Logo SVG code has been copied to your clipboard

The State of Agentic Sales 2026

What nearly 49 million prospects reveal about AI, sales teams, and pipeline.

by Amplemarket Research

Data from January 2022 to September 2026 ·

About this data

This report uses Amplemarket data collected for nearly 49 million prospects from January 2022 to September 2026. Analyses use different subsets and observation windows. These observational comparisons may also reflect differences between teams and workflows.

Outbound was changing before AI adoption accelerated

Email reply rates were already falling before AI-assisted outreach gained traction in this dataset. Calls and LinkedIn later took a larger share of outreach as teams expanded beyond email.

Email reply rates were falling before AI-assisted outreach took off

Share of contacted prospects who replied by email, by month of first contact
Email reply rates fell from 6.5% in 2022 to 1.8% in 2026, with 76% of that decline occurring before September 2024, when the chart’s AI period begins. In July 2026, reply rates were 2.0% for AI-assisted outreach and 1.7% without AI.

AI-assisted emails show more variation than pre-AI templates

How alike the first emails a team sent to different prospects were (0 = nothing in common, 100 = identical)
First emails within a team became less alike. Their similarity score fell from 70 before AI (2023) to 49 with AI (2026), compared with 66 for first emails sent without AI in 2026. Before AI, 51% of email pairs were near-identical. With AI, only 16% were.

Calls and LinkedIn now make up a larger share of outreach

Share of completed outreach tasks by channel, by month (LinkedIn profile visits, likes and follows not counted)
Calls grew from 0.7% of completed outreach tasks in 2022 to 11.9% in 2026, and LinkedIn grew from 7% to 13% (2026 covers January to August). Most of this shift happened after 2024.

Are sales teams using AI?

In two years, AI-assisted outreach grew to reach more than a third of newly contacted prospects. Agents now select about 1 in 5. Sellers also use connected AI assistants such as Claude and ChatGPT to research prospects and build campaigns.

More than 1 in 3 new prospects now receive AI-assisted outreach

Share of newly contacted prospects whose outreach was AI-assisted, by month
The share of new prospects getting AI-assisted outreach grew from 0% in June 2024 to 37% in September 2026. A year earlier, in September 2025, it was 19%.

AI now selects 1 in 5 new prospects

Share of newly contacted prospects by who chose them and who wrote their sequence, by month
In September 2026, AI selected 22% of new prospects, while 37% received AI-assisted outreach. Duo Copilot selected prospects and wrote outreach for 12.5%. Connected AI assistants accounted for 16.4%: 9.4% with assistant-selected prospects and 7.0% with rep-selected prospects. In-app AI accounted for another 7.7%. Two years earlier, AI’s share was below 1%.

Connected AI assistants support outreach to 17% of prospects

Share of prospects contacted each month in 2026 whose sequence a connected AI assistant built, loaded or edited
Through MCP, AI assistants like Claude or ChatGPT built, loaded, or edited the sequence for 0.8% of prospects in April 2026 and 16.6% in September 2026. Copy drafted in a chatbot and pasted into Amplemarket isn’t counted. In September, 3,253 sellers at 860 teams made 2.4 million tool calls (requests an assistant sends to Amplemarket).

The number of reps using connected AI assistants grew 7× in five months

Sellers using an AI assistant, per month

Reps use connected AI assistants most often to find and research prospects

Share of connected AI-assistant sessions in September 2026 that included each kind of task (a session can include several)
60% of AI-assistant sessions include finding and researching prospects. 1 in 18 sessions builds a whole outbound campaign, from searching for people to launching the sequence.

How is AI changing sales teams’ work?

Teams using AI reach more companies and build longer sequences across email, LinkedIn and phone. Their first emails are less repetitive. Broader account coverage, however, does not necessarily mean reaching more people within each account.

Reps using AI build sequences with more steps and more channels

Steps per sequence, by channel

Channels used in the sequence

Before AI = 2023; AI-assisted = January–July 2026
Sequences grew from 5.6 steps before AI to 7.9 with AI, and still run about 16 days. Most of the extra steps are on LinkedIn. 21% of AI-assisted sequences combine email, LinkedIn, and phone, vs 14% of sequences without AI and 5% before AI.

Sequences with more steps have higher meeting rates

Meetings booked within 30 days per prospect, compared with 1–6-step sequences (1.0×), by steps in the sequence, January–July 2026
Sequences with 13 or more steps book 2.0× as many meetings per prospect as sequences with 1–6 steps. Sequences with 7–12 steps book 1.2× as many.

AI-assisted email, LinkedIn and call sequences have 4.6× the interest rate of email-only outreach without AI

Interested prospects within 30 days since outreach, comparing email-only sequences with multichannel sequences. January–July 2026
Compared with email-only sequences without AI, AI-assisted sequences that combine email, LinkedIn and calls get 4.6× the share of interested prospects. The same three channels without AI get 2.7×, and email alone with AI gets 1.6×. A prospect counts as interested when an email or LinkedIn reply is marked interested, or a call is logged as interested.

Reps using AI reach 19% more accounts than equally active teammates without AI

Companies first contacted per rep per quarter, vs equally active teammates without AI in the same team and quarter (= 100), 2026
On an index where equally active teammates without AI reach 100 companies, reps using AI reach 119: 99 selected by reps and 20 selected by the agent. The comparison covers 1,805 rep-quarters (one rep over one quarter) at 342 teams, matching teammates within the same team and quarter.

Accounts reached through 6–10 people have 4.5× the meeting rate of single-contact accounts

Meeting rate per account vs accounts where one person was contacted (same team and company size, Q2 2026)
Accounts where three people were contacted had 2.2× the meeting rate of single-contact accounts; those with six to ten contacts had 4.5× the rate. Yet accounts worked only with AI reached fewer people: 1.45 on average, versus 1.89 without AI.

AI-assisted first emails are less repetitive

Share of pairs of first emails from the same team and month that are not near-copies (under 95 out of 100 similar)
In 2026, 84% of pairs of AI-assisted first emails from the same team were not near-copies, compared with 56% without AI and 49% in 2023.

What is the impact of AI in sales?

AI is involved in a growing share of outbound pipeline. At the median CRM-connected team, 3 in 10 new outbound opportunities trace back to AI, and AI-sourced meetings become opportunities at similar rates to other meetings. After agent adoption, buyer interest increased more consistently than meeting volume.

Reps see 43% more interested buyers after their team adopts an AI agent

Per rep, months 2–4 after a team switched on Duo Copilot vs the 3 months before, compared with teams that hadn’t switched it on (1.0× = no change)
Relative to changes among teams that had not adopted Duo Copilot, interested buyers per rep rose 43% in months 2–4 after adoption and 36% in months 5–7. The gain came from agent-selected prospects; interest from reps’ own prospects changed little (1.05×). Meeting gains were less consistent: 1.11×, with the median team at 0.98×.

Agent-selected prospects accept LinkedIn connection requests 50% more often

LinkedIn connection-request acceptance within 30 days of first touch, relative to rep-selected, rep-written outreach.
Agent-selected prospects accepted LinkedIn connection requests 1.5× as often as prospects selected and messaged by reps. Acceptance was particularly high among people who had posted about a relevant topic or engaged with a competitor. AI-written requests to rep-selected prospects showed no uplift in this comparison (0.94×).

3 in 10 new outbound opportunities trace back to AI

Share of a team’s new outbound opportunities whose prospect was picked by an AI agent or got AI-written outreach (typical CRM-connected team)
At the median CRM-connected team, AI’s share of new outbound opportunities grew from 1% in Q3 2024 to 11% in Q3 2025 and 30% in Q3 2026. The mean was 35% in Q3 2026, when 91% of teams in this analysis had at least one AI-sourced opportunity.

AI-sourced meetings become opportunities at similar rates to other meetings

Share of meetings that became a new CRM opportunity within 60 days, same team and quarter. Based on nearly 27,000 meetings
37% of AI-sourced meetings became a new opportunity within 60 days, vs 33% of other meetings at the same team and quarter (1.12×). Based on 158 teams.

Reps combining calls and 3+ AI tools book 2.4× as many meetings as reps using neither

Reps combining calls with three or more AI tools booked 2.4× as many meetings as reps doing neither.
Reps who make calls and use three or more AI tools book 1.66× their team’s average meetings. Reps who do neither book 0.70×. Per prospect, every group books at about its team’s rate. The difference is volume: these reps work 1.61× their team’s average number of prospects.

What should sales teams do next?

  1. Use AI to expand account coverage.Reps using AI reached 19% more companies than equally active teammates. Track whether agent-selected accounts add useful reach beyond the accounts your reps already work.
  2. Build account depth deliberately.Accounts where six to ten people were contacted had 4.5× the meeting rate of single-contact accounts. Yet AI-only outreach reached fewer people per account. Track account coverage and account depth separately.
  3. Test a broader channel mix.Email, LinkedIn and phone appeared together in 21% of AI-assisted sequences, compared with 14% without AI. Measure which combinations produce interested prospects for your team.
  4. Review relevance alongside variation.In 2026, 84% of AI-assisted first-email pairs were not near-copies, versus 56% without AI. Check whether messages address each recipient’s role, needs and context.
  5. Measure the outcomes of additional follow-up.AI-assisted sequences averaged 7.9 steps, versus 5.6 before AI, over roughly the same 16 days. Longer sequences had higher meeting rates; test whether extra steps improve your team’s results.
  6. Track the path from interest to meetings.Agent adoption was associated with 43% more interested buyers per rep, while meeting gains were less consistent.
  7. Track AI involvement in your CRM.At the median CRM-connected team, 3 in 10 new outbound opportunities traced back to AI, up from 11% a year earlier. AI-sourced meetings became opportunities at similar rates to other meetings.
  8. Start with the workflows sellers use most.Finding and researching prospects appeared in 60% of connected AI-assistant sessions. Complete campaign workflows appeared in 1 in 18 sessions.
  9. Assess meeting volume and conversion together.Reps combining calls with three or more AI tools booked 2.4× as many meetings as reps doing neither. Their meeting rates per prospect were similar; they worked more prospects.