Outreach Automation — AI Lead Engine in Make & Lemlist

An outbound engine built in Make that turns a raw scraped lead list into personalised, ready-to-send sequences. Leads are cleaned, qualified, enriched, written up by GPT and pushed into Lemlist and Instantly, with Google Sheets kept in sync as the single source of truth.

Cold outbound breaks down at scale for two reasons: personalisation is manual, and lead data goes stale across the tools that hold it. Writing a genuinely relevant opener takes minutes per prospect, and once a list is split between a spreadsheet and two sending platforms, nobody knows which version is correct.

Ingest: scraped leads land in Google Sheets and are cleaned and de-duplicated before anything else runs.

Qualify: a router checks each lead against target criteria, so only genuine fits move down the pipeline and the rest are parked rather than emailed.

Enrich: company and role signals are pulled in and normalised into structured JSON the model can reason over.

Write: GPT generates a personalised opener and follow-up variants from that context, with a fallback path for records where the data is thin.

Parse and iterate: the response is parsed as JSON and iterated per lead, so a single run handles a whole batch.

Sync: every lead is written back to Google Sheets and pushed by API into Lemlist and Instantly, keeping the sending platforms and the sheet in step.

Error handlers and fallback routes sit on each API call, so one bad record cannot stop the scenario, and every step writes its status back to the sheet, which makes each run auditable.

Make, OpenAI GPT, Google Sheets, Lemlist, Instantly

Personalised outbound at list scale without a copywriter in the loop, and one place to look when you want to know where any lead actually is.