those customers with perceived and real value for money.

Meeting Demand Involves More than the Quantity of Supply

That, you see, is where consultants have missed the mark. The consultancies’ problem, bluntly stated, is that they are selling AI expertise and proficiency while possessing very little of either. Even then, that might not be an intractable problem if enterprise clients remained under the false impression that the consultants knew what they were doing. Unfortunately for the consultancies, many enterprise clients know that they’re not getting value for the money they're spending on AI consulting services.

In fact, the disconnect is often jarring between what clients want and expect from the AI consulting engagements and what they ultimately receive. The result is that the consultancies are seen to be overpromising and underdelivering as opposed to underpromising and overdelivering. To belabor the obvious, this is no way to engender customer loyalty or build a thriving AI practice. Clients whose promises are not being met eventually discontinue the commercial engagement. Even worse, many clients are so disillusioned by their experiences that they’re in no mood to reward the consultancies with market-growing endorsements or testimonials. Disappointed customers do not provide emphatic referrals.

Consider this excerpt from the article:

Clients quickly encountered a mismatch between the pitch and what consultants could actually deliver. They found that consultants, who often had no more expertise on AI than they did internally, struggled to deploy use cases that created real business value.
Sometimes consultants built successful proof of concepts, but couldn’t scale them across the business. That was the case at pharma company Merck, said Chief Information and Digital Officer Dave Williams.
“We love our partners, but oftentimes they’re learning on our dime,” he said.

“We Learn on Your Dime!”

“We learn on your dime!” That’s not exactly the tagline that PwC or McKinsey wants to shout from the rooftops in their next AI-related adverting blitz. It’s not the sort of pitch that induces a prospective client to reach for the phone, much less the wallet. The following excerpt, referring to a consultancy’s disappointing follow though, vividly describes the problem:

“They overpromised,” said Magesh Sarma, chief information and strategy officer at AmeriSave Mortgage. When it came to building real use cases, he said, “we discovered that they really also had no idea how to do these things.” He added, “They were just as good or as bad as what we would have been able to do in house.”

People talk about the virtues of truth in advertising, but nobody ever takes the plunge. When it comes to advertising, nobody — especially the advertiser — wants the constraining embrace of truth. You can see why.

If Magesh Sarma’s experience was the inspiration for a consultancy’s promotional campaign, the slogan would be: “We’re just as bad at AI as you are — but you pay us for it!” It has a certain ring to it, but the ring might first involve a peremptory call to fire the consultancy’s advertising agency or marketing department.

I think we can deduce how such an unfortunate situation arose. For many enterprises, AI is a little like sex for high-school students: Plenty of people are talking about it, but few are actually doing it. Even fewer enterprises are doing AI with any degree of competence or meaningful results.

For consultancies, the fact that enterprises are all talking about, and interested in, AI means that the consultancies must talk about it, too. More to the point, consultancies must propose credible AI counsel, including use cases, implementation guides, performance metrics, and roadmaps for continual process improvements. Those are the sorts of services enterprises expect of consultancies. Unfortunately, given the nascence of AI — which is, let’s be frank, still more like dreamy aspiration than quantified substance at this point — the consultancies really don’t have much experience from which to devise and codify best practices. Instead, they have, as Sarma puts it so candidly, “no idea how to do these things.”

Explaining, but Not Excusing

The prudent course of action, of course, would be for consultancies to hold off on the braggadocio and the bluster, to avoid making promises that they can’t possibly keep. Under no circumstances should they have overpromised and underdelivered, so why did they do it?

Well, the answer is simple: The monsoon of AI hype — from corporate publicists, PR agencies, the business and trade press (not all, of course), and vendors selling AI infrastructure and cloud-based services — has permeated the consciousness of enterprises large and small. They have come to believe that AI is happening, even that it has already happened at their competitors. Consequently, they feel they need to get moving, that they’re late to a party that, in fact, hasn’t actually started.

So, enterprises feel the need to reach out, now, to third-party experts (consultancies). But most enterprises, including the organizations’ competitors, are just as benighted and inexperienced when it comes to practical application of AI as is everybody else.

In these circumstances, consultancies are, somewhat paradoxically, also at a disadvantage. They thrive when they’re able to smoothly adapt or replicate best practices at organizations that have achieved success in the gainful application of a given technology. AI is not at a stage where enough of those organizations exist, putting the consultancies in a position of selling generalities, often of the vaguest variety, rather than pragmatic and results-oriented particularities.

It’s not quite a chicken-and-egg situation, because we know what needs to come first — success stories from organizations who have used the technology productively — but the market hasn’t evolved to the point where enough of those stories exist. Since we don’t have enough of those AI trailblazers and their tales of AI achievement, consultancies and other enterprises can’t learn and profit from their exemplary experiences.

For consultancies, the good news is that the situation should improve as and when early adopters of AI figure out where the technology does and doesn’t generate tangible business benefits. At present, however, AI nirvana remains an illusory oasis rather than an easily attainable destination for most enterprises and the consultancies who’ve oversold them on instant gratification.