Feathery's AI assistant makes bespoke proposals an operations problem

For most advisory teams, a genuinely bespoke proposal has been a multi-stage exercise that pulls in client data, planning work, and compliance review long before a client sees it. Feathery, the San Francisco provider of AI-driven data collection, document processing, onboarding, and operational workflow systems for wealth firms, has launched an AI-powered Proposal Generation tool, Wealth Solutions Report reports, that folds client data, planning documents, advisor expertise, and firm-approved templates into a single workspace where Robin, an AI assistant, drafts the proposal and updates it through natural-language chat as team members revise.

Peter Dun, Feathery's co-founder and CEO, says wealth firms already know how to tailor advice to a client's situation, but the process is slow and demands coordination across departments; the goal is to remove that operational drag so firms can deliver more personalization with less manual work.

The launch follows last month's announcement of $30 million in total funding, including a recently completed Series A from Portage Ventures, Index Ventures, Allstate Strategic Ventures, Clocktower Ventures, Erie Strategic Ventures, and Bain Capital Ventures.

The tool's real test is whether a firm can codify its pitch without boiling it down to a template. For most advisory teams, the bottleneck has never been the advice but the calendar that reserves bespoke work for flagship relationships. If Robin compresses that pipeline, the constraint shifts to the quality of the inputs and the advisor's own judgment. The closer a tool moves to client-facing output, however, the more it carries the firm's brand, and the open question is how much automation a client will tolerate before 'personalized' starts to feel templated.

Practice leaders may find the more useful result is the discipline Robin imposes. To get real value from the assistant, a firm must explain its winning proposal precisely enough for a machine to reproduce—which templates work, which data matters, which language closes. The first firms to test that translation will learn quickly whether their best thinking survives being turned into instructions.