From a chat function to a deeper integration

PXR's use of AI did not begin with a grand strategy. "We started with a chat function, and we still use that heavily," the partners note. But the way the firm works with AI has been changing quickly.

By now the technology is embedded deeply enough in the firm's daily processes that a return to a pre-AI way of working is no longer realistic. Take it away, the partners suggest, and PXR would have a problem.

The way that we use AI is really evolving. We're going much deeper now, and we have to go much deeper, because if AI is supposed to be a part of our product offering to our clients, then we have to integrate it in a very different and much deeper way.

That shift from AI as an internal helper to AI as part of the product itself is what sets the firm's current work apart.

Why due diligence is the right place to start

Due diligence has long been one of the most labour intensive parts of any transaction. For years, large teams would work through a target company's data room line by line, spending an enormous number of hours reviewing contracts, hunting for irregularities and flagging clauses that could create legal risk.

As dealmaking moved into the venture and startup world, that model came under pressure. Investors are careful with the resources they put into legal work, so the market adapted with lighter reviews that covered only the most critical points. The approach kept costs in check, while leaving many documents unread and creating blind spots that could later surface for either side of a deal.

Due diligence is, in Peter's words, "a straightforward case for AI." It is document-heavy, repetitive and time-pressured, which is exactly the kind of work that has traditionally consumed junior lawyers' nights and weekends.

How much fun did you have doing a due diligence in the past? Yeah, it wasn't the most fun exercise to do. Done well, they argue, AI does not just make the work faster. It's not only better. It's fun for us.

Why general-purpose models aren't enough

The temptation is to reach for an off-the-shelf model and point it at the data room. PXR tried that and hit a wall.

"We also use the general models," Grisar explains, "but we realised that they are not really made for a due diligence exercise." The problem is scale. Due diligence means "dealing with a huge amount of data, especially unstructured data," and general models simply cannot ingest enough of it. "The input capacity is limited. You can only dump a certain amount of documents and that just doesn't help when you have a data room with maybe a thousand documents."

Purpose and infrastructure, not just power, was the missing ingredient.

Building with GAIA

To close that gap, GAIA built a solution together with PXR. The goal is a tool the firm describes as "a relentless and highly intelligent associate". One that can "do the work of 15 associates with the experience of a senior lawyer."

For the firm, this represents a new product in its own right, a 2.0 version of legal due diligence, and a way to reshape how the service itself is delivered rather than a feature bolted onto an existing process.

The design principle is deceptively simple: feed the system the right intelligence and the right questions, framed from the perspective of an experienced lawyer, and let it work at machine scale. "We can ask hundreds of questions to hundreds of documents in no time, and at a fraction of the cost," the partners say. Just as important, the economics have to work for real clients. The aim is to "deliver the product that we want to deliver, and that the market and the budgets can afford."

Value on both sides of the deal

The payoff, PXR argues, is not only speed. A thorough, AI-assisted review gives investors more confidence in their investment and it gives target companies something useful too.

The depth of that review has changed what a good report looks like. In earlier years, findings were often trimmed down, because a long and detailed report signalled expense to a client who knew it reflected hundreds of billed hours. With AI now carrying the weight of the review, PXR has returned to fuller documentation, this time as a lasting resource that clients can look up long after the deal has closed.

It also helps the target companies to improve, because they have a more detailed agenda on what to work on and what to improve in the future. The result is a healthier starting point for the relationship between investor and company.

Companies that already manage their contracts on the GAIA platform tend to arrive at any due diligence better prepared, with room to sort their data and address potential issues proactively before a process even begins. For investors, there is value in a lighter footprint, since a smoother review keeps them from becoming a burden to a company they have yet to invest in. As the feedback from the market has shown, investors can now tell founders something refreshingly simple: "Just dump your data. AI will sort it. You don't need to bring it into a structure beforehand."

Why the right partner matters

For PXR, the difference lies in who is building the technology. "What really makes it different for us in working with GAIA is that it has engineers on the one hand, but also lawyers on the other," the partners say. That combination is rare, and it is essential. "It's just not enough to understand the engineering part. It's the translating of legal processes, legal thinking and legal work into technical products and we haven't found that in any other provider."

The working relationship counts for a great deal as well. PXR points to the fast iterations with GAIA, the ongoing discussions that let the firm shape and adjust the product to its needs as they emerge along the way.

Confidentiality is non-negotiable

Because PXR is a law firm, the stakes around data are higher than for most software users. "We're bound by professional secrecy, so for us this is a very sensitive topic," the partners emphasise. The bar is not just regulatory compliance. "It's not only GDPR compliance. It's a very high level of confidentiality and data security that we need to ensure, simply because we are bound by it."

Looking ahead

PXR's message is clear: AI in legal services is no longer an experiment at the margins. Built with the right partner, purpose-designed for the work, and held to a lawyer's standard of confidentiality, it can become a genuine part of the product a firm offers its clients, starting with the exercise nobody used to look forward to.

Legal due diligence is only the first chapter. The partners already see the same approach reaching into adjacent work such as the disclosure process in M&A, where large volumes of contracts meet tight deadlines and a real need for meaningful access to detail.