Sales reps spend 60% of their time not selling. AI is giving it back

Sales reps spend 60% of their time on tasks that are not selling, according to Salesforce's research (Salesforce). Research, data entry, follow-up emails, CRM updates. AI is now taking a real share of that load. The reps who gain most are the ones who use the time for customers, not for more admin.

What the numbers say for sales

Most of the numbers below come from Salesforce, which sells AI agents for sales teams. Read them as the company's own research, not independent findings.

Salesforce's State of Sales 2026 found 87% of sales organisations use AI, and 54% have already used AI agents. Nearly 9 in 10 plan to by 2027. Top performers were 1.7 times more likely to use prospecting agents. Sales teams expect agents to cut research time by 34% and content creation by 36% (Salesforce State of Sales, 2026).

Among reps already using agents, 85% said AI frees them for higher-value work. But 72% of reps also said they feel overwhelmed by the skills the job now requires (Salesforce). The tools arrived faster than the training.

Independent data points the same way. McKinsey's 2026 survey found revenue gains from AI are most often attributed to marketing and sales (McKinsey, 2026). In India, SBI has generated about ₹22,000 crore of business through AI-generated analytical leads (Financial Express, 2026). EY India expects GenAI productivity gains of 38–40% in banking sales and customer service by 2030 (EY India, 2025).

What the day-to-day work looks like with AI

The best way to picture this is a normal week. These are illustrations, not survey data.

• Before a call. An enterprise rep in Bengaluru asks an AI assistant to summarise a prospect's latest annual report, recent news and job postings into one page: what they are trying to do this year, and where the product might fit. Twenty minutes of reading becomes five minutes of checking.
• After a call. A rep selling to hospitals records notes (with patient and personal details left out), then asks AI to draft the follow-up email, update the CRM fields and list open questions. She edits the email so it sounds like her before sending.
• Pipeline review. A sales manager in Mumbai asks AI to flag deals with no activity in two weeks and draft a check-in plan for each. He decides which ones to chase.
• Insurance and financial products. BCG found admin takes over 50% of an insurance agent's or broker's time (BCG, 2025). Drafting policy comparisons and renewal reminders with AI gives some of that back.

In every case, the rep keeps the relationship. AI does the preparation and the paperwork.

1. 1.AI preps the account

   One-page brief: priorities, buyer, objections.

2. 2.Verify the facts

   Check every claim against a real source.

3. 3.You run the call

   Read the room and keep the relationship.

4. 4.AI drafts the follow-up

   Email, CRM fields and open questions.

5. 5.Edit and reinvest time

   Make it sound like you; book more calls.

The skills to build

Research prompting. Ask AI specific questions about a prospect, not "tell me about this company". Ask what changed in the last year, who the buyer likely is, and what objections they may raise. Then check the sources.

Personalisation without fakery. AI can write a hundred "personalised" emails in a minute. Buyers notice when the personal detail is shallow. Use AI for the first draft, then add one line only you could write.

Knowing where AI is weak. In a Harvard and BCG study, consultants with AI were faster and better on tasks AI handled well. On a task designed for AI to fail, they were right 60–70% of the time with AI, against 84% without it (HBS AI Institute). Pricing, discount approvals and contract terms are not places to trust a confident answer.

Working with agents. As more teams use agents that act on their own, reps need to set clear instructions, check what the agent did, and step in when it goes off track.

Try this week: pick three accounts you will speak to next week. For each, ask an AI assistant for a one-page brief: the company's priorities, likely objections, and two questions worth asking. Verify every fact against a real source. After the calls, note which brief helped and which invented something.

What stays human, and the risks

Trust is the product in sales, and it is easy to damage. KPMG and the University of Melbourne found 66% of employees rely on AI output without evaluating it (KPMG, 2025). A wrong figure in a proposal or a made-up case study can lose a deal that took months to build.

Customer data needs care. Client names, contact details and deal terms do not belong in a public AI chatbot. Use the tools your company has approved.

Volume is a trap too. If AI lets you send ten times more emails, your buyers get ten times more noise. The reps who win use AI to send fewer, better messages.

What stays human: reading the room on a call, knowing when to stop pitching, negotiating, and being the person the customer trusts. Employers know this. In Microsoft and LinkedIn's 2024 research, 71% of leaders said they would rather hire a less experienced candidate with AI skills than a more experienced one without (Work Trend Index, 2024). The combination of AI skill and people skill is what they are paying for.

What to do next

1. Track your non-selling time for one week. Write down every task that is not talking to a customer.
2. Hand one task to AI. Start with call prep or follow-up emails, the easiest wins.
3. Build a checking habit. Every number, name and claim in an AI draft gets verified before it reaches a customer.
4. Reinvest the time. Block the hours you save for extra calls or deeper account planning, or they will fill up with more admin.

AI Fluency is taught live in small cohorts, with a project you keep at the end of each month, and sales professionals are welcome. Start here if you want to build these skills with others.