Pennant Raises $5M Seed for Proxy Voting Intelligence

Ryan Nowicki Stewart has helped investors cast votes and companies understand the shareholders casting them. Pennant puts those experiences into a software business, giving institutions a place to connect their policies, research and previous decisions.

The New York company announced a $5M Seed round on October 20, 2026, led by Urban Innovation Fund with Y Combinator and individual angels participating. Co-Founder & CEO Ryan Nowicki Stewart and Co-Founder & CTO Tomas Taylor are building AI-assisted corporate-governance software, starting with proxy voting. Pennant says its clients already include pension funds, institutional asset managers and governance and legal advisers.

The commercial opportunity sits in the work surrounding a shareholder decision. An institution wants research that reflects its own policies, people who can review a recommendation, and a record that preserves why a particular vote made sense. Pennant is entering that workflow while established governance providers are also investing in more customized research and voting policies.

What Pennant's $5M Seed finances

Pennant describes its ambition as a world model for corporate governance. The practical starting point is a platform connecting institutional research, voting guidelines and prior decisions, with people retaining a review role. The phrase describes the company's direction; the announcement provides no independent technical benchmark establishing a particular model architecture or level of accuracy.

Urban Innovation Fund led the round, with Y Combinator participating alongside strategic angels. The lead firm's stated mandate combines Seed capital with support for startups in regulated industries, including fintech and regtech. That is a useful fit for a company selling into institutional governance, where understanding the customer's operating environment belongs in product development and the sales process.

Pennant's announcement identifies the angels through current or former organizational affiliations. It expressly says they invested in their individual capacities, and those affiliations do not imply institutional investment or endorsement. The named employers therefore should not be read as additional corporate investors in this financing.

Pennant has not disclosed a valuation, a detailed spending allocation or a reconciled cumulative funding figure in the primary materials reviewed. Its current hiring supplies a more concrete view of the work ahead: the company is recruiting in New York for product engineering, data and AI engineering, and applied-AI and GTM fellow roles. Those openings put engineering and commercial development alongside the broader governance ambition.

How policy becomes a decision people can revisit

Pennant's product site describes a workspace bringing meetings, filings, proposals and voting outcomes into a source-linked review process. It applies custom voting guidelines to recommendations, gives reviewers a rationale to examine, and retains comments, approvals and overrides. An institution can therefore evaluate the proposed action alongside the policy and evidence associated with it, according to the company's product description.

The override is an important product choice. Investment organizations can encounter a case where applying a general rule mechanically would miss something a reviewer considers material. Keeping the exception and its explanation makes that human intervention part of the working record. A later reviewer can examine the prior decision's basis rather than rely on a colleague remembering the conversation.

Institutional memory has to remain useful as people, policies and portfolio companies change. A buyer evaluating Pennant would need to examine how a prior rationale stays attached to its original circumstances, so a colleague can distinguish a recurring policy from an exception made for a particular case. The record needs enough context to support a new judgment.

Customer evidence remains selective. Pennant reports serving pension funds, asset managers and governance and legal advisers, while its Y Combinator profile describes paying enterprise customers. Publicly reviewed materials do not identify those contracts or disclose revenue and independently measured productivity gains. Institutional adoption is the company's reported starting point for expansion.

Why the founders' different paths matter

Ryan Nowicki Stewart and Tomas Taylor met as Cornell undergraduates and later worked on the same BlackRock desk building institutional technology. Their official biographies describe a subsequent division of experience that is particularly relevant to Pennant's chosen market. The company brings the work of using governance tools together with the work of building enterprise software.

Ryan Nowicki Stewart went on to cast proxy votes at State Street and advise public companies on shareholder engagement at PJT Partners. That places investor-side decisions and issuer-side preparation within his professional background. The commercial implication is that Pennant can approach institutions and their advisers with an understanding of how each party encounters the same vote, while still having to earn each buyer's trust separately.

Tomas Taylor's later work included founding Mindware, where he built enterprise-agent infrastructure. His experience adds a different concern to the governance assignment: how software moves through workflows and remains useful to people doing consequential work. Pennant's emphasis on sources, review and recorded rationale makes that connection visible in its product description.

For a stewardship lead, the handoff between technology and voting judgment is a daily assignment. A useful research system has to make room for the person responsible for departing from its recommendation and the colleague who later needs to understand that choice. The founders' combined experience gives Pennant a specific starting point for designing that handoff.

Customization is becoming a competitive requirement

Pennant is entering a market where established providers are making their own investments in institutional choice. In a September 15, 2026 announcement, Glass Lewis described a multi-perspective research model planned to begin in September 2027, alongside work helping clients adopt custom voting policies. Glass Lewis also said benchmark guidelines and Proxy Paper research would remain available for the 2027 proxy season.

That timetable matters because buyers are comparing a developing startup with providers that continue to supply research and change their products. Pennant's opportunity depends on the particular work it can improve for an institution: organizing evidence, applying policy, coordinating reviewers and keeping a usable account of exceptions. A general claim that AI is arriving in governance leaves those purchasing decisions unexplained.

For public companies and advisers, more varied investor policies also make shareholder preparation more specific. Understanding how a particular owner approaches a proposal becomes relevant to the discussion before a meeting. Pennant's stated customer base spans that side of the process as well as the investors deciding their votes.

The New York hiring effort now has to support those connected workflows as the company grows. The next analyst opening a prior vote will need to find the policy, evidence and explanation that made the decision intelligible to the person who recorded it.