Build scored sales lead profiles and outreach drafts
The full prompt framework, a scoring system, and custom Gem instructions for turning prospect research into outreach decisions. Task: The author built a sales research pipeline using Gemini Deep Research, NotebookLM, and Google Sheets to evaluate potential clients for AI consulting. A custom Gem named 'Sales Lead Researcher' executes two commands: 'Research [Company Name]' to gather intelligence and 'Score [Company Name]' to generate a 0-16 qualification score, decision-maker profile, and outreach draft. The workflow requires pre-defined scoring dimensions (Budget, Need, Timing, Accessibility) and specific service positioning before execution. Human involvement includes defining the initial research lens, reviewing the generated scores against evidence, and approving the final outreach message before sending.; tools: Gemini. Human role: The original post does not say how a person stayed involved.; limitations: The workflow relies on Gemini Pro ($20/month) and has daily usage limits on Deep Research. It is not suitable for open-field discovery; it only amplifies existing awareness of specific target companies. The quality of output depends heavily on the user's pre-defined qualification framework and service specificity. If the user does not have a clear value proposition, the AI produces 'beautiful, thorough, completely useless' reports. Human review is critical to validate the 'Need' score, as high maturity can falsely appear as high fit..
Tools
Published on Agent Work Atlas:
Overview
The full prompt framework, a scoring system, and custom Gem instructions for turning prospect research into outreach decisions.
Workflow
The author built a sales research pipeline using Gemini Deep Research, NotebookLM, and Google Sheets to evaluate potential clients for AI consulting. A custom Gem named 'Sales Lead Researcher' executes two commands: 'Research [Company Name]' to gather intelligence and 'Score [Company Name]' to generate a 0-16 qualification score, decision-maker profile, and outreach draft. The workflow requires pre-defined scoring dimensions (Budget, Need, Timing, Accessibility) and specific service positioning before execution. Human involvement includes defining the initial research lens, reviewing the generated scores against evidence, and approving the final outreach message before sending.
Reported result
The pipeline was tested on three companies, producing scores of 14/16, 13/16, and 9/16, which corresponded to 'reach out now', 'not relevant due to advanced maturity', and 'keep on radar' respectively. The author reports that the process takes under 20 minutes per company, compared to 45 minutes manually. However, the post does not provide data on actual conversion rates, reply rates, or revenue generated from these leads. A measurable acceptance check would be to track the percentage of '14/16' leads that result in a booked meeting within two weeks.
Human role
The original post does not say how a person stayed involved.
Limitations
The workflow relies on Gemini Pro ($20/month) and has daily usage limits on Deep Research. It is not suitable for open-field discovery; it only amplifies existing awareness of specific target companies. The quality of output depends heavily on the user's pre-defined qualification framework and service specificity. If the user does not have a clear value proposition, the AI produces 'beautiful, thorough, completely useless' reports. Human review is critical to validate the 'Need' score, as high maturity can falsely appear as high fit.
Try it
Act as a Sales Operations Specialist. I need to evaluate [Company Name] as a potential client for [My Service Offering]. 1. **Preparation**: First, ask me to define my specific service offering and the four scoring dimensions (Budget Capacity, Need Intensity, Timing, Accessibility) with 0-4 anchors for each. Do not proceed until I provide these. 2. **Research**: Use Gemini Deep Research to gather data on [Company Name], focusing on [My Service Offering] signals, recent funding, leadership changes, and tech stack. Cite 50+ sources. 3. **Scoring**: Analyze the research to assign a 0-4 score for each dimension. Calculate the total 0-16 score. Identify the key decision-maker and their recent public activity. 4. **Drafting**: Create a personalized LinkedIn outreach draft under 80 words referencing the decision-maker's specific activity. 5. **Review**: Present the score, evidence, and draft to me. **Do not send anything.** Wait for my explicit confirmation to approve or adjust the score/outreach. If inputs are missing, ask clarifying questions. If Deep Research hits limits, switch to standard Gemini Chat with provided context.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.