By Shawn Johnson. Published 2026-10-13.
The short answer. They will trust it when they get to review it. Pipedrive surveyed 1,000 working professionals in July 2026. The most wanted AI capability was drafting emails, summaries and notes for a person to review, at 43.3 percent. A combined 22.4 percent said they are uncomfortable with AI finishing tasks alone. 25.3 percent named one barrier ahead of price and ahead of training. They cannot tell when the output can be trusted. That reads as a build requirement. It matters most on a field team. An in person visit leaves no recording, so the rep is the only source. Call to CRM is built that way. The rep calls a number after a meeting. June asks the team's own questions. He answers out loud. She confirms what she heard. The call runs three to five minutes, and structured notes land in the CRM the company already runs.
What do salespeople want AI to do for them?
They want it to write the draft and let a person check it.
Pipedrive asked 1,000 employed adults in customer facing roles what AI capability they want most. The winner was drafting emails, summaries and notes for human review, at 43.3 percent. A combined 22.4 percent went further and said they are uncomfortable with AI completing tasks on its own.
One caveat worth stating plainly. Pipedrive surveyed working professionals, not field sales reps. The preference is theirs. The field sales argument below is ours.
Is this resistance to AI?
No. Adoption is already settled.
Salesforce surveyed 4,050 sales professionals in August and September 2025. 87 percent of sales organizations already use some form of AI. That covers prospecting, forecasting, lead scoring and drafting emails. 54 percent of sellers have used agents, and nearly nine in ten plan to by 2027. 94 percent of sales leaders who have agents call them critical for meeting business demands.
So the people in these surveys already live with AI every day. They are asking for one thing back. Show me the draft before it counts.
What is actually stopping adoption?
Reps cannot tell when a draft is right.
In the Pipedrive survey, 25.3 percent named one barrier ahead of every other. They cannot tell when AI generated output can be trusted. A rep reading a summary in his truck has no way to grade it. That is a different problem from cost, and it needs a different fix.
A rep who cannot tell whether a line is right has two choices. He can sign his name to it anyway, or he can leave it alone. Most reps leave it alone. The record stays thin and nobody learns why.
Why does this hit a field team harder?
Because there is nothing to check the draft against.
Picture a rep walking out of a building products distributor after an hour in the conference room. He heard that the second yard is on hold until spring. He heard which competitor the branch manager keeps bringing up. None of that is written down anywhere, and nobody recorded the meeting.
An inside seller has a recording of his call. He can point an agent at it, read the summary, and check the parts that matter. The field rep has none of that. He is the single source of what happened in that room, and his read on it fades inside a day.
So a tool that writes on his behalf and hopes gets ignored. A tool that asks him and reads it back gets used.
How should an AI tool for field sales be designed?
So the rep supplies the facts and confirms them before anything is written.
That is what Call to CRM does. After a meeting the rep calls a phone number. June asks the team's own qualification and risk questions, the ones a good manager asks on a ride along. The rep answers in his own words. June confirms what she heard. Then structured notes reach the CRM the company already runs.
No app. No login. No typing. June is never in the customer meeting, because the rep calls her after it.
Nothing enters the record that the rep did not say out loud and agree to. We built it that way on purpose. It is the design a field team keeps using in month four.
What does a leader get out of it?
A record the team believes, and better questions getting asked in the field.
Every AI product in the stack gets better when somebody confirms the field half instead of guessing at it. The forecast rolls up from lines a rep stood behind. The branch manager coaches from what happened rather than from a reconstruction three weeks later. And the rep who answers the same five questions after every meeting starts hearing them during the meeting.
FAQ
Do salespeople trust AI to write for them?
Conditionally. In Pipedrive's survey, 43.3 percent of working professionals asked for AI that drafts and lets a person review. A combined 22.4 percent said they are uncomfortable with AI acting on its own.
What is the biggest barrier for a sales team adopting AI right now?
Trust in the output. In the same survey, 25.3 percent of the people surveyed could not tell when AI generated output can be trusted. That ranked ahead of cost and ahead of training.
How should an AI tool for field sales be designed?
So the rep supplies the facts and confirms them. There is no recording of an in person meeting, so the rep is the only source. A design that asks him and reads it back is the one that survives contact with a real territory.
Does this replace the AI we already bought?
No. It feeds it. Platform agents work well on the half of the pipeline that leaves a digital trail. Call to CRM covers the half that happens in a customer's building.
Sources: Pipedrive, Clarity, not automation, is becoming sales teams' biggest productivity advantage, July 8 2026, survey of 1,000 employed adults administered through Pollfish. Salesforce, State of Sales 2026, survey of 4,050 sales professionals fielded August through September 2025. Both fetched and verified live 2026-10-12.