Where AI fits into the workflow
Prospecting 8 min
0/28
Business problem
Salespeople throw AI at a single task (writing an email) and wonder why the results are weak. AI works best not as one tool, but as a layer running through the whole workflow — provided each stage gets the right task and the right data.
Proven method
AI has 6 natural spots in the workflow:
- Discovery — generating queries, company lists, categories (AI thinks 'broad'),
- Research — extracting data from a URL, synthesizing information (AI reads fast),
- Enrichment — completing data from multiple sources (AI connects),
- Qualification — assessing ICP fit against criteria (AI classifies),
- Scoring — assigning points with a justification (AI calculates + explains),
- Personalization — creating the first sentence from research (AI writes). The rule: AI gets a specific INPUT and returns a specific FORMAT (list, table, JSON, sentence). The more precisely you define the format, the better the result.
Market example
- BAD
- Unstructured prospecting without ICP filters and generic outreach to random accounts.
- GOOD
- Weak: 'help me with prospecting' (AI doesn't know what you want). Good: 'based on the company URL list 3 problem hypotheses and 1 personalization, format: bullets'. The same AI that returns word salad on a vague command returns ready research on a precise one. The difference isn't a stronger model — it's a better-defined task.
Ready-to-use prompt
You are a prospecting process architect. Context: My workflow has 9 stages. I sell [PRODUCT] for [SEGMENT]. Task: Design which of the 6 AI areas (discovery, research, enrichment, qualification, scoring, personalization) I connect to which workflow stage: 1. For each of the 6 areas: 1 sentence on what AI does, 1 sentence on what the result format is, 1 sentence on what the human checks 2. Point out the area that will save the most time in my case Format: 6 blocks of 3 sentences.
Model rekomendowany: Claude 3.5 Sonnet / GPT-4o Copy and paste into ChatGPT, Claude or Perplexity
ALTERNATYWNE SCENARIUSZE
Variations & alternative scenarios
Small-team version: point out which AI area a junior should handle and which a senior should do.
Budget version: rank the 6 areas by return (time saved / API cost).
Practical exercise
Take your 5 activities from lesson 1.1 and assign each to one of the 6 AI areas (or 'stays with the human'). Write down which area you'll start automating first.
Zapisano w przeglądarce