Tech Stack
Everyone talks about AI saving time. Nobody shows the receipts. Here's exactly how I saved 15 hours per week for every member of a 20-person sales team.
The 15 Hours — Broken Down:
1. Lead Research: 4 hrs → 10 min (saved: 3.8 hrs) Before: Reps manually Googled every lead, checked LinkedIn, estimated company size, and wrote notes in the CRM. After: An n8n workflow triggers on new lead creation. Claude enriches the lead with company data, recent news, tech stack signals, and a suggested opening line. By the time a rep sees the lead, the research is done.
2. Follow-Up Emails: 3 hrs → 15 min (saved: 2.75 hrs) Before: Reps wrote individual follow-up emails, often forgetting leads entirely. After: The system tracks every conversation. When a follow-up is due, Claude drafts a contextual message referencing the last interaction. Rep reviews, clicks send. Or enables auto-send for cold leads.
3. Data Entry & CRM Updates: 2.5 hrs → 0 (saved: 2.5 hrs) Before: Reps manually logged calls, updated deal stages, and tagged leads. After: The system listens to call transcripts (Twilio recording → Whisper transcription → Claude extraction). It auto-updates the CRM with call summary, next steps, deal stage, and follow-up date.
4. Report Generation: 3 hrs → 0 (saved: 3 hrs) Before: The team lead spent every Friday afternoon building a report in Google Sheets. After: An n8n workflow runs every Monday at 8 AM. It queries Supabase for pipeline data, generates a narrative summary with Claude, and drops a formatted report into Google Sheets + Slack. Nobody touches it.
5. Meeting Prep: 2 hrs → 15 min (saved: 1.75 hrs) Before: Reps scrambled before client calls, re-reading email chains and CRM notes. After: Before every scheduled call, the system generates a one-page brief: client history, deal status, last conversation summary, suggested talking points, and potential objections. Delivered to the rep's inbox 30 minutes before the meeting.
6. Lead Qualification: 1.5 hrs → 20 min (saved: 1.2 hrs) Before: Reps eyeballed leads and made gut decisions about who to call first. After: AI scoring ranks every lead by conversion probability. Reps start their day with a prioritized list. No more calling cold leads while hot ones go stale.
Total: 16.25 hrs saved, rounded to 15 hrs/week per rep.
What made this work: Not one big AI tool — a *system* of small, connected automations. Each one saves 1-3 hours. Together, they transform the entire workflow. The stack is boring on purpose: n8n, Claude, Supabase, Gmail, Google Sheets. No exotic tools. No vendor lock-in. Everything the team already uses.
The meta-lesson: AI doesn't replace your team. It removes the tasks your team hates — so they can do the work they're actually good at.
Want to build something like this?
I architect and deploy end-to-end AI systems — from MVP to revenue.
Let's Talk