End-to-end pipeline automation for 250+ AEs
The company's 300 account executives were spending an average of 2.4 hours per day on tasks that weren't selling: researching prospects before calls, personalizing outreach templates, updating Salesforce fields after meetings, scheduling follow-up sequences, and hunting for relevant case studies or product content to attach to emails. A time-motion study commissioned before the engagement found that 31% of an AE's working day — across all 300 reps — was consumed by these activities. At an average OTE of $180,000, this represented approximately $16M in annual salary cost generating zero direct revenue.
The CRM data quality problem was severe and self-reinforcing. AEs who were time-pressured updated Salesforce minimally — deal stages moved, but contact details, account intelligence, and interaction notes were sparse. This meant that any tool relying on CRM data as context was starting from a poor foundation. The VP of Sales estimated that 40% of accounts had materially incomplete or stale data. Pipeline forecasting was unreliable as a result. Sales managers spent significant time in 1:1s reconstructing deal state from memory rather than reviewing accurate CRM records.
"31% of an AE's working day was consumed by tasks generating zero direct revenue. At 300 reps, that's $16M in salary cost not selling."
We built a sales agent with direct tool access to Salesforce, LinkedIn (via API), the company's email system, their content library, and their meeting scheduling infrastructure. The agent can read and write across all five systems — not just retrieve information, but take action. The agent's primary workflow: when an AE prepares for a call, they trigger the agent with the account name and meeting context. The agent pulls Salesforce history, scrapes recent LinkedIn activity from the prospect and their company, searches the content library for relevant case studies, and generates a pre-call brief with talking points, risk flags, and suggested questions — all in under 90 seconds.
Post-meeting CRM updates are now agent-handled. The AE records a 2-minute voice note after each call. The agent transcribes, extracts key facts (decision timeline, stakeholders mentioned, next steps, objections raised), maps them to Salesforce fields, and drafts the follow-up email sequence — ready for AE review and one-click send. Salesforce data quality improved measurably within 60 days. Field completion rates went from 58% to 94% for active opportunities. Sales managers could now run accurate pipeline reviews from CRM data rather than reconstructing from memory.
For prospecting, the agent generates personalized outreach based on prospect LinkedIn activity, company news, job postings (as a proxy for strategic direction), and prior interaction history. Personalization is substantive — referencing the prospect's recent post, a company milestone, or a shared connection — not cosmetic first-name insertion. Reply rates on agent-drafted outreach were 34% higher than the AE-written baseline in an A/B test run over 6 weeks with 1,200 prospects.
"Reply rates on agent-drafted outreach were 34% higher than the AE-written baseline. Personalization was substantive, not cosmetic."
The 30% reduction in admin time per rep freed approximately 45 minutes per day per AE for selling activity. Within the first quarter, the team's outbound activity volume increased 28% without headcount change. The $2.4M incremental ARR figure represents new business closed in Q1 by the pilot cohort of 60 AEs — attributed to increased outbound volume and improved prospect conversion from better-prepared calls. The CRM data quality improvement had a second-order effect on forecasting accuracy. The CFO reported that pipeline forecast variance dropped from ±34% to ±12% in the quarter following deployment — enabling more confident resource allocation decisions.
AE adoption was the highest-risk factor in the engagement. Sales teams are notoriously resistant to workflow changes. The agent was positioned not as a replacement for AE judgment but as a research and prep tool — the AE always reviews and approves before anything sends. Within 8 weeks, 89% of the pilot cohort was using the agent daily for pre-call briefs. The usage pattern that took longest to adopt was post-meeting voice notes for CRM update — AEs habituated to the keyboard found voice input awkward initially. Adoption reached 76% for this feature by week 12.