Every Side of the Table
Andrew Keller has spent his career in private markets legal work—as an associate at Kirkland & Ellis, in-house at BlackRock and CVC Credit Partners, and now at Covenant. Here's what sitting in every seat taught him about what's broken, what's changing, and why this moment is ripe for disruption.
Andrew Keller still remembers the hours he lost on a confidentiality provision.
He was a mid-level associate at Kirkland & Ellis, grinding through an LP agreement. The provision itself was narrow—a list of parties the LP wanted included. But he kept working it alone, afraid to surface a half-finished question to the senior associate or partner above him. "I was afraid of bothering them," he says, "so I'd sit with a question rather than ask it." The work stretched far longer than the provision was worth. “The irony is that the provision barely moved the needle."
It's a small story. But it captures something that anyone who has done this work will recognize, and it stays with Andrew because he's now seen the same dynamic from every other angle too. As in-house counsel at BlackRock and CVC Credit Partners, he was the client absorbing those hours. At Covenant, the AI law firm purpose-built for private market investors, he's part of the team rebuilding what the model could look like instead.
"If I'd had AI then, I could have gotten my bearings in minutes instead of hours—what was actually at stake, what this LP had accepted on the same point in past deals, what the language meant in plain terms. That context would have told me where the provision sat in the deal. Instead the billable model rewarded the grind."
The view from the other side
The shift from outside counsel to in-house gave Andrew something most lawyers never get: a clear-eyed view of what legal work actually costs, and what it does and doesn't deliver to the people paying for it.
At BlackRock, he was managing outside counsel across one of the most complex private debt funds in the world. Each strategy had its own fund terms, its own quirks, its own negotiating history. But certain things had to stay consistent across the entire platform: confidentiality language, expense provisions, regulatory language that couldn't vary by team or geography. "Keeping that consistent across dozens of strategies and jurisdictions is quite difficult to track by hand," he says.
The more fundamental issue was what happened to negotiating intelligence over time and who actually owned it. Every deal, every side letter, every position taken on a contested term represents hard-won knowledge: what the LP accepted, what it pushed back on, what the market looked like at that moment for that strategy.
He watched the consequences play out firsthand when BlackRock acquired another manager in his strategy area. Over time, that team's people moved on. "We still had to manage the funds and investments that team had made ten years earlier, with all the knowledge of the little intricacies and history of a deal gone. When an investment goes sideways or an LP raises a historical question, context and history matter a lot." The knowledge existed. It was in the documents, the emails, the data. The problem was that without a way to surface it, it was effectively invisible exactly when it mattered.
This is the gap that shapes how Covenant works. The standard LP approach when reviewing a new fund is to redline the prior vintage—Fund III versus Fund IV—and focus on what changed. "That's fine as far as it goes," Andrew says, "but it's narrow. It only looks at one manager, one lineage." Covenant's platform applies intelligence from across a client's entire portfolio to the live negotiation in front of them, surfacing what that LP negotiated three funds ago with a different manager running a similar strategy, making visible the precedent that would otherwise have walked out the door with the lawyer who negotiated it.
"Say the LP won a creative solution on a particular point three years ago. Today nobody remembers it. The PM moved on, the lawyer who negotiated it isn't on this deal. That win is invisible exactly when it would be useful."
The real product was always pattern recognition
"When you bring in the senior partner on a hard point, you're not just paying for raw intelligence. You're paying for their market knowledge, the hundreds or thousands of deals they've seen across a career. The best partners can reach into that experience and pull out a creative solution on the spot, because they've seen the problem before in some other form. That instinct is the product."
Covenant's approach is to institutionalize that instinct directly for the client. The deals in a client's own history combined with the pattern recognition built across Covenant's entire book of private markets work becomes the resource the lawyer draws on in a live negotiation rather than one partner's memory.
The bet
Joining Covenant wasn't a simple decision. But the opportunity to be inside a fundamental shift in how private markets legal work gets done, rather than watching from the outside was hard to ignore.
"The chance to come in almost at the ground floor, to help build something. And then with AI—looking back even just a year ago at where the industry was, it was so nascent compared to what we're experiencing today. To be at the frontier of where the legal industry is headed was really exciting. It was a risk. But it was the right risk."
If he were starting at BlackRock today, Andrew says he'd approach it entirely differently. "I'd be automating from day one, building agents and workflows for the repeatable work, going after the data aggressively, making sure the institutional knowledge the team was building didn't stay trapped in people's heads. The lean in-house team running on agents and its own data is what legal looks like in a few years. I'd want to be the one building it rather than catching up to it."
Who's actually nervous, and why
When Andrew tells former colleagues what he's doing, the reaction is telling. The most common response—half joking, half not—is some version of: you're coming for my job.
He thinks the anxiety looks different depending on where you sit. For junior associates, it's structural. The mechanical work that fills early-career hours like subscription document reviews, LP transfers, redlining off a senior lawyer's markup is exactly what AI agents handle today and handle faster every quarter. "What a hundred associates were doing last year can be done by 20 associates running strong agents." That shifts the economics of the traditional firm in ways that are hard to absorb gradually.
For senior lawyers, the picture is more nuanced and ultimately more interesting. "AI can't replace knowledge and experience. You absolutely need human lawyers with real deal experience overseeing the agents, and experienced humans to interface with clients and counterparties, to negotiate, to read a room." What changes is the leverage model. A partner with senior associates and strong agents can take on significantly more than was ever possible with the traditional structure.
The version of legal that Andrew is building at Covenant—experienced private markets lawyers working with a platform that carries the institutional memory of every deal they've ever touched, delivered at the speed that modern deal-making actually requires is what he wishes had existed when he was on the other side of the table.
"We're an AI-first business. The service we offer happens to be legal. But the starting point is AI and that changes the cost structure, and the trajectory of what we can do."
