The AI Consulting Grift
In 2017, I realized that V-Squared’s success with AI strategy was built on internal strategists’ shoulders. Yes, the frameworks and implementation support we provide have value. However, it only works when we hand off to people we have helped train, but who work for the business. If our consultants take ownership of the AI strategy from start to finish, it fails.
Internal training and upskilling are some of the most valuable work we do for clients. Training internal employees to take over from us has become our business model, and I wouldn’t change that. BUT…it isn’t as profitable as open-ended strategy or technical services engagements. V-Squared successfully swims in the same pool as much larger sharks because we operate in a niche that other consulting companies don’t want to.
Here's an insider’s secret.
Even though AI strategy consultants make very high hourly rates, AI strategy engagements don’t come with massive contracts. Ours run between 6-18 months with 2-4 consultants working on each engagement. We hand off to the business at the end of the contract and go into advisory mode. One of us remains connected to the business, providing 20-80 hours of advisory services per month, usually for a year or less.
Even our largest contract is small by large consulting companies’ standards. Enterprise strategy or rebranding playbooks can cost tens of millions. Engagements that span multiple years and include implementation support or technical services will push that into the hundreds of millions.
The bottom line is that to optimize revenue, we would need to disconnect our business model from our clients’ outcomes. An excellent example of that comes from McKinsey’s work with Warner Bros, Discovery, and CNN+.
During the early phases of this process, McKinsey was consulting with CNN. They convinced the company that a niche streaming service would be profitable in a short amount of time. As a result, CNN’s leadership committed to spending nearly a billion dollars on CNN+. A few months after launching and $300 million later, CNN+ was shuttered. It was a disaster that was on pace to lose massive amounts of money when subscribers failed to materialize.
In less than 3 years, McKinsey extracted close to $200 million from the companies. Their strategic recommendations resulted in hundreds of millions in wasted spending, never mind how much the merger itself cost, and it doesn’t appear as if any growth or new value was delivered.
Technology Is The New Grift
Consulting companies played the same game with the cloud over a decade ago. They told CEOs that they needed to invest and move all their workloads into the cloud. Slide decks promised massive ROI. The same companies got the contracts to do the migrations and technical work. They got even more contracts to manage and maintain the new cloud infrastructure and services.
Then, a few years ago, consulting companies pivoted. ‘Not all workloads belong in the cloud, and some should be migrated back on premises.’ You’ll never guess who got paid to deliver that strategic assessment and who is also getting the contracts to do that work…
Why get paid once for a strategy when you can deliver an imperfect strategy and get paid a second time to clean up your own mess? Internal strategists and leaders can’t get away with these types of schemes because they have a higher level of accountability. They’re also most closely coupled with implementation and execution. That feedback loop means pivots happen sooner.
Unfortunately, we’re seeing AI strategy follow a similar arc as cloud strategy. For large consulting companies, more projects mean more revenue. If the AI strategy positions AI as the solution to every major business challenge, that creates more contracts. It’s no surprise that revenue from AI consulting is rising so quickly. Just look at the quarterly investors’ meetings notes for companies like IBM and Accenture.
However, a recent survey from Canada shows that many businesses are rolling back their previous decisions to leverage AI broadly. This table shows the rise of boring, but high-value use cases and the slow decline of sprinkling AI on everything.
J&J announced a strategic pivot from prioritizing thousands of use cases to focusing on a handful of high-value use cases. When we’re brought in to support an AI strategy, the first thing we do is slim down the number of AI use cases. It’s an increasingly common theme.
Most businesses are doing way too much. You know it’s getting out of hand when even NVIDIA’s CEO comes out and says that companies shouldn’t be using LLMs for many of the use cases currently under development and in production.
Consulting companies largely enable the hype cycle, and the massive amount of spending on AI projects has delivered a windfall of earnings growth. In the worst case, companies are cutting back across the business to free up the money to spend on AI projects. For many, that impacts service levels for core business functions. Even in the best case, businesses are wasting tens of millions of dollars chasing AI rainbows with very little to show for it, while high-value use cases go unaddressed.
What Should We Do?
Internal leaders and strategists must own the data and AI strategies. Giving that control over to a consulting company is playing with fire. Yes, large consulting companies can support data and AI strategy development and implementation, but ownership must still remain in the hands of employees, vs. external consultants. Otherwise, the temptation to optimize for the external firm’s revenue becomes too much for most to resist.
Every contract must include knowledge transfers and handing over ownership of business-critical processes. Data and AI are part of the core business, so internal personnel need to learn how to support them from strategy to execution and continuous improvement. Without the handoff, an external company controls too much of the business’s destiny.
Outcomes-based pricing aligns incentives between the business and consulting firm. Outcomes are different than artifacts. The AI strategy is an artifact. A model is also an artifact. An outcome is the revenue growth or cost savings that the project is expected to deliver.
Most consultants will jump at the opportunity to implement outcomes-based pricing. AI initiatives, when selected properly, frequently outperform and deliver more value than expected. That means the consultant’s compensation increases as well. A significant portion of the contract (40% - 70%) should be outcomes-based vs. guaranteed.
Consulting firms that repeatedly breach their clients’ trust should be held accountable, but I haven’t seen significant consequences for bad behavior. That must change. If there’s no downside to taking advantage of clients, it will keep happening. At some point, the board or shareholders must intervene to stop the grift.