Farsight

Incredibly excited to announce publicly that Deutsche Bank is running Farsight on-premise across Investment Banking and Capital Markets in Europe and the Americas. Press release in the comments with more details. Every bank has its own way of doing things, built up over years of prior work and team standards. A generic AI tool can't fully reflect that. Running inside the bank lets AI use that history while the bank keeps control of it. That's why Farsight is running on-premise, within Deutsche Bank's infrastructure. The bank's data, institutional knowledge, and workflow activity stay in-house, and Deutsche Bank keeps greater control over how that information is accessed and used. Each team's formats, processes, and standards become part of how the work gets produced. Firms that put their knowledge to work inside their own walls will build a data moat no one else can copy. Thank you to the Deutsche Bank team for their partnership and everyone at Farsight who helped make this happen.

Deutsche Bank has deployed Farsight on-premise across Investment Banking and Capital Markets teams in Europe and the Americas. For regulated firms, moving AI into specialized workflows depends on access to data, much of which is sensitive. "Farsight gives our teams a way to apply AI to their own workflows and materials while operating within Deutsche Bank's technology environment. That combination of control and customization is critical to making AI work in a global, regulated enterprise like Deutsche Bank." Read the full announcement in the comments.

We saw this exact same motion in AI for software two years ago, and it's about to repeat itself in AI for finance. The systems that eventually won in software were the ones that learned how you code, not just the ones that produced code. They learned your preferences, your conventions, how your team builds, and the nuance behind a single line of code. The learning and contextualization itself became the product, not the models underneath it. It allowed the outputs to go from generic slop to immediately usable, and it's why companies like Cursor emerged on top. Finance is at the same point now. Every firm has access to the frontier models, but the model alone doesn't know how to apply each firm's tens of years of unique experience they've built. How will the system continue to learn as the firm gains new experiences and evolves its preferences? The best AI solution is one that solves that problem, and it's a big part of what we're building at Farsight. My co-founder Kunal Tangri wrote up a full argument for it, link in the comments.

Junior bankers and investors never learn much from creating the first draft. They learn from the markup that comes back from their VP. That's where the judgment lives. But that markup vanishes the moment the work goes out, so as AI takes over the drafting, firms are quietly losing the one thing their people and their models actually learn from. Lose that, and you lose both how the next generation comes up and how the firm itself remembers what good looks like. Farsight's fixing that. A System of Judgment that keeps the record of how a firm decides, so it compounds instead of walking out the door. The labs will keep making models that can draft anything; teaching one how your firm actually decides is the part only you can do. We're starting with financial services, where the difference is in the margins. Read my co-founder Kunal Tangri’s white paper on it here: https://lnkd.in/ek46Gkwp #SystemofJudgment

The most valuable document in a bank isn't the deck you send the client. It's the marked-up draft you throw away on the way there. Everyone says the last mile for AI is judgment. And everyone's fix is the same one we've heard for 30 years: capture your institutional knowledge, which always just means index your documents and search them. But a finished document is the output of judgment with the judgment removed. You can search every deck a firm ever shipped and learn nothing about how it actually decides. The judgment is in the diff. The edits between the first draft and the final one, where a senior person's taste and relationships become concrete changes to the work. Almost every firm in finance and law throws that signal away, every single day. I wrote about what it would mean to keep it, and why the system that does is the next thing after the system of record. We're building it at Farsight, for financial services first. The System of Judgment. Link in comments.

For years, the industry has competed on access to data. The next frontier is turning that data into client-ready work that reflects 𝗵𝗼𝘄 𝗲𝗮𝗰𝗵 𝗶𝗻𝗱𝗶𝘃𝗶𝗱𝘂𝗮𝗹 𝗳𝗶𝗿𝗺 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝘁𝗵𝗶𝗻𝗸𝘀, 𝗮𝗱𝘃𝗶𝘀𝗲𝘀, 𝗮𝗻𝗱 𝗱𝗲𝗰𝗶𝗱𝗲𝘀. Today we're announcing a strategic partnership and investment with S&P Global Market Intelligence to bring Farsight's AI-powered decision intelligence and deliverable creation into Capital IQ Pro. This lets us scale Farsight's 𝗷𝘂𝗱𝗴𝗺𝗲𝗻𝘁 𝗮𝗻𝗱 𝗼𝗽𝗲𝗿𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗰𝗮𝗽𝗮𝗯𝗶𝗹𝗶𝘁𝗶𝗲𝘀 to S&P's global client base, helping bankers and investors t𝘶𝘳𝘯 𝘵𝘩𝘦𝘪𝘳 𝘧𝘪𝘳𝘮'𝘴 𝘣𝘦𝘴𝘵 𝘥𝘦𝘤𝘪𝘴𝘪𝘰𝘯𝘴 𝘪𝘯𝘵𝘰 𝘥𝘦𝘭𝘪𝘷𝘦𝘳𝘢𝘣𝘭𝘦𝘴 𝘵𝘩𝘢𝘵 𝘳𝘦𝘧𝘭𝘦𝘤𝘵 𝘦𝘹𝘢𝘤𝘵𝘭𝘺 𝘩𝘰𝘸 𝘵𝘩𝘦𝘺 𝘸𝘰𝘳𝘬. Excited for what this means for our team, and for the customers who'll get their hands on it later this year. 𝗦&𝗣 𝗣𝗿𝗲𝘀𝘀 𝗥𝗲𝗹𝗲𝗮𝘀𝗲: https://lnkd.in/eXq465w3