In our latest live session, Will Chen, Founder of Mike OSS, joined Platforum9 to explore why open-source AI may represent the next stage of legal technology adoption. Drawing on his background as an Oxford-trained lawyer and former Project Finance associate at Latham & Watkins, Chen shared why he left private practice to build an AI platform designed specifically to address the concerns he repeatedly heard from lawyers.
Why Current Legal AI Solutions Are Falling Short
Chen explained that conversations with lawyers across major firms revealed a common frustration with many existing legal AI platforms. While adoption has accelerated, concerns remain around security, confidentiality, pricing, and flexibility. Many lawyers also highlighted that existing tools are heavily text-focused, offering limited support for the broader range of documents lawyers work with daily, including Excel spreadsheets, PowerPoint presentations and image-based files.
These challenges became the foundation for Mike OSS, an open-source legal AI platform that firms can deploy within their own infrastructure, keeping complete control over sensitive client data.
Security and Data Control Are Becoming Critical
A central theme throughout the discussion was trust.
Chen argued that firms remain cautious about placing confidential client information into third-party AI platforms, particularly where questions exist around data ownership, model training and regulatory compliance.
By allowing firms to host the platform themselves, use their own API keys and choose whichever language models best suit their needs, Mike OSS removes many of the barriers associated with externally hosted AI solutions while giving firms greater flexibility over costs and governance.
The Economics of Legal AI Are Changing
The session also examined how the commercial landscape is evolving.
As AI providers increasingly shift towards usage-based pricing, firms are beginning to reassess whether buying commercial software or building internal solutions offers better long-term value. Chen suggested that this “build versus buy” conversation is becoming increasingly relevant as AI becomes embedded into everyday legal work.
He also noted that firms are paying closer attention to infrastructure costs, procurement decisions, and long-term scalability rather than simply adopting whichever platform enters the market first.
AI Challenges Traditional Legal Business Models
The discussion explored how AI continues to challenge the traditional billable-hour model.
While AI can automate many routine legal tasks, including research, document review and due diligence, many firms still face limited commercial incentives to maximise efficiency under existing billing structures.
Chen believes firms will increasingly need to rethink how they deliver value to clients as AI takes on more of the repetitive work traditionally performed by junior lawyers.
Open Source Is Already Everywhere
One of Chen’s key messages was that open source is not a niche concept.
Much of today’s enterprise software, including systems already used by law firms, is built on open-source technologies. Applying the same model to legal AI, he argued, allows firms of every size to access powerful tools without becoming dependent on expensive proprietary platforms.
By separating the software from the underlying AI models, firms gain the freedom to select the most appropriate models while retaining ownership of their data and technology stack.
Looking Ahead
The session concluded by considering what successful AI adoption will require over the coming years.
Chen believes firms must focus not only on productivity but also on protecting their knowledge, safeguarding client confidentiality, and maintaining control over their own technology environments. As AI adoption matures, issues such as data sovereignty, procurement and security will increasingly shape purchasing decisions.
For Chen, open-source AI represents a more practical, transparent and accessible future, one where firms can innovate without compromising on security or cost.
