{"id":96319,"date":"2026-11-06T08:00:00","date_gmt":"2026-11-06T16:00:00","guid":{"rendered":"https:\/\/www.salesforce.com\/?p=96319"},"modified":"2026-11-05T15:15:33","modified_gmt":"2026-11-05T23:15:33","slug":"koa-decisions-salesforce-decision-models","status":"publish","type":"post","link":"https:\/\/www.salesforce.com\/news\/stories\/koa-decisions-salesforce-decision-models\/","title":{"rendered":"Less Reasoning. More Doing. Meet Koa Decisions."},"content":{"rendered":"\n

Decision models are having a moment. In mid-September, TypeSafe AI came out of stealth with Jev<\/a>, a model built to pick from a set of options. A month later, OpenAI made its Decisions API<\/a> generally available, and several other AI companies<\/a> have released decision models of their own. <\/p>\n\n\n\n

Salesforce has used decision models for years, including HyperClassifier<\/a>, which routes requests inside Agentforce<\/a>, sending each inquiry to the part of the agent built to handle it. For example, when a customer asks a service agent, \u201cWhere\u2019s my package?\u201d this routing model sends the message to the part of the agent that handles shipping and delivery. Other parts handle FAQs, current promotional offers, and escalations. Decision models help agents quickly determine the right next step, whether that\u2019s selecting the appropriate workflow, making a yes-or-no decision, or assigning a score, without relying on a larger, more expensive AI model for every task.<\/p>\n\n\n\n

This month, we\u2019re introducing the next version of our HyperClassifier model, Koa Decisions. Building on HyperClassifier\u2019s proven capabilities, Koa Decisions dramatically expands where and how decision models can be used across Salesforce. The result is a more versatile model that brings fast, efficient decision-making to a broader range of workflows and applications.<\/p>\n\n\n\n

Koa Decisions reads a request and a list of categories you define, then makes a decision in a single step, in milliseconds. Our data science team built it by applying novel training techniques to Google\u2019s Gemma, an open-weight model. Among other advances, it adds multimodal support<\/a> for images and audio. <\/p>\n\n\n\n

We plan to extend it well beyond routing, into flows, Apex, and prompt templates<\/a> for Salesforce builders, plus an API for developers. Koa Decisions joins the Koa family of Salesforce models, which also includes a separate reasoning model<\/a> built with NVIDIA. <\/p>\n\n\n\n

Why decision models matter for enterprise agents<\/h2>\n\n\n\n

A large language model (LLM)<\/a> generates text, but much of an agent\u2019s work is simply making choices, such as which team gets a request or which knowledge article to show. That\u2019s what a decision model does: It returns a choice, like one option from a list, a yes or no, or a score.<\/p>\n\n\n\n

Say a customer types \u201cI can\u2019t log in to my account\u201d into a company\u2019s chat window. Should the request go to engineering, accounts, or sales? A general-purpose LLM can interpret the request, but it isn’t purpose-built to make these quick, targeted decisions, and it can take more than a second to work through the question.<\/p>\n\n\n\n

A decision model, by contrast, is designed for exactly this task. Given the same request and options, it identifies \u201caccounts\u201d in milliseconds, returning a clear answer and confidence score without the overhead of a larger, general-purpose model.<\/p>\n\n\n\n

An answer like that is often a single token, the basic unit of text an AI model<\/a> processes, about the size of a word. An LLM writes its answer one token at a time, with each new token based on the ones before it. A decision model reads the whole request at once and produces its answer in one step. A short answer takes less computing to produce, and other software can act on it right away. <\/p>\n\n\n\n

When we measured routing requests in Agentforce<\/a>, we found<\/a> that our decision model took roughly 26 milliseconds, compared with about 1.4 seconds for a general-purpose LLM. For a closer look at how Koa Decisions works, read our team’s research paper<\/a>. <\/p>\n\n\n

\n
Open Image Modal<\/span>\"Diagram<\/a>
A customer message goes in, and one decision comes out, such as the team that should handle it.<\/em><\/figcaption><\/figure>\n
<\/div>