amberstack/

Series A vertical SaaS, 60 employees

Cut SDR triage from 4h/day to 25min with a Claude-graded inbox

[PLACEHOLDER] this is template content - replace before launch
stack
n8n, Claude Code, Postgres, Hubspot
build time
9 working days
build cost
€11,400 fixed
result
Triage time down 88% (4h → 25min/day). AE no-show rate down 31%.

[PLACEHOLDER] this case study is template content. Replace the entire file before linking it from anywhere public.

The problem

[PLACEHOLDER] One paragraph on what the client was doing before. Specific: hours per week, dollar cost, named teams. No generic adjectives ("inefficient", "manual").

What I built

[PLACEHOLDER] A short technical description with one diagram or workflow screenshot inline. List the n8n nodes, the model used, the database schema if it's the interesting part.

// Example: the scoring prompt is loaded from Postgres at request time
const prompt = await sql`
  select template
    from prompts
   where name = 'sdr_score_v3'
     and active = true
   limit 1
`;

Results

[PLACEHOLDER] Concrete numbers, before/after. Time, money, accuracy. One screenshot of the dashboard or Langfuse trace if relevant.

  • Triage time: 4h/day → 25min/day
  • AE no-show rate: 8.2% → 5.7%
  • Cost per workflow run: €0.014

What it cost to run

[PLACEHOLDER] Monthly infra cost broken down: Hetzner VM, Postgres, Anthropic API, Langfuse. Be specific. Founders care about this column.