Rogo dispatching agents on someone's behalf to handle the action items from a client call will be an eye opening moment for a lot of ppl. It closes the loop on: conversation with client -> materials created based on client convo and incorporating all the user/firm context & fin data Rogo has -> send materials back to client. All without a human in the loop (or can keep in the loop to start). It's not easy to close the loop on entire worksteams within large enterprises. We're spending a whole lot of time ensuring we do this more and more to get to measurable outcomes for clients. It's a fun, but very hard, problem to solve in reality!

Excellent product marketers will be one of the hardest roles to hire for in AI-native companies over the next 12 months. There’s a very small pool of folks that can: 1) tell a story effectively for a product that can do almost anything 2) enable reps and own consistent GTM messaging when the product changes daily There’s no existing playbook for either. PMMs who can light the path here will be as high-leverage as they come.

Tastes delicious

Fired up to have you @bicacan ! We're looking for other great product folks to come build/scale @RogoAI with us. DM if this piques your interest.

This is an extremely well-articulated essay on moat building at the application layer. We at Rogo very much agree with Aatish's thesis here.

A visualization of how my Rogo Intelligence has grown over the last month. Rogo builds and maintains a profile for each entity. It pulls from your email, CRM, files, and meeting notes, plus third-party data, your activity on the platform, and its own research. Those profiles capture high-value insights that wouldn’t live within the rigid schema of a CRM. For example, when a sponsor bids $500m for a company, that purchase price ends up in the banker’s CRM. What probably doesn’t make its way to the CRM is the sponsor’s valuation rationale. Why couldn’t they stretch to $600m in a competitive process? What about the business gave them pause? The banking team could undoubtedly tell me, but the CRM usually can’t. Rogo Intelligence solves that. The entity profiles usually contain two types of info as a result: (1) Client-defined: Clients tell Rogo what to track for each entity. Many investment firms have already codified the highest signal datapoints in their evaluation process. Rogo puts collecting that data on autopilot. (2) Model-inferred: LLMs are exceptionally good at synthesizing massive amounts of information and interactions to unearth patterns that people would rarely observe themselves. By allowing the model to populate each profile with information it deems relevant to you, you take full advantage of the data at your disposal. The combination establishes Intelligence that can be searched and proactively surfaced to make you smarter on the next deal or investment.

Deal Room is Rogo's foundation for autonomous execution across M&A, capital markets issuances and financings, wealth management prospecting and client onboarding, and corporate and commercial banking loan origination, underwriting, and monitoring. I’m most excited about Deal Room because it creates leverage in a place most people are not applying AI today: communication and coordination. Most AI use cases focus on thought partnership and material creation. But that might represent only 30% of an actual deal process. The rest is managing inboxes, coordinating teams, updating trackers, chasing information, and keeping systems of record current. A few examples of what Deal Room enables: 1/ A PE associate receives eight diligence trackers from three third-party providers. Rogo monitors Outlook, saves each tracker as it arrives, consolidates the relevant tabs, and prepares the response to the sell-side. The draft stays in the associate’s Outlook until they review and approve sending it. 2/ The sell-side analyst receives that tracker. Rogo incorporates it into the master across all buyers, searches the data room and prior communications to populate every answer it can, and drafts an email to management that clearly identifies what still requires their input. 3/ A wealth advisor instills their sourcing process into Rogo. Rogo monitors prospects, identifies relevant information as it surfaces, and drafts timely outreach for the advisor to review. 4/ Deal Room also coordinates the work between your communications and systems of record. A CRM will capture a sponsor’s final purchase price, but it typically will not capture the valuation rationale that led to the bid. When a sponsor submits an IOI over email, Rogo automatically writes the purchase price to Salesforce or DealCloud. It also captures the valuation rationale from the calls, emails, and materials surrounding the process in Rogo Intelligence, which we launched two weeks ago. This is how Deal Room and Intelligence compound. Deal Room keeps the process and systems coordinated. Intelligence preserves the reasoning so it can inform the next deal automatically. The next time you work with that sponsor, Rogo might tell you: “This sponsor has focused on margin-expansion stories driven by operational AI in the last three deals you showed them. I put together a one-page summary of the relevant opportunity at Titan, based on management commentary and the contents of the data room, for you to review and send.” None of this works without the controls that allow financial institutions to safely delegate more work to agents. Deal Room combines autonomous execution with configurable permissions, full visibility, clear guardrails, and human review at the highest-stakes points.

This is a great example of the product work needed to make horizontal paradigms (ie, memory) maximally useful within a vertical. Some product considerations for Rogo Intelligence: (1) a .md file isn't enough for finance. Memory formed off a CIM in a live process can't inform work on another deal, and it can never cross a compliance wall. That takes lineage and permissioning on every single memory. So we store Intelligence as a structured, revisioned database, with every entry cited back to the chat, file, or email that formed it, permissioned, and tagged for sensitivity. That's also what lets safe entries be shared with broader audiences, manually or automatically. User-level knowledge compounds but not at the same rate as team- or org-level. We've built Intelligence in a way to safely enable the latter. (2) a database enables structured queries for things like precedent transactions, buyer lists, and portfolio monitoring. Plus generic reasoning over all of it for when structured queries would lack context. (3) the Intelligence UI is core to Rogo. Sometimes people want to peruse all of their Intelligence in one place, so we built a first-class home for it, organized around the entities they think in (companies, deals, people, etc.). But what matters more is surfacing that intelligence in visible, material ways at the right moments as they actually work. Showing them the impact of their interactions. (4) we passively write back to connected systems of record (CRMs, file systems) as Intelligence forms. That solves the data hygiene problem that plagues large institutions, and it captures the reasoning those systems were never built to hold. A CRM will have a sponsor's final bid, but maybe not the valuation rationale gathered throughout the process. That reasoning lets Rogo better prep you the next time you do a deal with that sponsor, without having to remember yourself. (5) the Intelligence belongs to the firm, not the model. We've made it portable across your AI tools. Rogo Intelligence will compound with Rogo's next release. Stay tuned... P.S. it actually generalizes well too. I run all product hiring through Rogo. My Intelligence tracks the hundreds of candidates I've got open to-dos with, pulls updates from Gmail, Slack, and Granola, suggests next steps, preps me before calls, and finds warm intros, all without a human in the loop.

Can't kill the Knicks

Spot on, and we think building that layer is the best product gig in the world right now. It'll be one heck of a challenge. If you're up for it, reach out! Roles below.