Salesforce published The ROI Report: Sales Cloud Agents on November 10. Its lead finding is that Sales Cloud agents save sellers an average of six hours a week, based on a survey of 24 customers who actively use those agents.
What Salesforce reported
The report covers customers who use Sales Cloud’s AI sales agents for administrative work across the sales cycle: prospecting, account research, pipeline management and logging conversations into the CRM. Salesforce lists these findings:
- Agents save sellers an average of six hours per week on individual, non-selling tasks.
- Almost three-quarters of the customers surveyed have deployed agents across prospecting, account research and pipeline management, and they save five hours or more per week on those tasks.
- 63% report more accurate and complete CRM data.
- Two in three report more thorough account research, and two in three report smoother handoffs across the revenue cycle.
- Deals move 10% to 20% faster for two-thirds of account executives.
- 57% of sales leaders agree that agents let the business increase revenue without adding headcount, and 75% say they kept their exact same hiring plans while increasing overall sales output.
Salesforce opens the report with its own estimate that the average seller spends 60% of their time on non-selling activities such as updating CRM systems and researching accounts.
Advertisement
How the survey was run
The methodology note says the data comes from an online survey of 24 Sales Cloud customers conducted on July 21 and 22, 2026, with 32 questions about their experience with Sales Cloud agents. Respondents came from two sources: existing customers that Salesforce invited directly, and Sales Cloud users recruited through a third-party research panel. All were screened to confirm they actively use the agents.
The figures are what active users report about their own teams. The page gives averages and shares and does not show the spread across the 24 respondents, and the report has no comparison group of teams working without agents. Salesforce links the full report from the announcement for readers who want the detail behind each chart.
Where the time comes from
The hours come mostly from three areas: prospecting, account research and pipeline management. Salesforce ties the CRM data finding to the same work. It says agents log conversations and account research into Salesforce automatically, so the record no longer depends on reps typing up notes from memory at the end of a long week.
One customer named in the report is Equipter. David Beiler, VP, Sales at Equipter, said, “Sales Cloud helped us shorten that window and focus our team on higher-value conversations. It’s made scaling outreach a lot more manageable.” The window he describes is lead follow-up, which he says had taken his reps four to eight hours, or even a day.
The report also describes where sales leaders now spend time. Two in three say they spend more time on partner and channel relationships, 67% spend more time engaging wider buying committees, and one in three spend more time on keeping relationships after the sale. Salesforce says buying committees at many companies have grown to eight to 10 people, and all of the customers surveyed say they can handle more leads and accounts with agents than without them.
How it fits with other agent evidence
Vendors and investors have published a run of agent data this autumn. Menlo Ventures found limited evidence of AI replacing CRM. Salesforce said earlier this month that Agentforce Coworker passed 185,000 users, and Vercel and Salesforce described the sales agents they built for their own teams. The ROI report adds customer-reported outcomes from Sales Cloud users to that set.
What it means for the sales leader
The CRM data finding is the one a sales leader can test inside their own org. Rep hours are hard to measure without time tracking, while field completeness on opportunity and activity records already sits in the CRM. Before switching on logging and research agents, RevOps can save a baseline report: the share of open opportunities with a next step, a primary contact role and an activity in the last two weeks. Running the same report after 60 or 90 days shows whether the agents moved it. Salesforce’s report connects that data to forecasting, saying a reliable forecast is hard to build when deals depend on notes typed from memory. Duplicate records are a related clean-up job, covered in why duplicate CRM records are now an AI agent problem.
The headcount findings need more care in planning. In the survey, 57% of leaders agree that agents let them grow revenue without adding people, and 75% kept their hiring plans unchanged. Those are answers about current hiring plans from 24 customers. A leader building next year’s capacity plan can use them as one input and still model ramp time, quota and coverage from their own numbers before changing headcount.
Lead follow-up is the cleanest operational test. Equipter’s example is a response window measured in hours, and every CRM records when a lead was created and when it was first touched. That gap can be charted weekly from the first day of a pilot.
What to do next
Teams evaluating Sales Cloud agents can ask Salesforce for the full report and for reference customers in their own segment and deal size. Save a CRM data baseline before the pilot starts. Track lead response time from the first week. Agree with the vendor in advance which of those numbers will decide whether the pilot expands, and run a seller survey at the end so the hours estimate sits next to the system data.
Source: Salesforce
