Feathery raises $30 million to run the workflows behind financial services
Feathery has raised $30 million, including its Series A, to scale what it calls an AI operating and decisioning system for financial services. Portage Ventures led the round, with Index Ventures, Allstate Strategic Ventures, Clocktower Ventures, Erie Strategic Ventures, and Bain Capital Ventures participating. Founded and led by Peter Dun and Zack Khan, the company already serves more than 300 firms across insurance and wealth management and processes tens of millions of submissions a month. The pitch is aimed at a persistent problem in financial services: client data arrives in fragments across forms, emails, and legacy systems, and turning it into structured, actionable records eats hours of skilled labor at every firm.
The capital lands in a busy corner of enterprise AI, where a wave of startups is promising to automate the back and middle office of regulated finance. Feathery targets a narrow slice on purpose. Its wedge is the unglamorous mechanics of onboarding, account opening, proposal generation, and the data migrations that follow an acquisition. Those workflows are high volume, error-prone, and tightly bound to compliance, which makes them both painful and defensible once a vendor is embedded. We see the round as an endorsement of a focused strategy, capturing the operational plumbing of financial firms rather than competing head-on with horizontal copilots that struggle to earn trust inside regulated environments.
Two systems: an operating layer and a decisioning layer
Feathery splits its product into two parts. The AI Operating System collects and structures client information, synchronizes the major systems of record a firm already runs, and normalizes data across every surface it touches. The goal is a single, clean view of the client that stays current as information flows in from applications, documents, and third parties. On top of that sits the AI Decisioning System, which analyzes the data moving through the platform, surfaces recommendations, and feeds those learnings back into workflows to push automation further over time. The design reflects a sound instinct: clean, unified data has to come before any decisioning layer can be trusted with consequential calls.
Peter Dun framed the opportunity around rising complexity. Financial service firms are dealing with more client data and expectations than ever before, he said, and Feathery turns those challenges into opportunities. The architecture matters because it addresses the sequence most AI projects get wrong. Firms that bolt a model onto messy, siloed records get unreliable output and quiet distrust from the people meant to use it. By owning ingestion and normalization first, Feathery earns the right to make recommendations on data it can vouch for. We would still want to see how much of the decisioning is genuinely autonomous today versus assistive, since that gap defines the real value and the real risk.
The customer list spans wealth, insurance, and brokerage
The named customers give the story weight. On the wealth side, Feathery counts Sequoia Financial, Allworth Financial, and Mission Wealth. In insurance, it lists Tokio Marine, Hiscox, and Banner Life. Among brokers and distributors, it names Baldwin Group, Hilb Group, and Hylant. That spread across three regulated segments is unusual for an early-stage company, and it suggests the underlying data and workflow problem generalizes well beyond any single vertical. Registered investment advisors, insurers, and broker-dealers all wrestle with the same fragmentation, which is why a platform that solves ingestion once can sell the same core into adjacent buyers with modest customization.
One use case stands out for our audience: the data migrations that follow mergers and acquisitions among RIAs and broker-dealers. Consolidation in wealth management has been relentless, and every deal triggers a painful transition of client records between incompatible systems. Feathery positions itself as the tool that absorbs that pain, mapping and normalizing books of business as firms combine. For private-equity-backed roll-ups in particular, that is a recurring, budgeted cost, and a vendor that shrinks it can attach to the deal engine itself. We think this M&A-transition angle is one of Feathery's most durable footholds, because it ties the product to an activity that shows no sign of slowing.
The data network is the real asset
Zack Khan pointed to what the company clearly sees as its moat. With this funding, we are doubling down on products that tap our data network across clients to help firms make faster, more accurate decisions, he said. The phrase cross-client data network deserves attention. As Feathery processes tens of millions of submissions across hundreds of firms, it accumulates patterns about how financial workflows actually run, and that aggregate view can sharpen decisioning in ways no single customer could achieve alone. Network effects of this kind are rare and valuable in enterprise software, since each new client both benefits from and contributes to the shared intelligence.
The same asset carries obligations that buyers must scrutinize. Pooling insight across competing firms in a regulated industry raises immediate questions about data segregation, consent, and what exactly gets learned versus exposed. Feathery will need airtight guarantees that one client's proprietary information never leaks into another's experience, and prospects should demand contractual clarity on how the network is trained and governed. Stephanie Choo of Portage Ventures argued that Feathery has earned the trust of enterprises by solving the real operational and regulatory complexity that defines financial services. That trust is the whole game. A data network is an asset only for as long as customers believe their information is safe inside it.
Where the AI claims meet reality
Enterprise AI is full of decisioning promises that dissolve under load, so a measure of skepticism is warranted. Feathery's strongest claims are operational: collecting, structuring, and syncing data across systems is well-suited to current models and delivers value even when the automation is only partial. The decisioning layer is the harder sell. Recommendations that touch underwriting, suitability, or account approvals sit close to regulated decisions, and firms will rightly keep humans in the loop until accuracy and auditability are proven. The company's language about feeding learnings back into workflows suggests a gradual march toward autonomy, which is the responsible path in an industry where a wrong automated call carries legal weight.
Competition is the other reality check. Feathery operates in a crowded field of startups selling AI operating systems and decisioning platforms to banks and insurers, several of which have raised larger rounds. Its defense rests on depth in specific workflows and the accumulating data network, both of which reward time in market. The $30 million gives it room to deepen those advantages and expand the go-to-market motion, though it is modest against better-capitalized rivals. We would watch retention and expansion within existing accounts as the truest signal, because in workflow software the firms that renew and buy more are the ones whose daily operations genuinely depend on the product.
What CIOs in financial services should take away
For technology leaders at wealth managers, insurers, and brokerages, Feathery is a useful marker of where practical AI value sits right now. The gains are concentrated in data unification and workflow automation, the layers beneath flashier autonomous agents, and that is precisely where regulated firms can adopt with manageable risk. Leaders wrestling with fragmented client data, slow onboarding, or painful post-acquisition migrations should evaluate whether a platform like this compresses cycle times enough to justify displacing incumbent tooling. The customer roster shows the approach works across segments, which lowers the reference risk of being an early adopter in your own category.
Diligence should center on the data network and its governance. The cross-client intelligence is Feathery's most compelling differentiator and its most sensitive liability, so buyers must understand exactly how information is isolated, consented, and used before signing. Leaders should also probe how much decisioning is truly autonomous today, keeping human oversight on anything that brushes a regulated determination. Priced against larger competitors, Feathery is making focus and depth its argument, and for firms whose pain is operational plumbing rather than headline AI, that argument is a strong one. The round gives the company fuel, and the next year of retention numbers will show whether the plumbing has become indispensable.
