← WorkSystem concept

AI Lead Engine

Sales Systems

A pipeline that handles research, enrichment, fit scoring, and drafted personalisation, so the human effort goes into the conversation.

  1. Research
  2. Enrich
  3. Score
  4. Personalise
  5. Outreach
  6. CRM

The problem

Sales teams spend most of their week preparing to sell — researching, retyping, rewriting, and updating records after the fact.

Key components

  1. 01

    Source and research

    Builds the target list from defined criteria and gathers the context a rep would otherwise look up manually.

  2. 02

    Enrichment

    Fills in firmographic and contact detail from the sources the business already licenses.

  3. 03

    Fit scoring

    Ranks accounts against the profile that has actually converted, not against a generic ideal.

  4. 04

    Personalisation drafting

    Drafts the specific opening based on real signals, and leaves the send decision to a person.

  5. 05

    CRM maintenance

    Keeps records current as a by-product of the pipeline rather than as a separate discipline.

System design

The work around the conversation

Selling is a conversation. Almost everything surrounding it is retrieval, formatting, and record-keeping. That surrounding work is what expands to fill the week, and it is the part that automates cleanly.

System design

The pipeline runs continuously rather than in campaigns. Accounts enter, get enriched and scored, and surface in a queue when they meet the threshold. The rep works a prepared queue instead of a raw list.

  1. Research
  2. Enrich
  3. Score
  4. Personalise
  5. Outreach
  6. CRM

Personalisation with a human in the loop

Drafted personalisation is generated from concrete signals — a role change, a job posting, a published case study, a product launch. Generic praise is worse than no personalisation, because it signals automation without providing value.

The send stays with the person. A rep reviewing twenty prepared drafts is doing a different job than a rep researching twenty companies from scratch, and the review is where their judgement is worth most.

Fit scoring against reality

Scoring should be built from accounts that actually closed, not from the profile the team believes it sells to. Those two lists differ more often than expected, and the difference is usually the most valuable output of the first month.

Expected effect

  • Preparation time per account drops sharply
  • Outreach references something real rather than a merge field
  • Fit scoring is applied consistently across the whole list
  • CRM records stay current without manual upkeep

Stated qualitatively on purpose. This is a reference architecture, and quantified results depend entirely on the business it is built into.

Every businessruns on a system.Most are accidental.

If any of this resembles how your operation works today, the next step is a conversation about where it actually breaks.