Introducing X’s Moonshot for Specialized Professional Intelligence
A new approach to AI architecture for large-scale private data
What if a lawyer’s next question in a deposition could be informed by a live analysis of every previous answer and millions of case documents in real-time? What if a doctor could instantly distill twenty years of a patient’s medical history into a precise, navigable timeline of cardiac risk? What if a company could audit its global expansion in seconds, identifying every regulatory gap between its internal policies and local law?
For the last few years, my team at The Moonshot Factory has been working towards turning “what ifs” like these into reality. We’re creating domain-specific AI architecture for specialized professionals — like lawyers, medical professionals and finance experts — that’s built on the secure and scalable foundations of Google Cloud Platform and powered by Gemini. Our moonshot is to catalyze a fundamental shift in the way professionals, who work with millions of sensitive and private documents, interact with information and harness the productivity benefits of AI.
We’re developing our platform in partnership with a small group of early customers in the legal industry, and look forward to making this architecture more widely available in 2027. We see enormous potential to help lawyers spend less time wrangling data and more time providing high-quality expertise to their clients. One partner recently shared that our platform surfaced crucial documents—found within data that had been analyzed for years—that had never been seen before. Another is using the platform to get near-instant answers for tasks that typically take days of research across terabytes of data.
As we expand and scale support for our early customers, we’re looking for software and machine learning engineers to join us. If you’re inspired to push technical boundaries and want to shape a product from the ground up, check out our open roles here.
Why Specialized Professionals Need Specialized AI
Specialized professionals need AI tools designed and built for the scale, security and rigorous workflow demands of their industries. These professionals need bespoke AI because they work with very large sets of sensitive, private data that public models aren’t trained on. A legal case or a complex medical history contains tens or hundreds of millions of pieces of information scattered across different formats: PDFs, handwritten notes, emails, databases, images and more. These databases are also constantly updated as new information arises. Making vast and constantly-evolving private datasets like these accessible, compliant, and in the formats that these professionals require demands a new approach to both AI and systems architecture.
The living knowledge systems we’re building filter and connect vast amounts of multi-modal data to elevate what matters and highlight the relationships between people, places and events.
Living knowledge systems for large-scale private data
We’re creating “living knowledge systems” for this kind of large-scale private data. These systems aim to filter and connect vast amounts of multi-modal data—from encrypted emails and sensor logs, to handwritten notes and terabytes of video—to create a bespoke living knowledge base for each legal case or patient file.
A bit like a detective's evidence board connecting people, places and events, these living knowledge systems will be constantly updated and highlight the relationships to tell a cohesive story. As a new email arrives, lab results are filed, or regulation changes, the knowledge system will elevate what matters, integrate it and make sense of how it relates to everything else.
We’re pairing these living knowledge systems with advanced reasoning capabilities that integrate specialized workflows that can be queried in real time. Whether it’s preparing depositions, analyzing medical records or piecing together mergers, we’re building tools designed to meet regulatory requirements that search living datasets and deliver precise, accurate answers that cite sources along the way.
Our living knowledge systems will integrate specialized workflows and can be queried in real time.
Join us!
While it’s still early days, we see a huge opportunity to build domain-specific AI, and our team has been making exciting progress on these engineering challenges. If you want to join us, see our open roles here or get in touch here.