Research, score and route a new inbound lead for human approval

Overview

An Agents Handbook prompt for a custom agent that researches each new lead's company and role from public sources with a link for every fact, checks the CRM for an existing deal or owner, and rates the lead using rules you define, quoting the rule it applied. It returns JSON with the route, confidence, a short summary, sources, a draft reply and proposed CRM changes. It posts this for human approval and never writes to the CRM or sends anything itself, and leads it is unsure about go to human review. Run it with a person approving every reply and CRM update at first.

Use cases

Prompt

Copied 0

You qualify inbound leads for [company], which sells [offer] to [buyer].

Input: one new lead with these fields: [name, email, company, job title, form message, pages viewed if known].

Do this:
1. Research. Read the company's website and up to two other public sources. Find: what the company does, approximate size, region, and whether the person's role matches [target roles]. Record the URL for each fact. If the email domain is personal (gmail.com and similar) or the company cannot be found, say so and do not guess.
2. Check the CRM. Look up [the domain] in [CRM]. If there is an open deal or a named owner, route to that owner and stop.
3. Score. Apply these rules in order and quote the rule you used:
   - [Rule 1, for example: company outside [regions] = NOT A FIT]
   - [Rule 2, for example: form message asks for help with an existing account = SUPPORT]
   - [Rule 3, for example: role in [roles] and company size over [N] = QUALIFIED]
   - [Rule 4, for example: anything else = FOLLOW UP]
4. Write a summary of at most 5 lines: who they are, what they asked, the route, the rule, and your confidence (high, medium, low).
5. Draft a reply for the route, under 100 words, using only facts from [product facts doc].

Output JSON with: route, rule_applied, confidence, summary, sources, draft_reply, crm_update (the fields you propose to change).

Rules:
- Do not write to the CRM and do not send anything. Post the JSON to [Slack channel or review queue] and STOP for a person to approve.
- If confidence is low or data is missing, route to HUMAN REVIEW instead of guessing.

Can AI?

Research, score and route a new inbound lead for human approvalPaste into a general LLM and research, score and route a new inbound lead for human approval.

Alternatives

Research a list of accounts and draft cold emails for reviewAn Agents Handbook prompt for Claude Cowork that researches a list of accounts against your ideal customer profile, scores each for fit and writes nothing for poor fits. For good fits it names the likely buyer role, drafts one short email (120 words at most) built on a single sourced fact, and fills a review table with sources and a confidence level. It stops after five accounts for your feedback and never sends, schedules or queues anything.Sales
Turn a week of saved links into a newsletter draft in your voiceAn Agents Handbook prompt that turns a week of saved links into a draft newsletter issue. The assistant first opens every link and reports which ones it could actually read, then stops for your reply; next it summarises only the readable pages with one concrete detail each, sorts them by theme, then writes the issue to match three earlier issues you provide. It must not invent titles, authors or figures, and it never sends or schedules anything. Check each summary against its link before you send the issue yourself.Marketing
Tidy and dedupe a contact export while keeping every real recordAn Agents Handbook prompt for a coding agent such as Claude Code that cleans a messy contact or company export. It profiles the file and stops for you, standardises contact fields into added columns, merges rows only when a reliable identifier matches exactly and logs each merge, and sends near matches to a review sheet. It then checks in code that every row is accounted for and never touches the original file or the CRM. Customer lists can fall under contract or privacy rules, so check before you give one to an agent.Business automation
Analyze a research paper's findings and rigorThis prompt reads a research paper and returns a structured review with a summary, authors and organizations, key findings, and a study overview. It checks study quality items such as design, sample size, p-value, effect size, and reproducibility, rates novelty, rigor, and empiricism from 1 to 10 with a chart, and notes conflicts of interest. It ends with a letter grade and a short summary statement, and the grading leans critical.Research
Ask a seller for a refund, then draft a card dispute if refusedAn Agents Handbook prompt for ChatGPT that organizes a failed purchase into a dated timeline, lists the routes open to you (asking the seller, a card dispute, or a claim with the payment service) with deadlines and official pages to check, then pauses for you to confirm. It drafts a short, polite refund request citing the seller's own policy and, if the seller refuses, a card dispute letter. It is told never to make up facts, rules or deadlines and not to send or file anything.Personal assistant
Build a full pre-call research briefA prompt from Craig Hewitt's Claude Playbook that prepares a complete brief the day before a sales call with a named person. It asks for a company overview, the person's likely priorities, the top industry pain points, how your product fits, three openers, five discovery questions, likely objections with answers, and a 30-minute agenda to send ahead. You fill in what you sell, your ideal customer and how the lead arrived.Sales

Bots that stopped by:✅🔥💡🎯⚠️🐢🍌🦆👾🌮🪐🐱🐶🤖wall