Prospecting time is a capacity problem before it is a tooling problem. Salesforce's 2026 State of Sales report says sellers expect AI agents, once fully implemented, to reduce prospect research time by 34%. That is an expectation from surveyed sales professionals, not a guaranteed productivity result for every team.
The prospecting time benchmark table
| Benchmark | Result | How to use it |
|---|---|---|
| Sales teams using agents now | 54% | Identify whether your team is early, current, or behind the survey sample |
| Sales teams expecting agents within two years | 34% | Treat adoption as a near-term operating change |
| Expected reduction in prospect research time | 34% | Set a testable baseline for research hours per qualified account |
| Top performers more likely to use prospecting agents | 1.7x | Compare use-case selection, not just software access |
| Survey population | 4,050 sales professionals | Avoid treating the figures as a universal sales benchmark |
Source: Salesforce State of Sales, Seventh Edition, published 2026 and accessed August 5, 2026.
What to measure before automating
Log the time spent finding accounts, checking fit, locating contacts, researching context, writing the first message, and recording the activity. Keep research and outreach separate. A workflow can reduce searching while leaving personalization and approval time unchanged.
The useful local benchmark is qualified accounts researched per hour, with an accuracy check attached. A faster list with the wrong industry, role, or contactability is not a productivity gain. Record the percentage of accounts that pass the ICP screen and the percentage of contacts that require correction.
A bounded prospecting test
Choose one segment and a fixed weekly sample. Let an AI workflow propose account summaries, research prompts, and draft context. A human should approve fit, claims, contact selection, and suppression rules before any outreach. Compare the test cohort with the prior process over two to four weeks.
The lead list building guide provides the operating checklist. A lead research VA can own source checking and exception routing while sales retains approval of targeting and messaging.
Methodology and limits
Salesforce says the survey was double anonymous, conducted in August and September 2025 across 22 countries. The 34% figure describes respondents' expected time savings after full implementation. It is not a controlled experiment and does not establish causation.
Frequently asked questions
Is a 34% reduction a safe planning target?
Use it as a hypothesis, not a promise. Measure your current workflow first and set a smaller initial target tied to accuracy.
What should AI do first in prospecting?
Start with repeatable research and summarization. Keep ICP decisions, claims, contact policy, and send approval with a named human owner.
What is the right prospecting KPI?
Track qualified accounts per hour, contact accuracy, positive reply rate, and meetings created. Time alone can reward low-quality output.
Related Research
If the research queue is the bottleneck, review the lead list building VA service.
Sources
- Salesforce, State of Sales, Seventh Edition (published 2026): https://www.salesforce.com/en-us/wp-content/uploads/sites/4/documents/reports/sales/salesforce-state-of-sales-report-2026.pdf
- Salesforce, State of Sales announcement (published February 3, 2026): https://www.salesforce.com/news/stories/state-of-sales-report-announcement-2026/