Claude for Deal Sourcing & CIM Screening

What this workshop covers Deal sourcing is not a volume problem. It is an attention problem. Every mid-market fund sees more inbound than it can read: broker teasers, banker CIMs, listings, off-market whispers. It burns its scarcest resource, senior attention, saying no slowly to deals that should have died in the first ten minutes. This is a 1-hour hands-on session showing PE and VC partners, principals, and analysts how to run their entire front-end deal engine, sourcing and CIM screening, using Claude AI agents inside Claude Cowork. Two prompts, each driving its own set of agents, do the work: one finds and scores targets, one screens and analyzes every CIM that lands. We will set up Claude Cowork from scratch, so even if you have never used it before, your team will leave running both engines live on your own pipeline. Limited slots available. First come, first served. Section 1: The Deal Sourcing Engine One prompt and a set of agents that runs the full outbound motion inside Claude. You give it your mandate: sectors, geographies, deal size, ideal target profile, exclusions. It then searches public sources, identifies founder-owned and family-owned businesses that fit, reads the acquisition signals that matter (succession gaps, professionalising hires, competitor exits, sector consolidation), scores every target on a consistent rubric, and returns a ranked, formatted watchlist. It then drafts the initial due-diligence summaries and the partner-approval note, held as drafts for a human to review, never sent. Section 2: The CIM Screening & Analysis Engine One prompt and a set of agents that runs the full inbound motion inside Claude. Upload a CIM in any format and the agents extract the business description, financials, team, market size, and deal terms in under two minutes; score the deal against your thesis on a consistent rubric; surface the red flags that kill deals at IC, such as customer concentration, margin deterioration, cap table issues, and missing financials; auto-draft a first-look IC memo structured to your fund's template; and draft the partner notification the moment a deal clears screening. Every verdict is a draft; the advance decision is always yours. Section 3: Building the Skills How to turn both engines into permanent Claude Skills, one for deal sourcing and one for CIM screening, so the workflow is not a prompt you paste each time but a capability your whole team invokes the same way, every time. We build both skills live, so every analyst screens and scores to the identical standard. Section 4: Root Files (the moat) The standing documents every session reads first: your thesis written as explicit rules (deal-breakers as hard PASS conditions), your pass-rationale log (why you really passed on recent deals, your revealed preferences, often stricter than your stated ones), and your house screening and watchlist formats. Collection and code are commodities; this context is what makes your engine yours, and no competitor can replicate it. Section 5: Scheduling How to put the engine on autopilot inside Claude, scheduling the sourcing run and the screening cycle to fire on their own cadence, so a fresh watchlist and a cleared inbound queue are waiting for you each morning without anyone pressing a button. What you leave with Both engines (the sourcing prompt and the CIM screening prompt with their agents), plus two working Claude Skills, your root files, and a schedule, all ready to run tomorrow on your own pipeline. 30-minute Q&A with Ayub. Who this is for PE and VC partners, principals, analysts, and investor relations teams at funds managing $50M to $5B+ who want to stop spending senior attention on deals that never deserved it, and run a repeatable front-end engine that sources, screens, and scores to the same standard every time.