What 30,000+ AI interactions reveal about how GTM teams actually work | Endgame

Every existing report on AI in sales is based on surveys. We did something different: we analyzed more than 30,000 real AI interactions from hundreds of GTM professionals across enterprise organizations. We combined that with interviews to validate that our quantitative analysis matched what teams experience on the ground. No self-reporting. No hypotheticals. Just the behavioral record of what sellers, leaders, and ops teams actually ask AI to do when nobody's watching.

This report is based on 31,000+ AI interactions from hundreds of GTM professionals at enterprise organizations. No surveys, no self-reporting. See Methodology for details.

How to read this report

Personas

• Sales Reps (55% of users): AEs, CSMs, SEs, SDRs, BDRs, Account Managers
• Sales Leadership (25%): VPs, Directors, CROs, Heads of Sales
• RevOps & Enablement (14%): RevOps, Sales Ops, Deal Desk, Enablement

The remaining 6% is Product & Marketing and unmapped roles, discussed in the conclusion.

Five jobs-to-be-done

Account Intelligence Conversation Readiness Deal Acceleration Pipeline Inspection Team Enablement

Each job and its component use cases are defined in Finding 1.

Finding 1: Teams spend 2x more AI time on understanding than producing

When you classify every AI interaction by the job the user is trying to accomplish, a clear structure emerges. GTM teams hire AI for five distinct jobs-to-be-done, each composed of specific use cases:

These split into two groups: selling-oriented jobs that help individuals, and organizational jobs that make the team smarter.

Five jobs-to-be-done: how GTM teams use AI

The three selling-oriented jobs account for 53% of all activity. Within that, teams spend more than twice as much AI time on contextual understanding (Account Intelligence + Conversation Readiness) as on producing deliverables (Deal Acceleration). The research behind great outputs, stitching together CRM records, call transcripts, competitive landscape, and stakeholder history, was too time-consuming to do well before.

Now that it's accessible in minutes, teams invest in it. Tim Miller, CRO at Monte Carlo, describes the shift: "I can pull up strategic insights across all our wins in seconds — the top 20 initiatives we're tied to, the best use cases for a specific industry. That quick identification helps us make deals move faster and larger."

The contextual research required for great outputs was never optional. It was just impossible at speed. Now that it's accessible, teams invest in it, and everything downstream gets better.

Team Enablement is the single largest category at 36.3%. The majority is automated digests: synthesized account updates, pipeline changes, and competitive movements delivered on a daily or weekly cadence without anyone asking. It also includes coaching and rep performance analysis. Pipeline Inspection (10.8%) is the active management counterpart: auditing deal health, flagging qualification gaps, monitoring which accounts need attention now.

Finding 2: Each persona has a distinct AI playbook

When you break the data down by persona, three distinct profiles emerge.

How each persona allocates AI time

% of each persona’s interactive activity · JTBD order is consistent across all three bars

Reps: 89% selling-oriented

Account Intelligence leads at 38%, followed by Deal Acceleration at 29% and Conversation Readiness at 23%. The workflow mirrors a rep's daily rhythm: understand the account, build the deliverable, prepare for the conversation.

"Instead of analyzing endless history and context, I can ask for help: highlight any historical interactions where things went wrong, why did we not go forward, help me build a hypothesis so I can re-engage. I cut down a ton of time."

Deal Acceleration is content production: emails, decks, proposals, business cases. Reps need AI that's fast, deal-attached, and grounded in account context. If it can't pull from real CRM data, call transcripts, and the sales methodology the team runs on, the deliverables will be generic, and reps will stop using it.

Leaders: 36% Conversation Readiness

Leaders show a different pattern. Their #1 job is Conversation Readiness, driven by intelligence briefings: synthesized summaries consumed for inspection and decision-making. Account Intelligence is second at 30%, skewed toward competitive intel and customer insights rather than account-by-account research. Deal Acceleration drops to 13%. Leaders consume context and set direction; they don't draft follow-up emails.

RevOps: 34% Pipeline Inspection, 33% Account Intelligence

RevOps is the most distinctive profile: equal weight on managing the machine and understanding what's happening across accounts. Within Pipeline Inspection, deal management leads: auditing qualification gaps, flagging stalled deals, monitoring stage progression.

Within Account Intelligence, RevOps gravitates toward competitive intel and customer insights at the portfolio level. Deal Acceleration is just 9%. RevOps isn't working individual deals. They're working the system. Like leaders, RevOps looks at whole-book patterns, but for process optimization rather than strategic direction.

Three personas, three playbooks. Reps need speed and context in the flow of selling. Leaders need intelligence delivered for inspection and cross-book visibility. RevOps needs tools that operate at portfolio scale. A deployment that only optimizes for one leaves the others underserved.

Finding 3: Every persona uses all five jobs — that changes the platform decision

Finding 2 showed that each persona allocates differently. But here's what those percentages hide: all five jobs show up across all three personas at meaningful rates. No job is exclusive to any one role.

The heatmap shows what percentage of each job's activity comes from each persona. If usage were evenly distributed, each row would mirror the user share shown in the header. It doesn't.

Who drives each job: share of activity by persona

% of each job’s interactive activity by persona · (%) in header = share of total users

The deviations tell the story. Reps are 55% of users but 80% of Deal Acceleration. RevOps is 14% of users but drives 53% of Pipeline Inspection, four times their expected share. Leaders over-index on Conversation Readiness (30% vs. 25% user share). Team Enablement is the most evenly distributed job at 37/31/32, because coaching and digests serve everyone.

This overlap is where the platform argument gets concrete. When a rep does Account Intelligence, they're asking "What should I know about Acme before Thursday's call?" When a leader does it: "What's the competitive landscape across our enterprise segment?" When RevOps does it: "Which accounts have gone dark in the last 30 days?" Same data, different questions. If they're in different tools, none benefits from the others' work. On the same platform, the rep's call notes feed the leader's competitive analysis, and the leader's priorities inform which accounts RevOps monitors.

Fragmented tools create fragmented visibility. A shared platform creates a flywheel where every interaction, from every role, makes the next one smarter.

Finding 4: RevOps is the organizational leverage layer

RevOps drives 53% of Pipeline Inspection with just 14% of users. The interesting question is what happens when you give systems-thinkers the tools to build actual systems?

Even in chat, RevOps asks a different kind of question. Not "How do I win this deal?" but "Which deals across the pipeline have stalled without next steps?"

The real inflection comes when RevOps moves beyond chat into programmatic AI, where automated workflows run across CRM, call recordings, and enrichment systems without human intervention:

What programmatic AI workflows actually do

1,710 programmatic (MCP) interactions

Three-quarters of programmatic activity is deal management and content generation: automated pipeline scans, qualification audits, deal reports, and coaching documents running across the book without anyone having to ask.

The programmatic data shows what RevOps builds when you remove the ceiling: automated deal reviews that scan the pipeline weekly, coaching analyses that surface patterns across every recorded call, account monitoring that triggers proactive outreach.

One RevOps professional with programmatic AI capabilities creates leverage across the revenue team, not by doing their work for them, but by building the intelligence infrastructure they all run on.

Ryan Vanshur, VP of GTM Intelligence and AI Solutions at Handle, describes the result: "I set up automated weekly summaries for each rep. Every Friday they get a digest of their activity and next steps. Leadership gets the same intelligence automatically, without anyone having to ask."

What this means for your team

1. Start with account research, not a transformation roadmap

Account Intelligence and Conversation Readiness together represent 37.5% of all activity: the context that feeds every downstream deliverable and deal strategy. In the data, every organization started in the same place: account research. Pre-meeting context, account briefs, competitive landscape.

The organizational, high-trust jobs like pipeline management, automated digests, and QBR prep come last every time. You don't need to boil the ocean. Start with the research pain point, build trust, and expand from there.

2. Deploy for all three personas on the same platform

The practical test: can your leader see the real context from a rep's last call without asking for a status update? Can RevOps build automated reviews on the same intelligence reps use daily? If those require different tools, you're paying for fragmented visibility.

3. Treat RevOps as the AI infrastructure layer

The highest-leverage AI investment isn't buying every rep a copilot. It's empowering RevOps to build the workflows, analyses, and automated reviews that serve the revenue organization. Give them programmatic tools rather than just a chat interface, and the ROI multiplies across every seller and leader they support. Matt Baker, VP of GTM Strategy at Accuris, saw this firsthand: "In our first 90 days, we automated over 400 account plans, leading to a 2.5x increase in pipeline generation versus the prior year."

4. Measure breadth, not just adoption

The teams getting the most from AI aren't the ones with the highest login rates. They're the ones where users engage with the widest range of jobs. Track how many of the five jobs each persona touches: that breadth signals AI has become embedded in the actual work, not just one corner of it.

Where this is heading

Programmatic volume is growing. Automated agent workflows barely existed in the early months of this dataset and have grown steadily since. As more RevOps teams move from chat to programmatic access, the automation rate in Pipeline Inspection is a leading indicator, not a ceiling.

Product & Marketing is appearing on the same platform. Though still roughly 6% of users, their activity concentrates in Account Intelligence: competitive analysis, positioning research, customer insights drawn from the same data sellers use. When ABM targeting and competitive positioning share the same intelligence layer as Deal Acceleration, the go-to-market motion gets more coherent.

The organizational share keeps climbing. Team Enablement is already the single largest job, driven by automated digests that run without manual prompting. As more teams adopt programmatic workflows, the share of AI activity that's organizational rather than individual will continue to grow.

Underlying these trends is a shift in what's feasible. The contextual research that drives better outcomes was prohibitively time-consuming just a year ago. As context infrastructure improves, the ceiling on what teams can ask and act on keeps rising. The organizations building that foundation now will compound the advantage.

Use case glossary

Methodology

This analysis draws from more than 31,000 AI-powered interactions from hundreds of GTM professionals across enterprise organizations, collected between mid-2025 and early 2026 and supplemented by user interviews to validate behavioral patterns. Of those, roughly 25,000 were classified into specific use cases using weighted pattern matching against conversation content. These interactions arrived through three delivery channels: 15,434 interactive (chat) conversations, 1,710 programmatic (MCP) workflows, and 7,931 recurring digests. The remaining approximately 6,100 interactive conversations did not map cleanly to a single use case and are excluded from the classified base.

Use cases are grouped into five jobs-to-be-done: Account Intelligence, Conversation Readiness, Deal Acceleration, Pipeline Inspection, and Team Enablement. Personas were inferred from anonymized role categories and behavioral query patterns, then grouped into four GTM functions: Sales Rep, Leadership, RevOps & Enablement, and Product & Marketing. Persona-level findings are drawn primarily from interactive conversations where individual intent is clearest. No personally identifiable information was used at any stage of this analysis. All data is presented in aggregate.