Flow view
Step 1 of 12·Diagnose·Week 1
01

Where the revenue is leaking today

Before we build anything, we name the leaks — in dollars.

Outcome we're aiming for

A Discovery Report agreed with leadership: every revenue leak quantified, backed by evidence from hundreds of real customer conversations — plus the automation rules we'll build in the next 30 days.

0/ 27tasks

AI Analysis — three approaches we can take

Pure Water already runs sales on Aircall. There are three different ways to add AI on top. We'll walk through all three in the kick-off so you can choose based on what you already pay for and how much custom logic you want.

Option A — HubSpot Conversation Intelligence (native)

  • What it does: native call summaries, keyword tracking and coaching insights inside HubSpot, fed by the Aircall ↔ HubSpot integration
  • Requires: HubSpot Sales Hub Professional or Enterprise + Aircall connected
  • Limitations: built for sales coaching and summaries — not for custom lead-scoring or automatic next-action routing
  • Best for: teams already on Sales Hub Pro+ who only need rep coaching and high-level summaries

Option B — Aircall AI / Aircall Transcription add-on

  • What it does: Aircall-native transcription, sentiment, topic detection and post-call summaries, pushed into HubSpot automatically
  • Requires: Aircall AI add-on on the Aircall subscription
  • Limitations: tied to Aircall's roadmap and field structure — the scoring and routing logic still has to be built on top in HubSpot
  • Best for: teams who want plug-and-play transcription with minimal custom build

Option C — Custom AI layer (our recommendation)

  • What it does: pull Aircall transcripts via API → run them through an LLM (OpenAI / Anthropic / Gemini) → push intent score, structured fields and next-best-action back into HubSpot
  • Requires: Aircall API access + an LLM API key (low monthly cost — pay per call analysed)
  • Limitations: we build and maintain the pipeline (included in our scope) — slightly more setup than a native add-on
  • Best for: Pure Water — gives full control over lead scoring, auto-routing of next actions, and the same pipeline powers Step 05

Our recommendation

Start with Option C (Custom AI layer). It's the only option that powers Step 05's automatic lead scoring + next-action routing — native tools can't do that out of the box. If Pure Water is already paying for Aircall AI (Option B), we don't replace it — we use its transcripts as the input into our AI layer, so you're not paying twice for the same capability.

What's inside the Discovery Report

  • Top customer intents — what callers actually ask for, ranked by frequency
  • Objection map — the real reasons deals stall, in customers' own words
  • Win vs loss conversation patterns — what closers say that the rest of the team doesn't
  • Missed follow-up list — named deals with estimated dollar value left on the table
  • Rep performance snapshot — talk time, objection-handling, follow-through
  • Recommended lead-scoring rules — ready to load into HubSpot in Step 04
  • Recommended automation triggers — ready to wire up in Step 05

Engagement options

  • One-time discovery audit — included in the 30-day scope. Runs once, informs strategy, builds the automation rules.
  • Ongoing monthly Call Intelligence Report (optional add-on) — same pipeline runs every month in our ongoing engagement, delivering a dashboard + report showing trend shifts, new objections emerging, rep drift, and fresh automation opportunities.

The leaks we typically find

  • Pricing-guide downloads go cold after the PDF is sent
  • Instant-quote requests wait hours before sales picks them up
  • Whole-house, under-sink and rainwater leads all get the same generic follow-up
  • Aircall conversations never make it back into the CRM — context is lost
  • Lost deals are never re-engaged
  • Past customers hear nothing until a filter actually fails

Deliverable

A Discovery Report quantifying lost revenue per stage, backed by evidence from real Aircall conversations, plus the automation rules we'll build next. Every pillar that follows is tied to a number — not a feature.