SAN JOSE, Calif., Oct. 21, 2026 (GLOBE NEWSWIRE) — Kognitos, the pioneer of neurosymbolic AI for business and finance automation, today announced Kognitos AI Employees and introduced new capabilities for Kognitos Context Graphs. Together, they let finance and banking teams automate work according to approved procedures while retaining the business context, decisions, and audit trails needed to understand exactly what happened and why.

The launch introduces 61 AI Employee roles backed by 515 standard operating procedures (SOPs), alongside Quill, an AI agent that drafts new SOPs for human approval. It also expands the Context Graph capabilities announced in August with 133 ready-made finance monitors that teams can turn into continuous controls with a single click, and REST and MCP access that allows customers to connect their own AI agents.

Finance and banking leaders have spent years piloting AI agents, but many have struggled to move them into production. An agent that decides each step as it goes can handle the same invoice differently each time, making it hard to explain to an auditor why a particular payment was made. Research examining 15 models across 18 months of releases found that while accuracy improved, reliability barely changed, with even the most accurate agents producing different results from one run to the next.

At the same time, the context behind finance decisions is often fragmented across ERP systems, payment platforms, and the experience of the people doing the work. For example, two vendor records may belong to the same company, while the reason a payment exception was cleared may exist only in the memory of the person who handled it. That missing context becomes a problem when AI is expected to make sense of the transaction.

“Every CFO wants AI in the back office that follows approved procedures rather than making up the process as it goes. They need AI teammates that do the work prescribed in their SOPs, with a record of every action that an auditor can examine,” said Binny Gill, Founder and CEO at Kognitos. “To deliver that, Kognitos has built an English-as-Code platform that executes SOPs deterministically, alongside Context Graphs that help people and agents investigate exceptions using the company’s actual business data and the knowledge its people have accumulated over time. Together, they allow us to automate more work without losing the controls and decision history that finance teams depend on.”

AI Employees: hire a role, not a prompt
Kognitos AI Employees are neurosymbolic AI agents for finance and banking, each assigned to a specific role. Each has a job description, approved SOPs, defined limits on what it can do without human approval, and clear responsibilities for handing work off to colleagues. Every AI Employee also operates within defined rules, ensuring it follows the approved SOP rather than inventing additional steps during execution.

The Kognitos AI Employees include:

  • 61 roles, ready to hire. The roles cover accounts payable (AP), reconciliation, cash application, the month-end close, KYC, and more. Teams choose from roles such as AP Specialist, Reconciliation Accountant, Cash Application Specialist, Close Manager, and KYC & Onboarding Analyst, backed by a library of 515 SOPs.
  • Quill discovers and writes the SOPs. Quill, the Kognitos builder agent, studies a team’s systems and recent transactions, compares this against written policy, and drafts each SOP in plain English. Nothing is published until a person approves it.
  • The English is what runs. AI employees are neurosymbolic: models read documents and requests, and a deterministic engine runs each SOP. Every SOP is compiled once and runs using Kognitos’s patented English-as-Code engine, the same way every time. During execution, AI Employees never add a step that isn’t in the approved SOP.
  • Audit-ready by design. Each run has a journal an auditor can read in English, noting every call to an ERP, bank, or inbox with its inputs and outputs, and the exact SOP version that ran. An auditor can pick any transaction and re-perform it, step by step.
  • When it’s unsure, it asks. When a case falls outside the SOP, the run pauses and asks a person in plain English. If the answer should apply next time, it becomes guidance only after an approver signs it off.
  • Context Graphs: from business questions to continuous controls
  • Kognitos Context Graphs connect financial records, such as bills, vendors, and payments, with the relationships between them and the knowledge employees have accumulated about how the business operates. The Cartographer agent builds this context from connected applications, giving AI Employees and finance teams a more complete understanding of the transactions and exceptions they are working with.

New and expanded capabilities include:

  • 133 ready-made finance monitors. Finance teams can use these monitors across 62 processes in 14 areas to check for issues such as duplicate invoices. A question such as “Are we paying any vendor twice?” can become a continuously running control with a single click, helping identify potential problems before a payment is made.
  • Access for customers’ own AI agents. Context Graphs are accessible through REST and MCP, allowing customers to connect their own agents and work from the same business context, subject to the caller’s permissions.
  • Business knowledge alongside system data. Context Graphs bring together information from connected applications, relationships identified by Cartographer and knowledge supplied by employees, so teams can investigate exceptions with more context about the transactions involved.

“Finance leaders have spent years establishing controls over how transactions are processed, exceptions approved and decisions recorded. They are not going to abandon those requirements simply because AI can now do more of the work,” said Amit Gupta, VP Sales at Kognitos. “Before trusting AI in production, they need to know it will follow approved procedures, ask for help when necessary and leave a clear record of its actions. This is particularly important in areas such as accounts payable, reconciliation and the financial close, where mistakes can have far-reaching consequences.”

How AI Employees and Context Graphs work together
Every AI employee works from a Kognitos context graph. Cartographer maps a team’s systems into the graph when the employee is onboarded, and the team’s teaching is stored there, along with a record of who taught it and why. As approved guidance and accepted knowledge build up, fewer runs need a person, and the team decides, action by action, when an employee may act without asking first. The same graph runs the monitors that check the work every day.

“CFOs don’t need AI that sounds smart. They need AI they can delegate to and defend in an audit. Kognitos AI Employees are hired into defined roles, run the same approved steps every time, and leave a record an auditor can re-perform. That changes the math on growth: finance can expand capacity without expanding headcount at the same rate, and put its best people on cash, capital allocation and strategy,” said Rajeev Madhavan, Founder and Managing Director, Radiant.

Recognition
Gartner named Kognitos a Sample Vendor for Context Graphs in the Gartner® Hype Cycle™ for Agentic AI, 2026, and a sample vendor for process agents, for KYC onboarding, in 3 AI Agent Archetypes. Kognitos was named #1 Overall and an Exemplary Provider in the 2026 ISG Buyers Guide for Automation and Orchestration, a Hot Tech 2026 company by HFS Research for deterministic AI automation, Most Innovative AI Product at the 2026 theCUBEd Awards, and Gold Winner, Neuro-Symbolic AI Platform at the 2026 Globee Awards, and is an MIT Sloan Innovation Winner.

Availability
Kognitos AI Employees and Context Graphs are available in preview to Kognitos customers. To see the 61 roles and hire a first AI employee, visit kognitos.com/ai-employees. To see the monitor catalog, visit kognitos.com/platform/context-graph.

About Kognitos

Kognitos builds the most advanced neurosymbolic AI employees for finance and banking: AI employees that follow your SOPs and show their work. They run on the first neurosymbolic AI platform for business automation, where AI reads documents and requests, and Kognitos’s patented English-as-Code engine executes each standard operating procedure deterministically, in plain English an auditor can read, the same way every time. Finance and banking teams hire from 61 AI employee roles for work such as accounts payable, reconciliation, the month-end close and KYC. They use Context Graphs to turn plain-English questions into controls that run every day and remember every decision, and 100+ apps whose numbers are computed, not generated. Headquartered in San Jose, California, Kognitos is backed by Khosla Ventures, Prosperity7 Ventures, Engineering Capital and Wipro Ventures. Learn more at kognitos.com.

Media contact: pr@kognitos.com


Primary Logo

Media gallery

About The Author
Kognitos Inc