Our Take
Big law is now spending like Kirkland & Ellis to own its data moat, but ownership without measurable client outcome is just expensive infrastructure.
Why it matters
Law firms are pivoting from buying third-party AI tools to building internal platforms that encode institutional knowledge. This shift signals a maturing market where competitive advantage comes from controlling the data, not licensing the model.
Do this week
Enterprise AI buyers: audit your current vendor contracts for exclusivity clauses before committing to proprietary build-vs-buy decisions.
HSF Kramer builds internal AI platform on Microsoft stack
Herbert Smith Freehills Kramer (HSF Kramer) has partnered with Acora, a data and AI specialist, and Microsoft to build what it calls a Sovereign System of Intelligence. The platform runs on Microsoft Azure and Microsoft Foundry, with governance through Microsoft Purview.
The system digitises HSF Kramer's legal expertise by combining internal data, governance, analytics, and AI in a scalable environment. It processes unstructured data from legal documents, spreadsheets, and emails using an applied ontology to surface insights more easily.
David Turner, global CTO at HSF Kramer (hired in 2024), described the system as "a moat" around the firm's internal knowledge. Raj Toora, principal strategist for emerging technology at HSF Kramer, led the Sovereign System of Intelligence development and emphasized the focus on balancing innovation, governance, user adoption, and client expectations.
The build-it-yourself trend is spreading across big law
HSF Kramer's move mirrors Kirkland & Ellis, which announced a $500 million investment in a proprietary AI platform earlier this year, separate from third-party AI tool spending. That spending level signals a shift: major law firms no longer believe generic large language models will deliver competitive advantage. They are betting that institutional knowledge, properly integrated, does.
The timing matters. Law firms face pressure to improve profitability while client budgets stagnate. If an internal system can encode expert knowledge and route work more efficiently, the ROI argument becomes easier to justify internally. The catch is measurable impact. Acora's chief AI officer, Darshna Shah, stated that "law firms don't need more end-user AI tools," but the article does not report deployment metrics, time savings, or client-side adoption rates.
HSF Kramer also selected Legora earlier this year as a third-party AI tool, which suggests the proprietary system is not meant to replace external AI entirely, but to sit alongside it.
Own your data governance before committing to a proprietary build
If you work in enterprise AI procurement for legal services, the HSF Kramer model raises a practical question: can your firm's data infrastructure support this kind of system, and do your vendor contracts allow it?
Building a proprietary platform requires clean, well-governed data and buy-in from lawyers who will feed it. Neither is guaranteed. The HSF Kramer project started as a process-alignment program and evolved into a platform, which suggests incremental commitment rather than big-bang risk. That is the pattern to replicate: pilot on internal data first, measure adoption and outcome, then invest in scale.
Second, audit your current third-party AI tool contracts for exclusivity or non-compete clauses. Some vendors restrict your ability to build competing internal systems. Clarify that before signing a multi-year deal.