The End Of Consulting: Data, AI, & Accountability For Outcomes
We’re all waking up to an outcomes-focused reality where we are accountable for delivering business and customer outcomes, not artifacts and technology. Only 3 spots remain in my new 12-week Data & AI Strategist Certification. My next AI Product Management Certification cohort is also open for registration.
Don’t miss out on your chance to prepare for high-value roles and future-proof your career.
Data and AI play by different rules than other technologies, and I explained the 5 hard truths in my last post. There is no finish line, and nothing is ever done. To succeed with AI, businesses must come to terms with continuous transformation and iterative improvement.
Every part of the business and product generates data, so any approach to the AI journey must be holistic. Other technologies didn’t have that level of impact. Tasks and decisions don’t produce code or apps. Data is the code every action and decision is written in, and AI extracts the logic from it.
There’s no clean starting line, and it’s never a good time to begin a data and AI maturity journey. However, data generation happens all the time. Businesses need only build contextual data-gathering processes to take advantage of it. The only barrier to starting is ingrained in business culture.
We never achieve perfection, so the AI journey will always start at an imperfect place. The process can’t be pushed off until we have the perfect data, the ideal infrastructure, a data or AI-literate business, and a fully developed data organization. In my data and AI strategy class, I teach frameworks that adapt data and AI to serve core business strategy from multiple, imperfect starting points.
Technology can’t be ignored just because it doesn’t fit the current business or operating models. It will advance under imperfect conditions, whether the business supports it or waits for ideal conditions that will never materialize.
Google is the prime example of technology moving around roadblocks. It shelved GenAI because integrating LLMs into search results disrupts ad revenue. People left Deep Mind and developed GenAI products themselves. This isn’t a recent trend. Major retailers ignored eCommerce, so a bookseller developed its business model around it, and Amazon scaled to fill the gap left open by brick-and-mortar retailers.
Unlike OpenAI and Amazon, most businesses can’t support years of losses on a journey to uncertain profits. Investors are losing patience with companies that are ramping up spending but can’t quantify the expected returns. Citi’s Head of Equity Trading Strategy, Stuart Kaiser, framed it nicely.
“Unless you can demonstrate AI revenues, investors are going to be skeptical. The bar is really high. A very small number of {companies} are actually demonstrating revenues.”
AI journeys must balance two needs: immediate incremental returns during the journey and multiple larger long-term revenue growth destinations.
To succeed, we must Monetize the AI Journey. However, most AI strategies and product delivery models focus only on monetizing a single destination. In this article, I will use case studies to explain why most AI journeys fail and introduce my approach to Monetizing the AI Journey that results in multiple large growth drivers.
The Money Tap Is Wide Open But…
There’s evidence of spending, but where is ROI materializing? So far, the GenAI boom has only benefited a small segment of businesses. NVIDIA and other chip makers were the earliest winners. Hyperscalers caught the second wave, and nearly all have reported boosts from GenAI workloads. Consultants and platform accelerators are cashing in on a third wave right now.
IBM has over $1 billion in commitments for GenAI consulting and watsonx purchases. Accenture booked $300 million last year. 40% of McKinsey’s business will be GenAI-related in 2024. KPMG went from no GenAI revenue to reporting over $650 million in business opportunities in just the last 6 months.
My consulting practice (V-Squared) began to pick up in 2022. We’re one of the oldest data and AI consulting companies in the world, and I haven’t seen anything like this wave before. It’s not just how quickly it began but how it’s forcing traditional consulting models to evolve. Money has been spent, and the ROI question is front of mind for C-level leaders.
Where’s the fourth level of returns? Businesses are spending on AI workloads in the cloud, buying AI platforms, and dropping billions on consultants. Some are even laying people off from one side of the business to free up more money for AI. We should see more ROI and material impacts in quarterly earnings, but we’re not. Profit and loss don’t lie.
The problems that block businesses from realizing those returns have been with us for decades. It took me 15 years to come to terms with the scope of the challenge and start the slow process of developing solutions. We must change the way we approach monetizing technology, or it will continue to be a cost center.
The 6th hard truth about data and AI extends to all technologies. Businesses don’t pay for technical artifacts. They pay for outcomes. Data creates visibility into longer value chains. For the first time, business leaders can connect the dots between money spent on consultants and platforms and the value created by the artifacts they deliver.
Is Money Invested Delivering The Promised Returns?
Walgreens went as far as hiring IBM’s former Chief Data Officer 2 years ago, yet it struggles to understand changing customer preferences. That’s a bread-and-butter data and machine learning use case that the company hasn’t successfully implemented.
Instead, Walgreens invested in Blue Yonder’s platform and planned to use its machine learning to improve inventory accuracy at its stores. If that was ever implemented, it doesn’t appear to have delivered much in terms of ROI. Why overlook the obvious, high-value use case (customer preference modeling) for something with dubious value?
I work with retail clients, and we began preparations for changing customer preferences in 2022. Inflation and rising interest rates are well-understood catalysts. My clients used data and simple models to measure their pricing power. They assessed price sensitivity across multiple customer segments and product categories.
As a result, they negotiated deals to keep price increases under control and updated their product mix to support customers who would begin to trade down to generic brands. Although I’m using Walgreens for this example, many other retailers failed to see change coming and implement straightforward mitigations. These are simple, low-cost data and machine learning use cases that don’t require expensive infrastructure or talent.
Businesses miss opportunities to Monetize the AI Journey when they focus on monetizing a single destination. Traditional consulting models and infrastructure are built to monetize the destination with large artifacts. Businesses need repeatable processes and incremental infrastructure migration roadmaps that Monetize the AI Journey multiple times.
McDonald’s also saw the impacts of inflation and higher interest rates on customer preferences. Experienced business leaders have seen this trend before and updated McDonald’s business strategy to take advantage of the opportunities that past cycles created.
It opened more stores in underserved locations to support people trading down to McDonald’s from fast-casual restaurants. It ran ad campaigns to remind customers how much further their money stretches at McDonald’s. Data, descriptive models, and experts delivered significant value without advanced AI.
However, without frameworks for Monetizing the AI Journey, businesses do not consistently apply that decision framework. When business leaders don’t understand why some initiatives succeed while others fail, they can’t put frameworks in place to repeat the successes. Data is a novel asset class because a single dataset can be monetized multiple times, making a repeatable process critical to Monetizing the AI Journey.
The AI Trap With A Side Of Fries
Despite tangible successes, failures to realize significant returns or monetize the destination caused the McDonald’s leadership team to withdraw from data and AI investments. McDonald’s has aggressively reduced its headcount and closed positions in the technology organization.
It canceled in-store automation initiatives like robotics for food preparation and cooking. The CEO said the unit economics for most robotics automation use cases didn’t work. It sold McD Tech Labs to IBM in 2021 and contracted with IBM to run its AI-enabled order-taking and drive-through automation initiatives. Last week, McDonald’s ended that test run with IBM early. Accuracy was one problem, but if you’ve taken my AI Product Management Certification course, you know the bigger problem.
I use McDonald’s drive-through automation initiative to study what happens when we fixate on monetizing the destination and prescribe a solution to do it.