Kirkland commits $500M to build AI platform

Kirkland and Ellis is committing five hundred million dollars to build its own AI platform, signaling a shift from using external tools to owning the systems that deliver legal work. This move ties directly to value based pricing and long term control over how services are produced and sold. For firm leaders, the implication is clear. Proprietary workflows and institutional knowledge are becoming competitive assets. At the same time, new guidance from CISA and evolving regulations like Connecticut's AI hiring law are raising the stakes on how AI is deployed, governed, and explained. We also cover Zscaler's move into AI access visibility, continued consolidation from Ascend, and why financing and process context are becoming part of the competitive landscape. Learn more at https://crestvale.io

Transcript

SPEAKER_00 0:00

The biggest law firm in the world just decided software, not lawyers, will define its future, and it is willing to cut partner profits now to get there. That changes how every firm competes, prices, and grows. This is the Crestvale Newsroom Daily Podcast. Kirkland and Ellis is committing $500 million to build its own AI platform. Not to experiment, not to layer on top of vendor tools, to own the system that runs the firm. The strategy is simple. Shared AI tools raise the floor. Proprietary systems win the highest value work. So they are building a closed platform that captures how their lawyers actually operate. Around 250 lawyers, including 100 partners, are feeding their workflows into the system. The goal is to turn judgment and experience into repeatable execution. That is a big shift. It moves the firm from selling time to delivering outcomes that can scale, and it ties directly to pricing. Kirkland is aligning this investment with value-based billing. If you can standardize how work gets done, you can price on results instead of ours. They are also making a clear trade. This is being funded out of current revenue. That means lower partner payouts now in exchange for a system that could dominate later. This is not about efficiency. It is about control. And control is what lets them decide how work is delivered, how it is priced, and how margins expand over time. Why this matters is straightforward. If top firms are turning expertise into software, then generic AI tools stop being a differentiator. They become table stakes. If you rely only on off-the-shelf tools, you risk looking interchangeable, and interchangeable work gets priced down. The move here is to start capturing your own workflows now. Document how work actually happens. Identify repeatable patterns. Build internal systems, even simple ones, that reflect how your firm operates. Because once the leaders lock this in, catching up gets expensive very quickly. Now, while firms race to build AI capability, regulators are drawing lines around how it should be used. The Cybersecurity and Infrastructure Security Agency just issued guidance on agentic AI. And the message is blunt. Autonomous systems expand your risk surface fast. The biggest issue is not bad models. It is over permissioned agents. When one agent can move across systems with broad access, a single failure can cascade. There is also the problem of unpredictability. Agents can follow instructions in ways that still create unsafe outcomes. Prompt injection and data poisoning are now operational risks. CSA is pushing a disciplined rollout. Start with narrow access. Increase autonomy gradually. Monitor everything. If you skip that, small pilots turn into firm wide exposure. For firms already experimenting with agents, this is a reset. Access control and audit logs are not optional. They are the foundation. Meanwhile, Connecticut is tightening rules on AI and hiring. The new law requires firms to disclose when automated systems are used in decisions. That includes hiring, promotion, discipline, and termination. And the disclosure has to be clear what tool was used, what data it processed, how it evaluated the person. This goes into effect in October 2027. The important part is liability. Using AI does not protect you from discrimination claims. You still own the outcome. Most firms will find that their existing HR tools fall under this. Applicant tracking systems, resume screening, performance scoring. If you cannot explain how those systems work, you have a problem. This turns AI and HR from a convenience into a compliance and litigation issue. The work starts now with audits and vendor reviews, and on the infrastructure side, ZidsCollar is buying symmetry systems to tackle a new kind of risk. AI agents are starting to act like privileged insiders. Traditional access controls were built for humans. They do not map well to autonomous systems moving across apps and data. The combined platform focuses on visibility. It builds an access graph that shows how every identity interacts with systems in real time. That allows for faster detection when something goes wrong. It also lets firms tighten access based on actual behavior, not static roles. The takeaway is simple. If you cannot see what your AI is doing across your environment, you cannot control it. Here is what else is worth knowing today. Ascend continues its acquisition run, adding another 100-plus staff firm. The private equity rollup model is accelerating, and independent firms are now competing with capital at scale. IBM is committing $5 billion to open source security. The signal is that AI-driven vulnerability discovery will outpace patching, which raises the bar for firms relying on open source components. CapChase is embedding financing directly into sales workflows. That means clients will increasingly expect flexible payment options as part of how services are sold. Solonus is pushing deeper into process-level digital twins. The point is that AI without operational context produces unreliable results at scale. CrowdStrike helped dismantle a botnet targeting developers. It reinforces that development environments are now a primary attack surface. Before we close out, here is a quick look at where markets landed. Equities closed higher in the previous session, with both SPY and QQ moving up together. The 10-year yield moved lower. In commodities, gold finished higher while oil pulled back. Bitcoin also declined, giving up some recent ground. Here is the takeaway. Start treating your firm's knowledge and workflows as assets to capture and control, not just expertise to deploy. Tomorrow we are watching how firms start packaging internal AI systems into client facing products and what that does to pricing models. If this was useful, follow the Crestvale Newsroom Daily Podcast so you don't miss it. Thanks for listening.