What Does AI-First Mean For SMEs: Jamie Dimon’s Demons & Big Business's AI Achilles Heal
My instructor-led AI Strategist Certification is on hiatus, but take advantage of my low-cost, self-paced version of the course. Click the link to learn more. My book introduces the frameworks I discuss in this article and is an excellent entry point for AI Strategy.
Jamie Dimon, CEO of JPMC, is going through some things, and it’s a familiar pattern for me. Once C-level leaders, or anyone else at and above that level in the business world, come to terms with AI’s impacts over the next 5 years, they have one of three responses.
Some lean into the skid and work to maximize their gains, even at the expense of others.
The second group hits the alarm and sends rational, understated warnings to impacted groups. They also position themselves to take advantage of opportunities. In some cases, they share the opportunities they see coming.
Jamie Dimon is in the third group. He’s sounding the alarm and offering actionable advice to mitigate the negative impacts.
Urging People To Take This Seriously
Explaining The AI Paradigm’s Impacts
Recommending Next Steps
He also explains where the opportunities are and how to begin positioning the firm to take advantage of them. Read Dimon’s annual letter, and you’ll see the signs. I have seen the realization cycle play out for dozens of CEOs, board advisors, and VCs.
What’s unique about Dimon is the respect he commands from other CEOs. He can be more blunt and direct, telling others at his level not only what the paradigms are but also what to do about them.
CEO, board, and investor speak a different language because they have a different audience in mind. I teach a communications framework for presenting to those groups to account for their decision-making needs. They don’t trust most people enough to take direct advice on what to do. They seek advisors who can provide the expert context necessary to make better decisions.
Jamie Dimon is trusted, so he can say things others cannot. Rather than focusing on the job impacts, he spends most of his time discussing data. I’ll begin part 2 of this series in the same place. Dimon has seen JPMC implement over 400 data, machine learning, and AI use cases in production. He understands the maturity model from first-hand knowledge.
Right, But We Don’t All Have JPMC Money
What does a massive company with billions invested in AI and a team of nearly 2000 data professionals have to do with AI-first in small and mid-sized enterprises (SMEs)? There’s no way to bring that spending level to the problem because SMEs have finite resources.
In part 2, I must discuss constraints. SMEs are resource-constrained, which changes how they approach AI-first. Most view constraints negatively, and I will dispel that myth in this article. Constraints are a competitive advantage for technology and transformation.
Budget and staffing are the most obvious constraints, but not the only ones. How can a limited budget or small teams be advantages? That comes down to how you define data and AI strategy or business strategy as a whole. Here’s the definition I teach.
Strategy is the study of leverage and advantage in competitive, zero-sum games.
One line does a lot of heavy lifting, but the last piece is most relevant to this article. Zero-sum games get a bad rap because, in legacy strategy frameworks, limited resources lead to conflict. Competition is defined by conflict between multiple players or actors, but it doesn’t have to be.
In the modern competitive marketplace, we see more collaboration than competition. The most significant players (AWS, Microsoft, Hugging Face, SAP, etc.) build ecosystems and marketplaces. They make space for partners and monetize the benefits for both parties.
Some marketplaces like Apple’s App Store, Uber, or Amazon Prime’s third-party seller marketplace are predatorial. The largest players take a disproportionate share of the gains. Those are and will continue to unwind or switch to a more equitable model. In Uber’s case, the equilibrium may push prices higher than customers will sustain.
There are two driving forces. The first is scale, and companies with scale can monetize it in multiple ways. They can exert dominance over the marketplace through scale and monetizing access to their ecosystem’s benefits.
Most people miss the second force, which is why SMEs have an advantage. JPMC spent billions and hired thousands because it is the most obvious path to scaling its AI capabilities. Even as Jamie mentioned the second force in mentioning Stripe, he avoided directly commenting on it with respect to JPMC’s approach.
Stripe used technology to disrupt payments by serving a large customer base more efficiently than larger incumbents like JPMC. Stripe could maintain margins, provide a substitute product, and deliver the same service levels at lower costs because its operating model was more efficient. Technology was the lever that delivered that advantage.
And We Don’t Have Technology First Business And Operating Models
Scarcity forces efficiency, but SMEs need a roadmap to efficient growth through technology. That’s the challenge I saw in 2015. My clients weren’t willing to invest as heavily as JPMC did, and my budgets and resources were constrained. Buy-in was a challenge because data and AI hadn’t proven themselves as levers for advantage.
My data and AI strategy and product management frameworks were built from this necessity. Throwing cash at the problem wasn’t an option, and I needed a repeatable way to deliver significant value on a shoestring budget. I was often the only technical resource for developing solutions, or I had access to a tiny data team.
Business units were at an early data maturity level and couldn’t integrate solutions into their workflows. Data wasn’t being gathered properly, and getting access to it was challenging. Infrastructure and additional data team resources were too expensive. I needed to show returns and impacts every quarter, so I had to build towards bigger initiatives in smaller steps. These are familiar challenges for anyone working at an SME.
These are not technical challenges. The solutions involve change management, transformation, and strategy. At companies like JPMC, these are established capabilities, but few SMEs have resources for them. Everything must be lightweight, flexible, and obviously valuable.
There are several constraints, and that’s good. We have a well-defined problem space. The most obvious option, spending our way out of the problem, isn’t available, and that’s good, too. It forced me to find a more efficient solution. This feels more difficult than it really is. As soon as we accept the constraints as givens and treat this like an engineering problem, it’s easier to work on solutions.
Creating Frameworks Based On Reality (Where The Business Is, Not Where We Wish It Were)
One of my frameworks is “Meet The Business Where It Is.” We must create solutions that work in more than just the ideal setup. Like engineering solutions, strategy solutions must account for edge cases, a range of users, and changing conditions. They must be designed to work with people instead of forcing people to change. There must be obvious, significant value in adopting them.
With very large companies, the option of spending their way out of solving these problems is too attractive to pass up. The assumption is that bringing all those resources to the problem leads to a rapid solution. In reality, every solution expands to take up the available resources.
JPMC’s AI organization has 1 leader for every 5 ICs. It’s heavy in the middle, and every unnecessary leadership layer slows the data team’s delivery. Technical ICs often say that strategy slows progress, but in reality, the bureaucracy slows technical delivery and sidelines innovation.
Scarcity leads to efficiency. The myth of massive resources is that transformation and maturity will happen faster. Just look at how long the journey took for businesses like Walmart or JPMC (both 6 years). I wouldn’t call their transformations rapid.
Airbnb was faced with an existential threat at the beginning of the pandemic. Its data maturity journey rapidly accelerated and was driven by necessity, AKA constraints. The company didn’t have the luxury of time, so its transformation happened much faster. While the technical initiatives began at the beginning of 2019, the most significant progress on the maturity and adoption fronts happened in early 2020.
The trends are consistent. Constraints lead to efficiency. Necessity leads to acceleration.