A downloadable agent you connect to your own data
For the deals with gaps, the agent drafts a specific next move for the manager to review and run. It reports movement by rep and by deal: what advanced, what sat still, and what regressed. To set it up, download it from Blueprint’s GitHub and point it at your HubSpot; the runbook takes you through the rest. Blueprint portfolio companies can submit the implementation interest form on this page to ask the Build Team to help customize the agent or set it up end to end.
The Sales Coaching Agent ships as a repository you download: a step-by-step runbook, three Claude skills, the scoring rubrics, and the supporting scripts. Setup runs in about a couple of hours. You connect it to your own HubSpot and call-recording data; it runs on your own machine against those systems, with no new vendor or subscription. The runbook walks you (or us, working with you) through every step.
HubSpot, Claude Code, Google Drive.
HubSpot CRM with deals, emails, and call recordings in one place; a Mac that can leave Claude Code running; a Google Drive folder where reports get delivered.
Connecting HubSpot and Drive, installing the skills and scripts, tuning the rubric language to your sales motion, a supervised first run, and scheduling the recurring cadence.
What you get (examples)
Anonymized deliverables from the pilot engagement. Open any preview to see the full PDF.
Two review modes built on the same scoring method
Both modes score deals against a rubric of selling behaviors, and both require verbatim evidence from emails, transcripts, and notes for every score, so the output reflects the work that was actually done rather than the fields that were filled in.
The Sales Effectiveness Audit
A retrospective review of your closed deals.
Scored across 10 dimensions: six quality (discovery depth, objection handling, stakeholder engagement…) and four strategy (pricing-tier discipline, next-step commitment, multi-threading…). The output is an executive brief on which behaviors correlate with winning in your data, and where documented work diverges from the CRM record. Useful as a one-time baseline before turning on the recurring agent.
The Sales Manager Agent
A recurring review of your open deals.
Scored on 8 coaching dimensions: current solution, pains, prospect outcome, org outcome, stakeholders, BANT, next step, and proof-point. Runs on a schedule you set (the default is Monday and Thursday nights, with a Friday week-over-week progress report). Each deal gets a specific suggested coaching action for the manager to review: who to contact, what to ask, and when to move on.
The output reflects the work that was actually done rather than the fields that were filled in.
What it surfaced at one portfolio company
We ran the audit at one of our portfolio companies before standing up the recurring agent. A few of the findings, to give a sense of what this kind of review can surface:
The competitor-intel field was populated on 97.5% of deals; conversation evidence supported it on 14%. The gap between the two turned out to be a useful coaching signal in itself.
Post-demo deals with one engaged contact won 3.3% of the time; deals with three or more engaged stakeholders won 33.3%.
Deals above $7.5K ACV grew 2.8× year over year while their win rate declined from 17.2% to 8.6%, pointing to a segment that needed its own playbook.
Pilot engagement: full-year cohort of 132 closed deals.
A note on scope. These are one company’s numbers, not benchmarks. The point of the audit is to find the patterns in your data. The dimensions, thresholds, and coaching priorities will differ by team and motion.
What a run looks like
Each run scores every open deal on the 8 dimensions (Yes, Partial, or No, with the supporting evidence) and produces a heatmap plus one suggested coaching action per deal. The excerpt below is from a real run at the pilot company, anonymized.
| Deal | Stage | Amount | Current solution | Pains | Outcome | Stakeholders | BANT | Next step | Proof-point |
|---|---|---|---|---|---|---|---|---|---|
| Deal 1 | Demo scheduled | – | · | · | · | · | · | ~ | · |
| Deal 2 | Demo scheduled | – | · | · | · | · | · | Y | · |
| Deal 3 | In review | $15,750 | Y | Y | ~ | Y | ~ | · | · |
| Deal 4 | In review | $11,000 | Y | ~ | ~ | ~ | ~ | Y | · |
| Deal 5 | Demo completed | $8,500 | ~ | · | · | ~ | · | · | · |
Legend: Y = documented with evidence · ~ = partial · · = not yet documented. Source: anonymized coaching run at the pilot company.
Deal 3 illustrates the kind of read the heatmap gives a manager: seven people attended the demo and the pain is well documented, but there’s no agreed proof-point and no next meeting on the calendar, a useful prompt for the next 1:1.
The suggested actions are specific to each deal
Deal 3: agree on a proof-point
“Email both stakeholders by EOD Friday: given your current tool doesn’t track activities the way you need, what’s the single workflow you’d need to see working live to make the case to leadership? Book 30 minutes for exactly that, and narrow the proposal to one recommended package.”
Deal 5: change the channel or close the loop
“Three follow-ups since the demo with no reply. One final move: reach the second decision-maker the prospect named, plus a closing-the-loop email. If no reply by Friday, mark closed-lost and revisit on the prospect’s timeline.”
Delivered to a Drive folder on a schedule
Every run lands in a dated Google Drive folder automatically. The Friday run adds a week-over-week progress report showing which dimensions improved, which deals advanced, and where documentation slipped.
Scored CSV
Every open deal, all 8 dimensions, the supporting evidence, and the suggested coaching action. The authoritative record.
Branded PDF report
The readable version for pipeline reviews.
Color-coded XLSX
The heatmap as a spreadsheet.
Progress report
Week-over-week movement across reps and deals.
Available to the portfolio
The package is in a private repository available to all Blueprint portfolio companies: the 11-step runbook, three Claude skills, six supporting scripts, the complete rubric set, and scheduled-task templates. Reach out to the Build Team and we’ll share access and work through the setup with you, including tuning the rubric language to your sales motion.
Where to start. If you want a sense of what the system would say about your team before committing to the recurring agent, start with the audit on a recent cohort of closed deals. It takes about a week, requires nothing from your reps, and produces the baseline the coaching agent then works against.
Want us to walk you through it?
You can set it up yourself from the runbook in the repository. If you would rather we do it with you, tell us here.