Will Sales Jobs Disappear to AI? — What Changes in SDR Work, Deals and Existing Accounts

AI is not wiping out sales jobs in one sweep. U.S. government projections put total employment in sales and related occupations down 1.4% over ten years, but telemarketers fall 21.4% while sales engineers, who explain technical products, and insurance sales agents grow. The dividing line is what the work consists of: delivering set information, or explaining, advising and taking responsibility.

1. The Short Answer — Look at Tasks, Not Job Titles

"Will AI put salespeople out of work?" has no answer at the level of the job title. Even within one sales role, a single day mixes tasks AI handles well and tasks it does not. This article reads four primary sources of different kinds side by side, task by task.

Put together, the four point the same way. Entry-point and administrative work — researching, writing, sending, recording — is starting to move to AI. Explaining, answering buyers' verification questions and staying responsible after the contract are where buyers want a person. Sales jobs are neither vanishing wholesale nor staying the same; each task goes its own way.

For AI and work in general, see Will AI Eliminate White-Collar Jobs?; for which occupations are most exposed overall, Jobs at Risk From AI; for differences between juniors and veterans, Will AI Hit Veterans or Juniors First? This article stays on the sales floor.

2. Government Data — Ten-Year Projections for U.S. Sales Jobs

Few national statistics publish occupation-level employment outlooks together with an explanation of the work. Below are the sales-related occupations from the U.S. Bureau of Labor Statistics (BLS) occupational employment projections for 2025–35. These projections build in every factor, not only AI but also the spread of online sales and demographic change.

The line between shrinking and growing jobs is the same line that separates what AI does well from what it does not. In its occupation profiles (the Occupational Outlook Handbook), BLS gives these reasons:

• Wholesale and manufacturing sales reps (little or no change): the profile expects online sales to mostly complement rather than replace face-to-face selling, while advances in AI and automation such as chatbots may limit employment growth. Even with almost no growth, about 123,400 openings are projected each year, mostly to replace workers who change occupations or retire.
• Insurance sales agents (growing): the profile notes that clients who research and buy insurance online reduce demand, yet agents will still be needed to help clients understand their options and choose the right policy.
• Advertising sales agents (declining): the profile points to the decline of print media, plus the ability to automate digital ad placement and the use of ad blockers. The advertising side of the change is covered in How AI Is Changing Marketing and Advertising.

In short, sales work that only delivers set information shrinks under pressure from online channels and automation. Sales work where the product is complex and buyers want to ask "what about my case?" holds steady or grows. Sales engineers and technical and scientific product reps also earn median pay above $100,000, at the high end of sales. Note that these are U.S. projections; for other countries, check each national statistics office.

3. Sales Tasks Companies Are Handing to AI — Anthropic's Usage Data

BLS projections look ten years ahead and do not show which tasks AI is doing today. Reports that AI companies publish from their own usage data fill that gap. Anthropic regularly publishes the "Anthropic Economic Index," which classifies how Claude is used by work task.

Its March 24, 2026 report, "Learning curves" (covering usage from February 5–12, 2026), named two uses that stood out as growing over three months in the API — the way companies build Claude into their own systems. One of them was "Business sales & outreach automation," made up of the four tasks below.

All four are core tasks of the SDR (sales development representative), who handles prospecting and first-pass qualification. The same report says API use is far more likely to be directive — handing the work off — with less need for a human in the loop than conversational use. Companies are not just chatting with AI; they are starting to pipe the sales tasks themselves into it.

Note also that answering questions in deals, negotiating and post-contract support did not appear among the growing uses in this kind of report. The "information-delivery" work that Section 2 shows shrinking overlaps with the work companies are starting to hand to AI.

4. Where Sellers' Time Goes — Salesforce's Survey of 4,050

The sellers' side comes from Salesforce's "State of Sales" (published February 3, 2026). It surveyed 4,050 sales professionals in 22 countries in August–September 2025 (sales leaders, sellers, partner sellers and sales operations). The main figures:

According to the same survey, sellers spend nearly one full day of their workweek on prospecting, and 48% say they lack the bandwidth to do adequate cold outreach. Only 40% of their time on average goes to selling itself; the other 60% goes to work other than selling. The tasks companies have started handing to AI in Section 3 and the tasks sellers here expect to shrink are nearly the same. What AI is cutting first is not the seller but the 60% of time that is not spent selling.

Data upkeep matters just as much. 74% of sales professionals say they are focusing on data cleansing, and 79% of top performers prioritize data hygiene, compared with 54% of underperformers. AI agents run on CRM data, so organizations with messy records see less benefit. Record-keeping does not disappear; it becomes the job of preparing what AI runs on.

5. What Buyers Want From Sales Reps — Gartner's Survey

The future of sales depends more on buyer behavior than on what sellers want. Gartner surveyed 645 B2B buyers in August–September 2025 and published the results in 2026 (releases of March 9 and May 20; the March release gives the sample as 646).

67% and 69% look contradictory, but they make sense once you see they refer to different stages of buying. While researching, buyers avoid people and gather information through AI and the web. Before deciding, they want to check that information with someone who can answer for it. It is the same pattern as the BLS insurance profile in Section 2: even when clients research on their own, choosing still takes a person's help. As the last row shows, reps are not trusted unconditionally either. Only reps who answer accurately get chosen as the person to check with.

In an August 25, 2025 release, Gartner also predicted that "by 2030, 75% of B2B buyers will prefer sales experiences that prioritize human interaction over AI." This is a forecast, not a survey result. It is most accurate to read it alongside the fact that today's buyers avoid people while gathering information (67%).

6. Task-by-Task Table — What Moves to AI, What Stays Human

The table below rearranges Sections 2–5 by sales task. The "Ease of handing to AI" column is this article's reading of the direction the sources point, not a conclusion stated by the sources themselves. No percentage or timeline forecasts are included, because none could be traced to a publisher.

Industry makes a big difference too. In SaaS and standardized services, where product explanations are easy to template, the top three rows automate faster. In manufacturing, construction, medical devices and other fields where on-site conditions and regulation matter, the bottom three rows carry more weight. The higher the deal value and the more people involved in the decision, the more of the work stays human.

7. What Sellers Can Do — Audit One Week of Your Time

How much AI affects you depends not on your job title but on how your time is split. Three steps let you estimate it yourself.

1. Write down one week of your time, split into the seven tasks in the Section 6 table. Your calendar and sent emails are enough; 30-minute blocks will do.
2. Work out what share of the week the three "High" rows take. In Salesforce's survey, sellers spend only 40% of their time selling on average. If the "High" tasks fill more than half your week, you are on the side where AI changes the content of your job the most.
3. Decide which "Low" tasks will get the freed-up time. Talk concretely with your manager about sitting in on deals, taking on existing accounts, or owning the setup and checking of AI agents.

Three places to move that time pay off:

To current and future SDRs: as Sections 3 and 4 show, SDR work overlaps most with what companies are starting to hand to AI, so a career spent only on entry-point tasks for years tends to narrow. SDR experience is not wasted. An eye for judging prospects carries straight over when you move to configuring AI agents. What matters is getting early exposure to the "Low" tasks in the table, such as sitting in on deals or owning existing accounts. Data showing that the entry door for juniors is narrowing first is in Will AI Hit Veterans or Juniors First?

8. What Managers Can Do, and How to Choose an AI SDR

For managers, designing the rollout of an AI SDR (an AI agent that handles prospecting automatically) as "a way to reallocate time" rather than "a way to cut headcount" fits the numbers in this article better.

• Fix the data first. 79% of top performers prioritized data hygiene (vs. 54% of underperformers). Putting AI agents on top of a messy CRM just sends the wrong copy to the wrong people.
• Move metrics from volume to verified outcomes. AI can inflate meeting counts and send volume without limit. Measure results people influence, such as opportunity conversion, deal size and renewal rate. On the risk of chasing volume, see also The Trap of AI Productivity Metrics.
• Build a path before cutting SDRs. Move them toward configuring and checking AI agents, sitting in on deals, and owning existing accounts.
• Raise your quality as "the person to check with." 69% of buyers want to validate with a rep, yet nearly half also doubt what reps tell them. Product training and materials that show the basis for answers help.

AI SDRs come either built into CRMs such as Salesforce and HubSpot or as standalone products. Product names, features and prices change quickly, so compare on the criteria below and check prices on each vendor's official page.

The basics of letting AI draft emails and replies are in Faster Email and Chat Replies With AI.

Summary

• Government data: U.S. sales and related occupations are projected at −1.4% for 2025→2035 (all occupations +3.5%). Telemarketers (−21.4%) and advertising sales agents (−7.2%) decline, while insurance sales agents (+3.3%) and sales engineers (+2.8%) grow (BLS).
• AI usage data: among the ways companies use Claude through the API, "sales & outreach automation" (lead qualification, data enrichment, cold emails, enablement content) at least doubled in three months (Anthropic, March 2026).
• Seller survey: the average seller spends 40% of their time selling. Respondents expect prospect research to drop 34% and email drafting 36% (Salesforce, 4,050 respondents).
• Buyer survey: 67% prefer a rep-free experience, while 69% want to validate AI-generated insights with a rep (Gartner, 645 buyers).
• What to do: audit a week of your time by task and move freed-up hours to deals, existing accounts and running the AI. Organizations should fix their data and measure outcomes, not volume.

Related: Will AI Eliminate White-Collar Jobs?, Jobs at Risk From AI, Jobs That Survive the AI Era, Veterans or Juniors, AI in Marketing and Advertising.

FAQ

Q. Are sales jobs actually declining in the statistics?
A. BLS projects total sales and related employment to fall 1.4% from 2025 to 2035, below the all-occupation average (+3.5%). The picture splits by occupation: telemarketers are projected to fall 21.4%, while insurance sales agents and sales engineers grow. The projections cover factors beyond AI, such as online sales, and are U.S. figures.

Q. Will selling to small businesses be replaced by AI?
A. The part most easily replaced is the "deliver information" stretch from first contact to sending materials. In Gartner's survey, 67% of buyers prefer a rep-free experience and 70% prefer a completely digital purchase. At the same time, many buyers (69%) want to verify information before deciding, so reps who can answer questions accurately keep their value.

Q. I'm an SDR. Should I change jobs?
A. There's no need to quit in a hurry, but avoid staying on entry-point tasks alone. Lead qualification and cold emails show up in Anthropic's usage data as tasks companies are starting to hand to AI. Start by writing down a week of your time by task, then look inside your company for a path into work that is hard to hand to AI: sitting in on deals, owning existing accounts, or configuring and checking AI agents.

Q. If we roll out an AI SDR, should I reduce my team?
A. Reallocating time first fits the survey numbers better. Sellers spend only 40% of their time selling on average, so shifting the time AI frees up into deals and existing accounts increases revenue-generating time without cutting headcount. An SDR's experience with prospects carries straight over to choosing an AI agent's targets and checking its copy.

Q. Is field sales safe?
A. It depends less on the visit itself than on what the visit delivers. Visits that only drop off materials and explain them overlap with what buyers would rather do digitally. Visits that work through requirements based on the buyer's on-site conditions, help get internal approval, and take responsibility for problems after the contract are exactly the work buyers want from a person.