{"id":41208,"date":"2026-09-22T08:13:02","date_gmt":"2026-09-22T15:13:02","guid":{"rendered":"https:\/\/salesforcedevops.net\/?p=41208"},"modified":"2026-09-22T08:13:04","modified_gmt":"2026-09-22T15:13:04","slug":"salesforce-aiforce-evals","status":"publish","type":"post","link":"https:\/\/salesforcedevops.net\/index.php\/2026\/09\/22\/salesforce-aiforce-evals\/","title":{"rendered":"Show Me the Evals, AIforce"},"content":{"rendered":"\n

Salesforce says Claude works better with its new layer than without it. I\u2019d like to see that.<\/em><\/p>\n\n\n\n

By Chris Pearson<\/strong><\/p>\n\n\n\n

Claude working through Salesforce\u2019s newly announced reasoning layer, AIforce, will do better CRM work than Claude left to sniff around your org, and the rest of your enterprise data estate, by itself was the big bold claim announced last week.  <\/p>\n\n\n\n

I\u2019m prepared (and cautiously excited) to believe it. I\u2019ve watched Claude sniff around an org by itself. It guesses, confidently, and it burns tokens doing it.\u00a0I just want to see it work – like I expect it to.
I’ve experimented with asking Claude various “end user” types of requests. Simple queries using standard fields and functionality usually work fine. The more customized your org is (and by that measure, the more tech debt you have) – the less of a chance Claude has of doing what you expect. <\/p>\n\n\n\n

This leaves me with more questions than answers. Notably: why hasn\u2019t anyone shown us by comparison how poorly Claude performs without AIforce? That comparison is the whole sale, and it is the easiest kind of claim to test. The materials I was given offer an adoption count instead.<\/p>\n\n\n\n

I\u2019d like to see the evals.<\/p>\n\n\n\n

\"Two<\/a>
Figure 1. The comparison at the center of the AIforce pitch, and the two numbers it would take to settle it.<\/em><\/figcaption><\/figure>\n\n\n\n

Figure 1.<\/strong> The comparison at the center of the AIforce pitch, and the two numbers it would take to settle it.<\/p>\n\n\n\n

AIforce<\/a> was a headline announcement<\/a> at Dreamforce 2026. The idea: people don\u2019t have to come to Salesforce anymore. Salesforce comes to them, in Claude, in Slack and in Lightning, carrying the data, business logic, permissions and governance already built into the org. Loosely translated: ask an LLM about your pipeline and the answer is supposed to be grounded in your org instead of a plausible-sounding guess. It shipped as Claudeforce, Slackforce and Agentforce Coworker. Amazon, Google and Microsoft are on the partner list, not in the box.<\/p>\n\n\n\n

That\u2019s a problem worth solving – but it is also not a new one.<\/p>\n\n\n\n

Table of contents<\/h2>