The mandate
A traditional, entrepreneur-led search fund pursuing its first platform acquisition in lower-middle-market business services, backed by a syndicate of search-fund investors.
The operating constraint
The searcher must run proprietary outreach and intermediated deal flow simultaneously: parsing CIMs, teasers, and broker packages, normalizing SDE and adjusted EBITDA across inconsistent presentations, and keeping investors informed. Critical context sits in inboxes, spreadsheets, and call notes. Each new opportunity restarts the analysis, and questions of customer concentration, owner dependence, and working-capital seasonality surface late — often after an IOI has already been submitted.
The multi-agent architecture
A planner agent decomposes each opportunity into a structured diligence graph. A sourcing agent applies layout-aware parsing to CIMs and broker packages, resolving entities and extracting financial tables into a schema-validated company profile. A screening agent scores each target against the investment thesis — scale, margin durability, recurring revenue, concentration — and records the rationale with citation-level provenance. A diligence agent reconciles reported figures against the QoE workstream, models a normalized working-capital peg, and sizes an SBA 7(a) or senior-plus-seller-note capital stack on a deterministic calculation engine. An adversarial critic agent challenges add-backs and unsupported claims before a memo editor assembles the LOI rationale and acquisition memorandum.
The work product
- A searchable, schema-consistent pipeline spanning proprietary and intermediated deal flow
- Source-linked target profiles with normalized SDE, adjusted EBITDA, and thesis-fit scoring
- A QoE-aligned diligence request list covering concentration, working capital, and add-back support
- Draft IOI/LOI rationale, capital-stack scenarios, and a searcher investor update
The operating shift
The intended result is a search process with persistent semantic memory: the searcher can revisit why a target cleared the screen, see precisely what changed between the teaser and the data room, and concentrate on owner conversations and deal structuring — earnouts, rollover equity, and step-up economics. Investment judgment, valuation, and all outreach remain with the searcher.
Human-in-the-loop governance
Human-in-the-loop checkpoints govern every consequential step: the searcher approves screening decisions, owner and broker outreach, IOI/LOI terms, and all investor communications.