The first real production numbers on agents came with a 40 point spread. salesforce customers…
the first real production numbers on agents came with a 40 point spread. salesforce customers report resolution rates from 50 to 90 percent on similar agents, and the gap tracks task scoping, not model quality.
a vendor number is a statement about your task design, and the one to believe least is for a use case you haven't scoped.
Context
Salesforce's metrics page says Agentforce customers have resolved 50% of customer service inquiries autonomously, and its customer pages give other figures: Nexo at 62% case resolution, Endress+Hauser expecting 50% of cases resolved, and the Kogan.com post saying resolution rate grew from 6% to 67% across several use cases.
Salesforce's own help site, the company's first deployment, is reported at more than 84% of questions resolved in an April 2025 post and over 75% on a later customer-story page. The Kogan post says Agentforce started with one use case, order management, because where-is-my-order inquiries were roughly 60% of inbound requests, and that product recommendations and TV troubleshooting were added later.
Every figure above comes from Salesforce's own pages, so these are vendor-reported numbers. They span 50% to 84% across the pages read here. A single published range of 50 to 90 percent across customers was not found, so the note's '50 to 90' is unsupported here, not refuted. The ceiling may come from another Salesforce page or a customer report that was not read.
The note's explanation, that the gap tracks task scoping and not model quality, is the author's reading. The Kogan post is consistent with scope mattering, since it ties the 67% to a set of deployed use cases, but it does not compare scoping with model quality, and no source read here measures that.
The note calls these the first real production numbers. Salesforce's help-site figures date to April 2025, so earlier production claims exist, and the first-ness is unsupported here.
Related work
- Agentforce Metrics: Real Impact & Results (Salesforce) ↗Source for the 50% of inquiries figure and the Nexo and other customer results.
- Salesforce Exec Shares Agentforce Customer Support Lessons Learned (Salesforce, April 22, 2025) ↗Source for the 84% help-site figure and the handoff design.
- How Kogan.com Automated 67% of Customer Inquiries and Tripled Resolution (Salesforce, July 10, 2026) ↗Source for the 6% to 67% resolution change and the use-case scope.
Watch next
- Look for a customer-side or analyst report that gives resolution rates across several Agentforce customers. Check how each page defines resolved.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 9 October 2026 at 01:02 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →