The first version of Kira was a chat interface. You typed, she responded. We shipped it to three lenders and watched them use it for a week. Then we rebuilt it.
What we learned from the chat version of loan decisioning
When you give a credit or risk team a blank chat box and tell them to describe their loan decisioning flow, they describe the happy path. They tell you what the policy is supposed to do, not what it actually does when the documents are messy, a handwritten payslip, a photographed bank statement, a scanned ID that other IDPs choke on. The Decision Engine, the visual flow that runs your credit policy on every application, surfaced those edge cases in a way that a conversation never did. Once the flow is on screen, you can see where an income figure needs income normalization or where a bank statement needs real cash-flow analysis before a rule can fire.
Why loan decisioning flows instead of prompts
Prompts are stateless. Every time you type a prompt, Kira has to reconstruct her understanding of your loan decisioning flow from scratch. Flows are stateful. Kira can look at your current Decision Engine setup, understand what's there, identify what's missing, and propose a specific edit, not a description of an edit, but the actual node change, with a preview.
That preview is the key. When Kira proposes a flow edit, you see the policy with and without her change side by side. A credit officer approves it, rejects it, or asks her to reshape it. There's no ambiguity about what she's proposing, and there's no implementation lag between approval and execution. The same logic applies whether the change is a new document intelligence step or a tightened underwriting rule.
The design constraint
We made a deliberate choice early on: Kira proposes, credit and risk teams decide. She never applies a flow edit without showing you the preview and waiting for approval. This feels slower than an autonomous agent. In practice, it's faster, because the approval loop is ten seconds and the rollback is one click, so people are willing to try things they would never try if the edit required a ticket.
This is the model in production at Alon Capital, where founder Rene de Jesus puts it plainly: "Floowed reads the documents, runs our credit policy, and surfaces a decision in minutes." The reading and the running both live in flows Kira can see and reshape, never in a prompt she has to reconstruct. Bring your own score or scorecard and it slots into the same flow, absorbed unchanged, we orchestrate the decision, we don't compete with your model. Start free, or book a demo to see Kira reshape a live loan decisioning flow.