Knowing where AI belongs is one challenge. Redesigning the work around it is another. But neither matters much if the partner responsible for making it real cannot bridge the gap between the technology and the business itself.
When leaders ask what most determines the success of an AI initiative, the instinct is to reach for the obvious answers: the model, the use case, the budget. But underneath all of them sits another decision: who you choose to build with. This is not a task to hand to a contractor for a project here and there. It is a decision about who gets close enough to the core of the business to change how the work actually gets done.
Robert Root, Transformation Partner at Southfield Capital, has learned that lesson firsthand. In a conversation with Kingsley Gate’s Alison Woodhead, he described a market that too often presents leaders with two tempting but flawed choices. At one end are the tool vendors, enthusiasts who saw what large language models could do and set off looking for problems to solve.
“There are a lot of hammers looking for nails out there, people just running around trying to sell solutions.”
Robert Root, Transformation Partner, Southfield Capital
At the other end sit the incumbents who felt they had to respond to AI and defaulted to what they know best, strategy on paper.
“Consulting companies who realized they needed to do something about AI started sending beautiful PowerPoint decks around everything that is possible. I did not find either of those models particularly helpful.”
– Robert Root, Transformation Partner, Southfield Capital
The two failure modes look different, but they share the same problem: neither owns the outcome. The vendor is selling a tool. The consultancy is defining a strategy. Both can leave the hardest part, embedding the technology into how the business actually operates, back to a team that went looking for help precisely because it lacked the capacity to do that alone.
That is often where the value disappears. In a 2025 MIT study that found roughly 95% of enterprise generative-AI pilots produced no measurable return, one of the clearest dividing lines was integration into real workflows. Solutions built with external partners were also more likely to succeed than purely internal builds. The lesson is not “never partner.” It is to choose a partner willing and able to own the integration, not just the install.
The unicorn: business-process depth and technology in one partner
The right partner needs a combination that is easy to describe and remarkably hard to find: deep understanding of how the business actually works, paired with real technology depth.
“Finding a partner with a good balance of business-process understanding and the technology at the same time, that is the magic. And I realize I have just described a unicorn, but that is the moment we are in.”
– Robert Root, Transformation Partner, Southfield Capital
Both halves matter because the most valuable AI applications are rarely generic. They sit at the intersection of technology and what a company uniquely knows: its data, its processes, its customers and the accumulated understanding of how the business creates value. A partner that understands the technology but not the business will tend toward an off-the-shelf solution. One that understands the business but cannot build will tend toward a recommendation. Either way, the opportunity gets flattened into something more generic than the business itself.
Southfield took that logic far enough to change its own approach. Rather than rely on a series of outside providers to build customized solutions across the portfolio, the firm acquired AI orchestration platform Contextual.io, bringing the capability closer to home. Root discussed the build-versus-buy decision in an April 2026 appearance on Nasdaq Trade Talks. The acquisition underscores the larger point: if AI is going to reach deeply enough into a business to change how it operates, the partner cannot simply understand the technology. It has to understand what makes that particular business work.
Why the wrong partner costs more in the middle market
The consequences are sharper for smaller, founder-led companies than for large enterprises, for a reason that is easy to overlook. In these businesses, a handful of people often cover the ground that entire departments cover elsewhere. There is no internal AI function to catch a partner’s mistakes, no deep bench to absorb a failed pilot, and no slack to run twelve experiments in the hope that two land.
The wrong partner does not merely waste money. It consumes the attention of the very people the business depends on most. That is also why these initiatives belong with the CEO rather than being delegated down the org chart. The decisions at stake are not simply which technology to buy. They are where AI belongs, how the work should change around it, and ultimately what the business should look like on the other side.
Choosing the partner responsible for helping answer those questions is not a procurement exercise. It is a strategic decision about the core of the business and deserves to be treated with the same seriousness as any other choice that shapes where value comes from.
How to spot the right partner: a few simple tests
The discipline comes down to a handful of tests. Be skeptical of anyone leading with the tool rather than your business. Be equally skeptical of anyone leading with a slide rather than a working system. Look for the rare partner who can hold both the process and the technology at once, and who is willing to own the messy work of integration, where value is actually won or lost.
And treat the choice as what it is: one of the highest-leverage decisions in the entire plan, not a line item buried beneath it. Unicorns are rare by definition. But in this market, that combination of business-process depth, technology capability and willingness to own the outcome is increasingly the thing worth searching for, and sometimes worth building or buying outright rather than hoping to rent. Because ultimately, the AI strategy is not the technology you choose.
It is the business you are able to build with it.
Related reading
This article is part of a Kingsley Gate Insights series drawing on a conversation between Alison Woodhead and Robert Root, and on Root’s April 2026 appearance on Nasdaq Trade Talks. Quotes are used with permission.
References
Southfield Capital, “Bob Root on Nasdaq Trade Talks: Delivering AI Solutions to Middle Market Businesses,” April 2026.
Nasdaq Trade Talks, “Integrating AI into Private Equity and Investment Banking Operations.”
MIT NANDA research on enterprise generative-AI adoption (“The GenAI Divide: State of AI in Business 2025”), as reported by Fortune, August 2025.
English 
Español de México