Kognitos says its new AI workers for finance teams follow a company's approved steps and keep a record an auditor can check, now in preview for its customers.
Kognitos, a software company, has launched AI "employees" for finance and banking. The idea is simple. Instead of an AI deciding how to do a job as it goes, each one follows the written steps a company has already approved. The company announced the launch on Oct. 21, 2026.
The launch includes 61 AI job roles built on 515 standard operating procedures. Those are written, step-by-step instructions for how a task gets done. A tool called Quill drafts new procedures, and a person must approve each one. Kognitos also added 133 ready-made monitors, which are daily checks a team can switch on with one click.
Kognitos argues that many AI agents never leave the testing stage because they are hard to trust. An agent that picks each step itself can handle the same invoice differently each time. That makes it hard to explain to an auditor why a payment was made. The company cites research on 15 models over 18 months. It found accuracy improved, but reliability barely changed.
Each AI employee has a job description, approved procedures, limits on what it can do without human approval, and rules for handing work to colleagues. Kognitos says its engine runs each procedure the same way every time, in plain English an auditor can read.
Each AI employee also works from what Kognitos calls a context graph. It is a map of a company's systems, plus a record of decisions and of who taught the AI what, and why. Over time, the company says, fewer runs need a person. The team decides, action by action, when an employee may act without asking first.
The roles target finance and banking work such as accounts payable, reconciliation, the month-end close and KYC, the identity checks banks run on customers. The new tools are available in preview to existing Kognitos customers. Customers can also connect their own AI agents through technical links called REST and MCP.
The material does not give a price or a date for the end of the preview. It also offers no independent test results. The claims about reliability and audit trails come from the company and its supporters.
If your business uses AI for bills or payments, ask whether it follows set steps and leaves a record you can check.
Otto, a new customer-tracking app for field sales teams, lets reps speak their notes instead of typing them. Price and launch details are not yet known.
Read more: Otto launches a sales app you talk to instead of typingVisa tested 22 AI models on realistic shopping tasks. They passed just over half, and often bought the wrong thing anyway.
Read more: Visa tested 22 AI shoppers, and most still make costly mistakesDARPA wants AI to hunt for design flaws in stock, bond, energy and cloud markets, so enemies cannot use them to cause economic damage that reaches households.
Read more: Pentagon wants AI to find weak spots in stock and energy markets
