Generating a list of prospects is only the first stage of lead generation. Once raw leads have been sourced, sales and business-development teams still need to determine which contacts are genuinely relevant to their ideal customer profile before investing time in research, personalisation, or outreach.
This use case explores how an AI-powered lead qualification workflow can assess prospect fit using available profile and company information, filter out unsuitable contacts, prevent duplicate records, and route qualified leads into the next stage of the sales process.
The Problem
Once raw leads are sourced, teams still need to determine which prospects are worth pursuing. This often involves reviewing LinkedIn headlines, job titles, company information, location, tenure, and industry details against the business’s ideal customer profile.
At scale, this process becomes difficult to manage manually. Outreach teams can spend time on contacts who do not match the intended audience, qualified leads may be inconsistently assessed, and duplicate records can enter the sales process more than once.
Without a structured qualification step, businesses may generate high volumes of lead data without a reliable way to identify which prospects should move into research, personalisation, or outreach..
The Solution
The LinkedIn Lead Qualifier Agent receives batches of raw leads from a sourcing workflow or connected system. It uses an LLM to assess whether each lead fits the Ideal Customer Profile, removes duplicate qualified records, logs approved leads in Google Sheets, and sends them to a downstream research or outreach workflow.
The workflow also supports on-demand and scheduled metrics reporting, helping teams track qualification activity without manually reviewing spreadsheets.
How the Agent Works
- Receive incoming leads. The agent receives batches of raw prospect records from a lead-sourcing workflow, spreadsheet, CRM, webhook, or connected system.
- Assess ICP fit. Each lead is evaluated against the defined Ideal Customer Profile using available information such as job title, company, industry, location, tenure, and LinkedIn headline.
- Filter unsuitable prospects. Leads that do not meet the required fit criteria are removed before they enter research, personalisation, or outreach workflows.
- Remove duplicates and log qualified leads. The workflow compares incoming profiles against existing qualified records, filters repeated entries, and stores approved leads in a central tracking system.
- Route and report. Qualified leads can be passed to a research or outreach workflow, while scheduled and on-demand reports provide visibility into lead qualification activity.
Technical Workflow
1. Intake
A webhook (/lead-qualifier) receives a batch of leads (this is the exact payload the first workflow’s “Construct Payload” step sends). A “Split Out Batched Profiles” node breaks the batch’s body array into individual lead items so each can be processed independently.
2. AI-based ICP fit scoring
Each lead is run through an ICP Fit Classifier, an LLM chain (GPT-4.1-mini) prompted with the ICP description, the offer description, and the lead’s headline, job title, tenure, location, company, and industry. It is instructed to return strict JSON: {“fit”: “Yes”} or {“fit”: “No”}, with no explanation.
3. Output parsing + merge
A code node strips any markdown fencing or stray text from the LLM’s response and safely parses the JSON (falling back to null on parse failure). A Merge node then recombines that fit verdict (by position) with the original full lead record, so each lead carries both its data and its yes/no judgment.
4. Filtering
Only leads marked fit: “Yes” pass the Filter ICP Fit Yes node. Everyone else is dropped from the pipeline here.
5. Deduplication
The workflow pulls existing rows from the “Qualified Leads” Google Sheet, builds a set of already-logged profile URLs (normalised to lowercase), and filters out any lead that is already there, so re-runs do not create duplicate qualified entries.
6. Logging qualified leads
Surviving leads are appended to the “Qualified Leads” sheet with a full field set, including name, headline, company, industry, job title, tenure, and connection information, plus a new Date Qualified timestamp, distinguishing this from the raw capture in the previous workflow.
7. Batching + second handoff
Qualified leads are read back from the sheet, chunked into batches of 50, and POSTed to another downstream webhook (/research-agent), passing the baton to whatever does deeper research or personalisation before outreach.
8. On-demand metrics endpoint
A separate webhook (leads-metric-fetcher, nicknamed “postman time” in the workflow) accepts a startDate and endDate, counts how many leads were found in that window from the raw sheet, responds with the count, and logs it to a “Qualified Leads Metrics” sheet.
9. Scheduled daily metrics
Every night at 11 PM, a scheduled trigger pulls all raw leads, counts how many were found that calendar day, and saves the tally to the metrics sheet automatically, with no manual pull needed.
10. AI-summarised reporting endpoint
A third webhook accepts a date range, pulls the saved daily metrics for that window, aggregates totals and a per-day breakdown, and feeds it to an LLM (GPT-4.1) that is prompted to write a one to three sentence executive summary. That summary, along with the raw totals, is returned via webhook, essentially creating a self-serve reporting API.
Technology and Integrations
Built with: n8n, OpenAI GPT-4.1-mini and GPT-4.1, Google Sheets, webhooks, and connected lead-sourcing, research, CRM, or outreach systems.
Outcome
The LinkedIn Lead Qualifier Agent creates a more structured process for moving from raw prospect data to qualified leads.
Instead of manually reviewing every contact, teams can apply consistent ICP criteria, remove duplicate records, store approved prospects in a central system, and route suitable leads into research, personalisation, or outreach workflows.
Together with a lead-sourcing workflow, the agent supports a connected process from prospect discovery through to qualification and sales engagement.
Need a Similar AI Automation for Your Business?
The LinkedIn Lead Qualifier Agent is one example of how AI automation can support a more structured lead qualification process.
Divverse Labs designs and builds custom AI agents, automation workflows, internal tools, and connected systems for a wide range of business processes.
From sales, marketing, recruitment, customer support, and reporting to finance, operations, and internal team workflows, each solution is designed around the way your business works.
Request a Quote for a Custom AI Automation Solution to get started.