Would AI Give You a Discount? Maybe — Just Not the Kind You’re Expecting

By now, almost everyone in the legal industry has heard: Kirkland & Ellis announced a $500 million, four-year investment to develop proprietary internal AI tools. The reaction was predictably mixed. With capable off-the-shelf solutions like Harvey and Legora already on the market, many observers questioned why K&E would build from scratch at that scale. Harvey co-founder Gabe Pereyra weighed in on the strategic logic behind the decision here.

But for K&E’s clients, the more pressing question is simpler: will any of this translate into lower bills?

Casey Flaherty argued in Bloomberg Law that the investment is unlikely to “translate to a fundamental change in the Kirkland business model — or in the Big Law model generally — that will lead to immediate discounts for clients.” It’s a fair read of the conventional wisdom. But I think it misses something important.

AI discounts are coming. They just won’t look like what most clients have in mind.

The Discount You Don’t See: Reduced Write-Offs

Standard law firm billing includes a built-in inefficiency: a portion of billable hours is routinely written off, typically because the time wasn’t used productively enough to justify billing. These write-offs are a quiet acknowledgment that not every hour is genuinely valuable to the client.

Properly implemented AI tools — paired with adequate attorney training — can change that calculus. When lawyers work more efficiently, fewer hours need to be written off. The effective cost of delivering the same work decreases. This isn’t a line-item discount, but it is a real reduction in the cost of legal services. The discount is baked into the bill before you ever see it.

The Discount You Don’t Notice: Cleaner End-of-Year Bills

Many clients are familiar with the ritual of year-end billing: an initial invoice arrives at a somewhat alarming number, followed by a negotiated reduction that brings it to something more palatable. The discount feels like a concession; the original number feels like an opening bid.

AI changes the starting point. When efficiency gains compress the original invoice, the need for that end-of-year haircut diminishes. Clients receive a more accurate bill from the outset — one that reflects the actual work rather than being inflated in anticipation of future negotiation. The law firm preserves revenue without padding; the client avoids sticker shock. Both parties get something closer to a fair transaction.

The Discount That Actually Matters: Winning

This is where the real calculus shifts. Most litigation is zero-sum: one side wins and one side loses, and the value of winning almost always dwarfs the legal fees involved. From that perspective, the most valuable thing a law firm can offer its clients isn’t a cheaper invoice — it’s a better outcome.

This is where proprietary AI becomes a genuine differentiator. Off-the-shelf legal AI products serve many firms with broadly similar tools. A firm that builds its own internal AI systems gains something different: the ability to tailor models to its specific practice areas, train on its own institutional knowledge, and control exactly how the technology is deployed. That customization compounds over time into a strategic edge that a $2 billion legal AI platform simply cannot provide to every customer simultaneously.

When your lawyers are better equipped than opposing counsel — when they can find the relevant precedent faster, draft more precisely, and anticipate arguments more accurately — that edge shows up in results. That is, arguably, the most meaningful AI discount any law firm can deliver: not a lower bill, but a better chance of winning.

A Different Kind of Discount

So yes — I think clients will get AI discounts from firms that invest seriously in building their own tools. But the value won’t always arrive as a percentage knocked off an invoice. It may come as fewer hours billed in the first place, cleaner bills that don’t require negotiation, or simply better legal outcomes.

Whether K&E’s $500 million bet pays off remains to be seen. But the clients asking “where’s my discount?” might be asking the wrong question. The more useful question is: what form will the value take? The answer is almost certainly more interesting than a line-item reduction.