Client result · Investor education
819 conversations. 583 replies. A 2.6-minute median first response.
A twelve-week Conversation Agent audit that exposed where valuable conversations were being dropped.
The assignment
Not a new funnel. A proper read of the one already running.
This business had an established WhatsApp Conversation Agent and a growing international lead flow. The question was not whether people would reply. It was what the replies revealed, where the handovers were stalling, and how much opportunity was sitting in the gaps between a useful answer and the offer.
The work audited exported conversation logs rather than relying on opening rates or a dashboard summary. It turned a long list of chats into a clear picture of intent, blockers and missed routing.
How the figures were counted
The data came from the conversation log, not a vanity dashboard.
Every figure was calculated directly from exported WhatsApp logs covering 4 June to 28 August 2026. Three internal test records were removed from every total. Where a lead pressed a button but the log showed no reply, that conversation was counted as stopped at that point. It is deliberately conservative.
What happened
The opening worked. The routing did not go far enough.
- Conversations opened
- 819 in 12 weeks
- People who replied
- 583, or 71%
- People who shared their investing stage
- 578, or 71%
- People who named a real blocker
- 450, or 55%
- Median time from first message to first reply
- 2.6 minutes
- Replies received outside weekday office hours
- 67%
- People shown the relevant offer
- 102
348 people named a problem and never saw the offer.
This was not a lead-quality issue. It was a routing setting. Six of seven blocker routes delivered a helpful answer and then ended the conversation before the relevant next step was introduced.
What the audit changed
Make the answer useful, then make the next step relevant.
The recommended change was not another campaign. It was to route all seven blocker paths towards the relevant offer after the video that addressed each concern. That moves the offer from 102 people towards the 450 people who had already described what they needed, without buying more traffic.
The audit also identified 40 typed messages that had not received a response. A free-text alert and a 24-hour nudge became clear human handover and recovery requirements.
What this does not claim
A useful audit is not a revenue forecast in a smart jacket.
This case study reports conversation, response and routing data. It does not attribute revenue, claim a close rate or turn a plausible opportunity into an invented forecast. The important result was the evidence needed to make the next commercial decision with confidence.
Less overthinking. More talking.
Your leads are already telling you what they need. Can your system hear it?
Message us and tell us where a useful conversation currently falls out of the process.
Prefer email? paul@clientfunnels.co.uk