Manager Due Diligence at Scale: Agent Systems for the Memo, Not the Decision
Wealth firms do serious DD on dozens of managers a year. Most of the work is artifact-gathering and structured comparison: the right place for agent systems.
A wealth firm that allocates across 50-80 outside managers does serious due diligence. Not the GP-meeting-and-handshake kind. The documented kind, with a memo trail, a follow-up file, and an investment committee that read the artifacts before they voted.
The work that produces that memo is mostly mechanical. Pull the pitch deck, the ADV, the DDQ, the LPA, the most recent letters, the most recent capital account statements, prior IC memos on the same manager or its peers, and any outside research the team has on hand. Read it. Compare it against the firm's standard DD framework. Write the memo. Surface inconsistencies. Surface unknowns. Route the unknowns to the analyst.
The judgment that produces the allocation is not mechanical. That is where the senior team earns their salary. But the work that surrounds the judgment is exactly where agent systems compound.
What the agent actually does
- Artifact ingestion: The agent pulls every document your team has on the manager: pitch deck, DDQ, fund docs, ADV, letters, prior memos, news, references, regulatory filings (FINRA, SEC, FCA depending on jurisdiction). It normalizes the format and stores everything in a single file with provenance.
- Structured DD memo drafting: Your firm has a DD framework. Strategy, team, terms, performance, operations, references, risks. The agent drafts a memo against that framework, with each section cited to source documents. Your analyst reads a draft with receipts.
- Inconsistency flagging: When the pitch deck says one thing and the ADV says another, when the manager's letter contradicts a prior letter, when the reference call notes diverge from the founder's claims, the agent surfaces the discrepancy. Your analyst gets a list of things that need a human eye, not a haystack.
- Peer comparison: The agent pulls prior memos on managers in the same strategy or vintage and produces a structured comparison: terms, fees, performance dispersion, key team, gates, side letters. Your IC reads a memo with context, not in isolation.
- Reference call summarization: Your analyst takes a reference call. The agent transcribes it, extracts the substantive feedback, flags the soft signals (hesitations, redirects, what was not said), and adds it to the memo with the analyst's notes preserved verbatim alongside.
Where the human stays
The investment decision. The recommendation. The IC vote. The relationship with the GP. The judgment about whether the fund is right for this client family at this allocation size at this point in the cycle.
Agents handle artifacts and structured outputs. Humans handle judgment. The line is clean and we hold it.
Why this matters more in wealth firms than anywhere else
Wealth firms have a structural challenge: the universe of managers worth allocating to grows faster than your DD bandwidth. You can hire more analysts, but each analyst's standards and outputs drift unless you systemize. Or you can systemize the artifact-gathering and memo-drafting work, so each analyst's judgment scales without their workload scaling linearly.
The systemization is the agent. The judgment is the analyst. The output is a memo your IC can actually read.
What the deployment looks like
Self-hosted on your infrastructure. Your manager files do not leave your perimeter. The agent reads from your document management system (Box, SharePoint, on-prem, your call), writes to your DD archive, and integrates with your CRM and reporting layer. No third-party cloud. No data leakage. Audit trails for every action.
We build the system around your existing DD framework and your existing data sources. We do not ask you to adopt our framework. We architect agents that operate within yours.
Bandwidth this typically frees
For a $1-3B wealth firm or wealth firm with an active manager research function, agent-supported DD typically reduces memo prep time by 50-70% per manager, with no quality loss when the human-in-the-loop is preserved correctly. Capacity that used to go to gathering and formatting goes to actual research and reference checks.
If this fits your shop
We build custom agent systems for RIAs and wealth firms, with audit trails and human escalation built in. If your DD pipeline is throttled by memo prep instead of by analyst judgment, book a strategy call. We will walk through your DD framework and tell you, honestly, where agents would compound and where they would just add surface area.
Considering agent systems for your firm?
30-minute strategy call. We map your highest-leverage workflows and give a clear build-or-not recommendation. No pitch deck.
Book a Strategy Call