Two months ago, we brought on Daniel Kronovet, a machine learning engineer, as Thesis Driven's head of technology. We're now embedding directly with real estate operators and developers to build their AI-powered capital markets infrastructure.
So far, we found the bottleneck in most capital raises isn't deal quality or investor appetite, but operational capacity. A two-person capital markets team at a mid-size sponsor is expected to do the same work that a large fund spreads across a 15-person IR department: researching investors, enriching contact databases, writing and sending outreach, drafting follow-ups, producing marketing materials, building content pipelines, and managing LP communications.
The greatest opportunity lies in compressing those workflows into AI agents that run autonomously with human oversight at critical decision points. Since we don't have the capacity to work with every operator, we're opening the hood on what we're doing in a workshop next week.
The agents we're deploying with these groups do things that would have required a full capital markets team 18 months ago:
Building and enriching investor databases from scratch using Clay and web intelligence tools
Drafting hyper-personalized outreach sequences grounded in each LP's actual mandate and recent activity,
Generating full content strategies from a single fund thesis document (white papers, newsletters, LinkedIn, webinar scripts)
Preparing custom pre-meeting briefings, and writing investor follow-ups that reference specific conversation points
The agent runs a compounding system: sourcing feeds outreach, outreach feeds events, events feed content, content feeds inbound.
The workshop provides an introductory how-to for setting all that up.
About the Workshop
Next Friday, June 26 at noon ET, Paul Stanton will lead a live 2-hour workshop on building and deploying an AI capital markets agent from scratch using Claude, Clay, and Anthropic's MCP toolset. The session is split into two halves: first, you'll build the agent piece by piece (brain, memory, skills, email access, API connections); then, you'll watch it execute a full capital raising campaign through a live case study.
Friday, June 26 | 12:00 – 2:00 PM ET | Zoom | $299
The session is built for GPs, capital raisers, investor relations professionals, and fund managers who want to stop hearing about AI in capital raising and start operating with it. No coding background required.
What the Session Covers
Part One: How to Build the Agent (approx. 60 min)
The brain: writing system instructions and custom prompts that define how the agent thinks about real estate capital raising, investor-deal fit, and communication tone
The memory: loading your fund thesis, track record, deal memos, and prior investor communications so the agent operates with deep knowledge of your specific strategy and voice
The skills: building reusable task-specific prompt templates (researching an investor, scoring mandate fit, drafting a 3-email outreach sequence, writing a newsletter, preparing a pre-meeting briefing) that become the agent's playbook
The communication layer: connecting the agent to email, Slack, and messaging tools so it drafts and stages outreach in your actual inbox
The toolbelt: wiring in Clay for investor enrichment, Google Drive for document access, your CRM for pipeline tracking, and web search for real-time market intelligence through MCP integrations
Part Two: Deploying the Agent (approx. 60 min)
Live Demo 1: Investor research and list building. The agent builds a structured, enriched, prioritized investor universe from a target segment in minutes
Live Demo 2: Personalized outreach at scale. The agent drafts hyper-personalized email sequences for each target, grounded in their specific mandate, recent activity, and connection to the deal thesis
Live Demo 3: Content creation and thought leadership. The agent produces a white paper outline, newsletter edition, and LinkedIn post concepts from a single fund thesis document
Live Demo 4: Meeting prep and investor relations. The agent generates custom pre-meeting briefings and post-meeting follow-ups referencing specific conversation points
Format and Access
The live session runs 2 hours on Zoom. All registered participants receive an invitation to the Thesis Driven private Circle community, where session recordings, agent templates, skills files, prompt libraries, and step-by-step setup guides will be posted within one week. Participants who cannot attend live will get the recording. A Claude Pro or Team account is recommended to follow along in real time but not required.
For any GP running a raise without a full capital markets team, or any capital markets professional who wants to operate at the speed and scale of a team three times their size, this is the session.
Paul Stanton is a Partner at PTB, a boutique investment bank focused on joint ventures, capital raising, and M&A transactions. Paul is a Teacher and Writer at Thesis Driven, developing and growing next-gen real estate & proptech concepts.
