update
May 20, 2026
By Teun
Anthropic sales leader uses Claude Cowork to score 4,000 accounts
Anthropic’s Travis Bryant says he uses Claude Cowork to prepare customer briefs, build weekly forecasts, and run an overnight scoring process for a 4,000-account territory. He says the system now handles data assembly and formatting work that previously took cross-functional teams hundreds of hours.
Travis Bryant, Anthropic’s head of US mid-market go-to-market, says Claude Cowork now handles work that used to consume much of his week. In a post describing his workflow, Bryant said he uses the tool to prepare customer briefs, assemble weekly forecast reports, and score a 4,000-account book overnight.
Bryant said his role covers 4,000 accounts split between mid-market tech companies and industries such as financial services, healthcare, retail, and manufacturing. He described the job as spanning three cadences: daily customer call prep, a weekly forecast rollup for Anthropic’s sales leadership, and quarterly territory and prospect-list work across the full account base.
Before Claude Cowork, Bryant said the surrounding work for those decisions often meant pulling data from multiple systems, reformatting reports, and rebaselining numbers whenever data changed. “Claude Cowork has shifted that balance,” he wrote, saying it handles the data assembly and formatting so he can spend more time on customer conversations and strategic decisions.
Bryant said he tried Claude Code first but did not feel comfortable working in the terminal. Claude Cowork, he said, presents the same engine in an interface he can actually use, which made it easier to hand off tasks and trust the work would get done.
Much of the time savings comes from scheduled tasks. Bryant said one morning task checks his Google Calendar and books a conference room for external meetings that do not already have one. Another task prepares customer call briefs before each meeting by pulling spend data from BigQuery and pipeline status from Salesforce into a document he can review when he opens his laptop.
He said the scheduler mattered as much as the underlying skill. Once a task stops being something he has to remember to trigger, he said, it stops falling through the cracks.
The bigger weekly gain comes from a Friday forecast workflow. Bryant said a scheduled skill pulls opportunity records and submitted commits from Salesforce’s Forecast tab, token spend from BigQuery, and notes from internal documents. It then builds a one-page web report in the format Anthropic’s sales leadership expects, with top-line metrics, top deals, movers and decliners, and a forecast snapshot from each first-line manager.
By Monday’s forecast call, Bryant said the report is already published to an internal link. His job is to add commentary. “Claude builds the what; I do the why,” he wrote.
Bryant said the largest project he has run through Claude Cowork is account propensity scoring for the whole mid-market segment. Every fiscal year, he said, each account needs a score to help the assigned account executive prioritize the territory.
He said he and Claude first built two five-dimension scoring rubrics, one for tech accounts and one for industries. The tech rubric included agent opportunity, internal transformation, AI commitment, white space against existing spend, and industry fit. The industries rubric used different factors, including knowledge-worker density and public AI commitments measured by mentions on a company’s open jobs page.
After setting the rubric, Bryant pointed Claude Cowork at the 4,000-account list. He said the system ran overnight, scoring each account one by one using deep web research, Salesforce data, and BigQuery data, and produced both a numerical score and a written rationale for every dimension.
He then asked Claude Cowork to turn the results into an interactive dashboard. According to Bryant, each account executive can click into their territory’s slice, see ranked accounts with the rationale behind each score, and hover over an account to surface possible use cases and comparable case studies for prospecting.
Bryant said none of the prompts required technical language. The pattern, he wrote, was to define the dimensions, test a territory, review the output, adjust the weights, and repeat. Anthropic’s Sales plugin, he said, ships with baseline skills that can be customized to match how a team actually works.
He said the two patterns he would carry to other sales teams are scheduling routine prep and running larger strategic projects as overnight Claude Cowork jobs. Examples he listed included TAM sizing, account research, and comp benchmarking. In his telling, the tool has replaced data assembly and report formatting with time he can spend on customer relationships and judgment calls.
