AI WhatsApp Re-engagement

Existing re-engagement journeys were largely email-based. Email remains useful, but it can be a relatively passive channel when a lead has a specific question or needs a timely response. The business needed a way to follow up based on a lead's status, handle replies and make sure worthwhile conversations reached the right person.

I designed and built the system independently over approximately two to three months, spending substantial time testing scenarios and refining prompts and conversation rules. This was iterative prompt and workflow development — not model fine-tuning.

The workflow connects Wati, Claude and Salesforce. Leads can enter status-specific follow-up journeys, with configurable delays and messaging designed around the reason they had not previously progressed.

When a lead responds, an LLM helps handle the conversation within defined scenario objectives. Once the objective is met, the system creates a callback task in Salesforce and provides the relevant conversation transcript to the lead owner.

The goal is not to replace the salesperson or pretend the bot can close every conversation. It is to identify when someone is ready for human follow-up and make that hand-off useful.

The system is live and has generated substantially more callback opportunities than the earlier email-led approach, with reactivated leads also progressing to contracts. The exact commercial figures are private and should not appear on the public site.

The practical benefit is a more connected journey from automated follow-up to a human conversation, with less context lost between the two.

The most effective AI workflow is usually the one with a clearly defined boundary. In this case, the bot's job is to re-open a useful conversation and hand it back to a person at the right moment — not to act as an unrestricted sales agent.