Long Lake’s $6.3 Billion Amex GBT Deal Tests AI-Led Buyouts

Alex Taubman describes Long Lake Management as purpose-built for a specific kind of AI deployment: buying services businesses, embedding applied AI into their operations, and using the resulting productivity gains to expand rather than simply cut costs. Elad Gil framed the firm’s announced agreement to acquire American Express Global Business Travel for $6.3 billion as what he believes may be the world’s first “AI take private,” following roughly 30 prior acquisitions by Long Lake under the same premise.

The operating premise is not that a services company needs a thin AI layer on top of its existing software. Taubman said Long Lake has invested heavily from the beginning in a horizontal AI platform called Nexus, which sits between models on one side and a company’s data sources, skills, and workflows on the other. Long Lake is model-agnostic, he said, and about 80% of the infrastructure is shared across verticals. The remaining work is deployment: mapping workflows, understanding and cleaning up data sources, integrating systems so models can access them, and building tools around the business’s actual pain points.

That deployment burden is central to the thesis. Taubman said Long Lake’s first acquisitions took more than a year before the firm could identify the real potential of AI and see it in business outcomes. Now, he said, the platform can be deployed within days of partnering with or buying a company, producing immediate time savings.

Gil put the implication plainly: can Long Lake buy a company and, within weeks, get margin lift because the employees move onto a platform already built for similar businesses? Taubman accepted the operational premise but corrected the emphasis. What Long Lake sees immediately is time savings. The strategic question is what to do with that saved time.

Long Lake’s answer, according to Taubman, is growth. He said the firm is “not really” focused on cost saving, but on growth and customer experience. His view is that AI is “incredibly positive sum”: if people become more productive, the company wants more productive people, not fewer of them. Happier customers lead to faster growth, and faster growth can create more jobs.

The clearest concrete example came from Long Lake’s homeowners association business. Taubman said companies Long Lake invested in were typically growing 0% to 5% annually in volume. Long Lake’s HOA company, he said, is now the fastest-growing company in that industry and is growing organically at more than 20% a year.

Taubman attributed the change to giving team members extra capacity to serve more customers, improving customer acquisition economics because incremental customers can be served at lower cost, and offering better products and services. For him, that is the deeper shift: applying what he called a “software style playbook of go to market” to “sleepy industries.” The goal is not merely to automate administrative labor. It is to change the unit economics of growth in services businesses that historically scale by hiring roughly in proportion to revenue.

Elad Gil asked why Long Lake buys companies rather than selling AI software to them. The traditional Silicon Valley route, he noted, would be to identify a vertical like HOA management, build software for the industry, and sell it as a vendor.

Alex Taubman answered with alignment. Software companies can be valuable partners, but when a vendor is only selling software, it does not have the same relationship to the business outcome. By owning the company and the customer relationships directly, Long Lake believes it can drive better results.

That ownership also changes who the engineers are building for. Long Lake’s engineers treat the company’s own field employees as their customers. The internal feedback loop is much tighter than in a vendor relationship: engineers are in the field with team members, observing workflows, hearing pain points, and building tools inside Nexus to solve them. Taubman compared it to the old skunkworks idea that engineers and the factory should be co-located to accelerate innovation.

Members of Long Lake’s engineering team were, at the time of the interview, likely spread across 20 states, sitting with employees in architecture, HOA management, HR services, and specialty tax businesses. The work is not just technical. It involves “a deep amount of change management,” including sitting with teams, understanding where work breaks down, and translating those problems into AI-enabled tools.

Gil identified that as one of the major barriers to AI adoption generally: changing processes and organizational design. If Long Lake owns the company, it can make those changes. A software vendor, by contrast, can influence them only indirectly.

Taubman’s view of the employee experience is also part of the feedback loop. He said Long Lake has seen very high retention across its acquisitions and wants to become “the best place to work in every industry” where it operates. The formula, as he described it, is to give the best people the best tools and the best customers. Once employees have stopped spending 25% or 30% of the day on mundane work, leaving for a competitor would mean taking that work back on. Taubman compared the prospect to “giving up email.”

He also said Long Lake can pay people more because they are more productive and are making more money. In his account, AI-enabled productivity improves retention on both sides: employees are less likely to leave because the work is better, and customers are more likely to stay because response times improve, errors fall, and service quality rises. He cited board reporting, budgeting, and email as areas where errors have gone down in Long Lake’s businesses.

Elad Gil argued that AI roll-ups require three competencies that are rarely assembled in one organization: private-equity-style acquisition capability, AI engineering capability, and change management. Alex Taubman said Long Lake was designed from day one as a cross-functional company combining technology DNA, M&A, and operating transformation.

The early team came through network relationships. Taubman said 100% of Long Lake’s first 20 people were hired through networks where the firm knew them well. He named Palantir, Ramp, Robinhood, and Glean among the kinds of companies from which early technical talent came. Rasmus, Long Lake’s co-founder and CTO, was introduced through an early investor and board member, in a network of people who had known each other for more than 15 years.

Gil emphasized that this is unusual. In his experience, business people often lack deep technical networks, and technologists and business people frequently struggle to hire each other early. That mismatch can weaken founding teams in precisely the kinds of hybrid businesses Long Lake is trying to build.

Taubman said the appeal to engineers was the chance to bring AI “into the real world.” His thesis is that labs and model companies are making extraordinary progress, backed by enormous investment, but there remains a large gap between model capability and real-economy deployment. Many members of Long Lake’s engineering team had previously been founders, CEOs, or CTOs of applied AI companies. Some had learned, in Taubman’s telling, that selling software into services industries is difficult. Long Lake’s alternative is “if you can’t beat them, join them”: become the operator, not merely the vendor.

On the investment side, Taubman said Long Lake’s M&A team includes people from firms such as GTCR, Blackstone, TPG, and H.I.G. He described traditional private equity firms as extraordinary at what they do. The distinction, in his view, is that they are not AI-native. For M&A professionals who believe AI can transform services businesses, he said, there are not many firms that look like Long Lake.

Elad Gil described the announced American Express Global Business Travel transaction as a $6.3 billion take-private and said he believes it may be the first of its kind organized around AI transformation. He called Amex GBT the world’s largest corporate travel platform and a 110-year-old business; Alex Taubman called it a 111-year-old company. Taubman said it began in 1915, when American Express created it to help Travelers Cheques customers get out of Europe during World War I. He also said Carlson Wagonlit, which Amex GBT acquired late last year, was founded in 1876, and that “Wagonlit” refers in French to sleeping cars on trains.

For Taubman, that history matters because these businesses have already lived through a century of technology transitions. He said Long Lake sees Amex GBT as an extraordinary franchise for another century.

Long Lake did not arrive at travel opportunistically, according to Taubman. The firm uses a “prepared mind” approach and keeps a whiteboard of roughly 15 to 20 industries it considers high-value areas of focus. Travel had long been on that list. The characteristics he cited were mission-critical service, high cost of failure, and the fact that most trips are revenue-generating. The customer trust built by Amex GBT over more than 100 years, he said, is “extraordinary.”

Taubman was careful about the limits of what he could say because the transaction involves a public company still going through the process. But he repeated what he said had already been stated publicly: Long Lake’s vision is to double down on Amex GBT’s existing AI transformation strategy. He described the future for Amex GBT’s customer as “your travel counselor with AI superpowers”: faster response times, faster disruption resolution, and better customer outcomes.

That language is consistent with the broader Long Lake operating model as Taubman presents it. Nexus is not described as an abstract automation engine detached from employees. It is meant to sit inside workflows and increase the capacity of the people already serving customers.

Elad Gil contrasted Long Lake’s approach with a familiar private equity pattern: buy a business, add debt, cut costs, hold for a few years, and sell. He suggested Long Lake’s orientation looks closer to a Berkshire Hathaway-style model, with long holding periods and continued investment.

Alex Taubman pointed instead to the scale of the market and to operating-company precedents such as Danaher. He said the relevant opportunity across services is more than $20 trillion in total addressable market and that Long Lake wants to be the market leader in every segment where it operates. Danaher, in his telling, compounded over time by developing a differentiated operating model, first in manufacturing and later in life sciences, that drove better growth, customer satisfaction, employee retention, and productivity. Long Lake wants to follow that pattern in services, with its AI platform as the advantage.

The long-term holding period is not incidental to the AI strategy. Taubman said the work is difficult and cannot be completed in a year, or even in two or three. The transformation compounds over multiple years: better tools help attract better people; better people can be paid more; better-paid, better-equipped employees deliver better customer outcomes; better customer outcomes accelerate growth.

Taubman said he would want to own such a company forever, compound the advantage for decades, and extend it into ancillary areas and service lines. He described Long Lake’s desired role as a long-term owner and partner to world-class services companies.

That positioning is also part of Long Lake’s appeal to sellers, especially founders. Taubman said he grew up in a family of entrepreneurs: his grandfather started a family business that his father and uncle later ran, a 75-year-old company the family recently exited. When he designed Long Lake, he said, he had in mind the principle of building the product he would want to use. If he were selling an asset, he would want the next steward to be a long-term investor in the employees and customers, not simply a financial owner optimizing for a near-term exit.

Elad Gil observed that Long Lake often wins competitive processes, and in some cases appears to be the only buyer to whom sellers want to sell. Alex Taubman said the firm’s message has resonated with business owners and management teams because it combines two things that are rarely paired: long-term permanent capital and deep applied AI engineering expertise.

Taubman argued that AI remains very underpenetrated in real enterprise use cases, estimating penetration at around 1%. He also noted that 99% of businesses in America are small businesses, which generally lack the resources of large companies. Even large companies, he said, are struggling to determine how to drive maximum impact from AI. Against that backdrop, Long Lake’s offer to sellers is not just capital, but a cross-functional team that becomes a partner from day one.

The pitch is unusually hands-on. Taubman said Long Lake engineers may “live in your office for the next two years” helping fix problems. He named members of the engineering team and characterized that embedded support as a strong value proposition for founders and management teams.

Long Lake also encourages rollover equity and alignment from existing shareholders, founders, and management. Gil clarified the point: the owner of a business can receive a stake in the new business so they benefit from the upside created by the transformation. Taubman said Long Lake is open to and encourages that structure because it wants to “win together.” In the firm’s first four verticals or service lines, he said, Long Lake has significant rollover participation from original founders and leaders.

The capital-markets ambition follows from the operating ambition. Taubman said that as Long Lake proves out the model and capital markets better understand it, he hopes the firm’s cost of capital will fall. That would make Long Lake an even more attractive buyer, in his view, because it could pay more. He pointed to Danaher and TransDigm as examples of operators that established a valuable acquisition currency: if a company can operate deals better, and if the broader platform is diversified, faster-growing, and financially stronger, it does not have to lose deals on price.

Alex Taubman said most services founders are already growth-oriented. They built their companies over decades by selling, knocking on doors, and winning local markets. The constraint is not necessarily ambition. It is the pain of scaling labor.

In a labor-intensive services business, Taubman said, 20% growth may require hiring 20% more people. That means finding them, training them, and managing them. It also means that a large share of each incremental revenue dollar goes to additional labor. He described a case where the company might retain only 20 cents on each incremental revenue dollar after labor costs, calling it a high marginal tax rate on growth.

AI changes the psychology and the economics if it allows existing teams to become 30% or 40% more efficient and handle more customers. In Taubman’s account, the organization begins to resemble a software company: it can grow with high incremental margins, invest more in growth, and pay employees more while improving customer service.

That is why he frames the opportunity as positive-sum rather than as a narrow cost-cutting lever. The productivity gain, in his thesis, makes growth less punishing for services companies, gives employees tools that make their jobs less tedious, improves retention, and makes customers happier through faster and more accurate service.

Taubman said CEOs in Long Lake’s partner companies, some of whom have run their businesses for decades, would say they are having the best time of their careers because they are now growing “like a software company” while team members are being paid more and customers are happier. That is the model Long Lake says it wants to bring to larger services franchises: buy or partner with trusted operators, embed AI deeply into their workflows, and use the resulting operating leverage to compound over time.