For years, artificial intelligence in financial services was largely an exercise in making information easier to access. Software could summarize a report, extract figures from a spreadsheet, or help analysts draft an email. Useful capabilities, certainly, but the actual work still rested with the people using the tools.
That distinction is beginning to change.
A new generation of financial technology companies is developing agentic AI: systems designed to carry out complex, multi-step tasks, interact with business applications, and produce finished work for human review.
In finance, that could mean preparing an investment banking model, reconciling accounting records, investigating suspicious transactions, or allowing an AI agent to initiate a controlled payment.
The opportunity is substantial, but so are the expectations. Financial institutions cannot simply trade accuracy and accountability for speed. Any technology handling sensitive financial workflows must demonstrate reliability, clear audit trails, and appropriate human oversight.
Six US startups illustrate how this next stage of financial automation is taking shape.
1. Rogo — Rebuilding the Investment Banking Workflow
New York | $160 million Series D | Reported valuation: $2 billion
Investment banking is built around research, modeling, presentations, and transaction analysis. Much of that work is essential, but preparing the underlying documents can consume a significant portion of an analyst's day.
Rogo is addressing that problem with an AI platform built specifically for investment bankers and other financial professionals.
Its agent, Felix, operates within familiar applications such as Excel, PowerPoint, and Word. Rather than functioning solely as a research assistant, it is designed to support deal screening, financial documentation, buyer outreach, and due diligence.
What makes Rogo different?
Many enterprise AI products focus on locating information or summarizing existing documents. Rogo is moving further into the production of the deliverables that financial institutions use in their daily work.
That distinction matters. An analyst does not simply need to understand a company's financial statements; they may also need to transform that information into a valuation model, investment memorandum, or presentation.
Rogo has reported adoption by more than 35,000 professionals across over 250 institutions.
Funding and market momentum
In April 2026, Rogo raised $160 million in Series D funding, led by Kleiner Perkins, at a reported valuation of $2 billion.
Its reported annual recurring revenue has also grown beyond $50 million, suggesting increasing demand for specialized financial AI tools.
The opportunity is compelling, although a valuation of roughly 40 times reported revenue creates significant expectations for continued growth.
Why it matters: Rogo illustrates a shift from AI-assisted research toward AI-supported execution of investment banking workflows.
2. Basis — Bringing Autonomous Workflows to Accounting Firms
New York | $100 million Series B | Reported valuation: $1.15 billion
Accounting firms face a persistent operational challenge: complex workloads, demanding deadlines, and limited capacity to scale their teams efficiently.
Traditional automation has helped with individual tasks, but many accounting processes still require substantial manual effort.
Basis approaches the problem by developing specialized AI agents for accounting, tax, and audit workflows.
Instead of offering a general-purpose chatbot, its technology is designed to handle extended assignments and return completed work for professional review.
What makes Basis different?
Its central innovation is the use of long-horizon agents—AI systems capable of working through larger assignments over extended periods.
These agents are intended to perform more than isolated actions. They can work through multiple stages of an accounting process while reducing the need for continuous human intervention.
In one company demonstration, Basis reported that an agent completed a partnership tax return, Form 1065, without assistance.
That demonstration is notable, although broader independently verified results will be important in assessing real-world reliability.
Funding and market momentum
Basis raised $100 million in Series B funding in February 2026, led by Accel, at a reported valuation of $1.15 billion.
Its investors include GV and Khosla Ventures.
The company and its investors have cited potential efficiency gains of 20–50%, although those figures should be treated as reported estimates rather than established industry benchmarks.
Why it matters: Basis represents a move toward accounting systems that can complete substantial portions of professional workflows, rather than merely assist with individual calculations or documents.
3. Rillet — An AI-Native Alternative to Traditional ERP Systems
San Francisco Bay Area | $100 million Series C | Reported valuation: $1 billion
Enterprise resource planning systems remain central to financial operations. They record transactions, maintain ledgers, and support reporting, but finance teams often depend on additional spreadsheets and applications to complete reconciliations and analysis.
Rillet is challenging that model with an AI-native financial platform.
Its approach brings the general ledger and AI agents into the same environment, allowing automated workflows to operate directly against the underlying financial records.
What makes Rillet different?
Many AI finance tools are built as additional layers on top of existing ERP platforms.
Rillet takes a different approach: it combines the core financial system with automation capabilities from the outset.
That architecture is intended to give AI agents access to the financial context they need while supporting audit trails and human approval where required.
For finance leaders, the potential advantage is straightforward: fewer disconnected systems, less manual reconciliation, and a more integrated financial workflow.
Funding and market momentum
In August 2026, Rillet raised $100 million in Series C funding, led by ICONIQ, at a reported valuation of $1 billion.
Its total reported funding exceeds $200 million, and the company has cited more than 600 customers.
Rillet is also building relationships with major accounting firms as it competes for customers currently using established ERP platforms.
Why it matters: Rillet suggests that AI may not simply become another feature inside financial software. It could influence how the underlying software is designed.
4. Natural — Building Payment Infrastructure for AI Agents
San Francisco | $30 million Series A | More than $40 million raised
AI agents are increasingly capable of researching products, managing business processes, and recommending transactions.
Moving money, however, introduces an entirely different set of requirements.
Conventional payment systems are largely designed around human users who initiate and authorize transactions. Autonomous software needs a different approach to permissions, transaction controls, and financial accountability.
Natural is developing infrastructure specifically for that emerging requirement.
What makes Natural different?
The company provides financial tools designed for software agents, including agent-oriented wallets and controlled mechanisms for receiving, requesting, and transferring funds.
Its vault functionality is particularly significant: funds can be received without giving an AI agent unrestricted authority to send them.
Natural is also developing the infrastructure needed to support agent identity, settlement, and compliance.
The broader objective is to make automated financial transactions possible without abandoning the safeguards expected of conventional payment systems.
Funding and market momentum
Natural raised $30 million in Series A funding in July 2026, led by Forerunner Ventures, bringing its reported total funding above $40 million.
The investment reflects interest in financial infrastructure for autonomous software, although commercial adoption remains at an early stage.
Natural has not publicly established a transaction-volume track record in the supplied research, making real-world usage an important metric to watch.
Why it matters: If AI agents are to carry out commercial transactions, financial infrastructure must support controlled, auditable machine-initiated payments.
5. Bretton AI — Automating Financial Crime Investigations
San Francisco | $75 million Series B | Financial crime and compliance
Financial crime prevention is one of the most operationally demanding areas of banking.
Institutions must verify customers, investigate suspicious activity, evaluate sanctions exposure, and maintain records that can withstand regulatory scrutiny.
These processes require analysts to collect evidence from multiple systems, apply established rules, and document their conclusions.
Bretton AI, formerly Greenlite AI, is developing agents to automate significant portions of that investigative work.
What makes Bretton AI different?
Bretton AI focuses on compliance-sensitive workflows, including:
Know Your Customer (KYC) and Know Your Business (KYB) reviews
Anti-money laundering investigations
Sanctions screening and investigation
Ongoing customer and transaction monitoring
The company describes its approach as Trust Infrastructure, combining automation with regulatory guidance, model risk management, testing, and explainable outputs.
Its reported customers include Robinhood, Mercury, Gusto, and several banking institutions.
For regulated businesses, the value of this approach is not simply completing investigations faster. It is doing so while preserving the evidence and accountability necessary for review.
Funding and market momentum
Bretton AI raised $75 million in Series B funding in February 2026, led by Sapphire Ventures.
The company reported completing approximately 1.2 million investigations and growing its average contract value to $201,000.
These company-reported figures indicate commercial momentum, although operational performance will continue to depend on the quality of the underlying investigations and controls.
Why it matters: Bretton AI demonstrates how agentic technology may become useful in regulated finance when automation is paired with explainability and documented oversight.
6. Fragment — Financial Infrastructure for the AI Software Economy
New York | $12 million Series A | $23 million total funding
As software companies incorporate financial capabilities into their products, they face an increasingly complex engineering challenge.
Billing, balances, transaction records, and financial reconciliation must operate accurately even when triggered by automated workflows.
Building those capabilities from scratch requires significant technical expertise, particularly when financial correctness is non-negotiable.
Fragment is addressing that challenge by providing financial infrastructure for developers building fintech applications and AI-powered services.
What makes Fragment different?
Fragment focuses on the underlying financial systems rather than the end-user AI experience.
Its infrastructure is designed to help developers manage financial operations without building their own money-management foundations.
The company places particular emphasis on financial correctness and preventing invalid transactions.
Fragment has made strong claims about the reliability of its infrastructure. Those claims remain important to validate as deployment grows and systems encounter more complex production workloads.
Funding and market momentum
In October 2026, Fragment announced $12 million in Series A funding, led by FirstMark, bringing its total reported funding to $23 million.
The company has identified Ramp, BILL, and Whop among its customers.
Why it matters: As AI systems become capable of initiating financial actions, the demand for dependable underlying transaction infrastructure may become as important as the intelligence of the agents themselves.
What These Six Startups Reveal About the Future of Finance
Despite operating in different segments of the financial industry, these companies share three important characteristics.
1. The value is shifting from information to completed work
The first wave of generative AI helped people find information, generate drafts, and answer questions.
The next wave is increasingly focused on producing operational outcomes.
Rogo helps generate banking deliverables. Basis tackles accounting assignments. Bretton AI performs investigative workflows.
The competitive advantage is moving toward measurable results, not simply conversational capabilities.
2. Trust and control are becoming competitive differentiators
Financial institutions operate under strict requirements for accuracy, data security, and accountability.
As AI agents take on more responsibility, auditability, permissions, human approval, and explainable decisions become essential product capabilities.
The companies that can demonstrate these controls consistently may have a significant advantage over solutions optimized primarily for speed.
3. The next generation of financial infrastructure is taking shape
Rillet, Natural, and Fragment represent a broader architectural shift.
Instead of adding AI to existing applications, they are developing systems in which automated workflows are part of the underlying financial infrastructure.
That approach could create new possibilities for enterprise finance, autonomous payments, and software-driven commercial activity.
It also introduces new challenges in transaction safety, operational resilience, and regulatory oversight.
What to Watch Next
Investment has accelerated, but funding alone does not establish long-term success.
Several developments deserve particular attention:
Commercial adoption: Can emerging platforms translate early customer interest into sustained, measurable usage?
Operational reliability: Can AI agents consistently perform complex financial work while maintaining accuracy and auditability?
Competitive pressure: How will established providers in ERP, payments, and financial software respond as agentic capabilities become more widely available?
Natural's transaction volumes, Rillet's expansion into additional industries, and the competitive response from companies such as Stripe will be particularly useful indicators.
Final Perspective
The most consequential change in financial AI may not be that software becomes better at answering questions. It may be that software becomes capable of performing more of the work those questions were intended to support.
For banks, accounting firms, fintech companies, and enterprise finance teams, that creates an opportunity to rethink how work is organized, verified, and delivered.
The companies profiled here are approaching that opportunity from different directions. Some already report meaningful commercial traction; others are investing ahead of markets that have yet to mature.
Their long-term success will depend less on how impressive their demonstrations appear and more on whether they can deliver accurate, secure, and economically valuable outcomes at scale.
The next chapter of AI in finance will be defined not only by what agents can do, but by what institutions can reliably trust them to do.
Editorial note: Funding, valuations, customer figures, and performance claims reflect information reported in the October 16, 2026 research brief. Company-reported metrics are not independently verified in this article. This publication is for informational purposes only and does not constitute investment advice.
