Explorium + OpenAI: score leads with confidence

You know the feeling: a new lead comes in, everyone has an opinion, and the “quick qualification” turns into a mini debate that steals your morning. Then someone copy-pastes a few LinkedIn details into a spreadsheet, guesses at fit, and you hope you didn’t just bury a great account under “maybe.”

This hits sales ops hardest, honestly. But GTM leads and growth-minded founders feel it too, because pipeline only moves when prioritization is consistent. With this ICP scoring automation, you get a clean 0–100 score plus a short summary for every company, so outreach decisions stop being personal preference.

Below you’ll see exactly what the workflow does, how the scoring logic works, and what you’ll need to run it in n8n without turning your process into a science project.

How This Automation Works

Here’s the complete workflow you’ll be setting up:

n8n Workflow Template: Explorium + OpenAI: score leads with confidence

Click to explore

flowchart LR

    subgraph sg0["On form submission Flow"]
        direction LR
        n0["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/form.svg' width='40' height='40' /></div><br/>On form submission"]
        n1@{ icon: "mdi:robot", form: "rounded", label: "AI Agent", pos: "b", h: 48 }
        n2@{ icon: "mdi:wrench", form: "rounded", label: "MCP Client", pos: "b", h: 48 }
        n3@{ icon: "mdi:brain", form: "rounded", label: "Anthropic Chat Model", pos: "b", h: 48 }
        n4["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/httprequest.dark.svg' width='40' height='40' /></div><br/>HTTP Request"]
        n1 --> n4
        n2 -.-> n1
        n0 --> n1
        n3 -.-> n1
    end

    %% Styling
    classDef trigger fill:#e8f5e9,stroke:#388e3c,stroke-width:2px
    classDef ai fill:#e3f2fd,stroke:#1976d2,stroke-width:2px
    classDef aiModel fill:#e8eaf6,stroke:#3f51b5,stroke-width:2px
    classDef decision fill:#fff8e1,stroke:#f9a825,stroke-width:2px
    classDef database fill:#fce4ec,stroke:#c2185b,stroke-width:2px
    classDef api fill:#fff3e0,stroke:#e65100,stroke-width:2px
    classDef code fill:#f3e5f5,stroke:#7b1fa2,stroke-width:2px
    classDef disabled stroke-dasharray: 5 5,opacity: 0.5
    class n0 trigger
    class n1 ai
    class n3 aiModel
    class n2 ai
    class n4 api
    classDef customIcon fill:none,stroke:none
    class n0,n4 customIcon

Why This Matters: Lead scoring breaks when it’s subjective

Manual ICP scoring sounds simple until you do it at volume. You start with “just check the website,” then someone looks for hiring signals, then someone else wants tech stack clues, and suddenly you have three different scoring styles across the same week. Even worse, the rationale disappears. Next month you can’t remember why Company A was “hot” and Company B was “not now,” so you re-litigate the decision and burn time again. The opportunity cost is real: your best reps spend hours qualifying instead of talking to buyers.

It adds up fast. Here’s where it usually breaks down.

People score the same lead differently, which means routing and outreach becomes inconsistent.
Firmographic and hiring context lives in too many tabs, so details get missed or misread.
Notes are unstructured (“seems good”) so you can’t audit decisions or train new SDRs.
When volume spikes, “quick checks” become shortcuts, and you end up chasing low-fit accounts.

What You’ll Build: Automated ICP scoring from enriched company data

This workflow turns a company name into a defensible ICP score in one run. It starts when someone submits a company through a simple form trigger in n8n. From there, Explorium enriches the account with firmographic details, hiring signals, and tech context (the stuff you normally hunt for across five sources). Then an AI agent applies a consistent three‑pillar scoring framework, totaling 100 points, and produces two outputs you can actually use: a numeric score and a short, plain-English summary of “why.” Finally, the workflow sends the structured result to a document conversion endpoint, so you can store it as JSON or a shareable doc-style brief.

In practice, the workflow begins with intake, pulls enrichment data from Explorium, and has the LLM score Strategic Fit, AI/Tech Readiness, and Engagement/Reachability. After that, the final score and rationale get packaged into a clean format that your team can copy into a CRM, a sheet, or a routing rule.

What You’re Building

What Gets Automated

What You’ll Achieve

A form submission triggers scoring the moment a new account is added.
Explorium enrichment pulls firmographics, hiring trends, and tech signals automatically.
An AI agent scores the company using the same 3-pillar framework every time.
The output is formatted into structured JSON (or doc-ready content) for sharing and storage.

Cut “should we pursue this?” discussions from about 15 minutes to a quick glance.
Every lead gets a 0–100 ICP score that’s easy to compare across your pipeline.
You get a short summary explaining the score, so reps know what angle to lead with.
Routing gets simpler because thresholds are clear (good fit vs. medium vs. poor).
Scoring becomes auditable, which makes coaching and iteration much easier.

Expected Results

Say your team qualifies 20 inbound companies a week. Manually, even a “fast” check is about 15 minutes per company once you factor in tab-switching and writing notes, so that’s roughly 5 hours weekly. With this workflow, submitting a company takes about a minute, then enrichment + scoring runs in the background and returns a score plus summary. You still review the result, but it’s a quick decision, not a research task.

Before You Start

n8n instance (try n8n Cloud free)
Self-hosting option if you prefer (Hostinger works well)
Explorium for firmographic, hiring, and tech enrichment.
OpenAI to produce the score and summary output.
Explorium MCP credentials (get them from your Explorium account).

Skill level: Intermediate. You will connect accounts, paste credentials, and adjust a scoring prompt without writing code.

Want someone to build this for you? Talk to an automation expert (free 15-minute consultation).

Step by Step

Company intake through a form trigger. Someone submits a company name (and any optional context you add, like target segment or territory). That single input becomes the “source of truth” for the scoring run.

Enrichment via Explorium. The workflow calls Explorium’s MCP tool to fetch the data humans normally look up: company size, industry, hiring signals, and tech footprint. That enrichment is what makes the final score feel grounded, not guessy.

LLM-based scoring using a consistent rubric. An AI agent evaluates the company across three pillars (Strategic Fit, AI/Tech Readiness, and Engagement/Reachability) and returns a 0–100 score, a verdict label, and pillar breakdown. The LLM is the “judge,” but the rubric is your guardrail.

Structured output for docs and downstream tools. The workflow sends the formatted result to a document conversion endpoint so it can land as JSON or a doc-like summary you can share internally. From there, it’s easy to push into Google Sheets, a CRM, or a routing system.

You can easily modify the scoring thresholds and pillar weights to match your ICP. See the full implementation guide below for customization options.

Step-by-Step Implementation Guide

Step 1: Configure the Form Trigger

Set up the intake form that starts the ICP scoring flow.

Add the Form Intake Trigger node as your trigger.
Set Form Title to Company ICP scoring.
Add a form field with Field Label set to Company Name, Placeholder set to Apple, and mark it as required.
Set Form Description to =This automation takes company's Linkedin Profile URL and Airtop Profile (authenticated for Linkedin) and returns the company's ICP score.

You can test the form immediately from the Form Intake Trigger node by opening the production URL and submitting a sample company name.

Step 2: Set Up the AI Scoring Agent

Configure the AI prompt that generates the ICP scoring report.

Add the LLM Scoring Agent node and connect it to Form Intake Trigger.
Set Prompt Type to define.
Set Text to the full prompt provided, including the expression {{ $json['Company Name'] }} in the company name line.
Set the System Message in Options to the provided business analyst instructions to enforce Markdown output and scoring rules.

⚠️ Common Pitfall: Ensure the output is Markdown only—extra text or JSON will break the document conversion in Doc Conversion Request.

Step 3: Connect the AI Language Model and Tool

Attach the LLM and external tool used by the agent. These are connected as AI sub-nodes.

Connect Anthropic Chat Engine to LLM Scoring Agent via the AI language model connection.
Set the model in Anthropic Chat Engine to claude-3-7-sonnet-20250219.
Credential Required: Connect your anthropicApi credentials in Anthropic Chat Engine.
Connect Explorium MCP Tool to LLM Scoring Agent via the AI tool connection.
Set SSE Endpoint to mcp.explorium.ai/sse and Authentication to headerAuth in Explorium MCP Tool.
Credential Required: Connect your httpHeaderAuth credentials in Explorium MCP Tool.

The AI tool and language model are sub-nodes for LLM Scoring Agent. Make sure credentials are added on these nodes (not on the agent itself).

Step 4: Configure the Document Conversion Output

Convert the Markdown report into a Google Doc via HTTP request.

Add the Doc Conversion Request node and connect it after LLM Scoring Agent.
Set URL to https://md2doc.n8n.aemalsayer.com and Method to POST.
Enable Send Body.
Set Authentication to predefinedCredentialType and Credential Type to googleDocsOAuth2Api.
Add body parameters: output with value ={{ $json.output }} and fileName with value ={{ $('Form Intake Trigger').item.json['Company Name'] }} ICP Report.
Credential Required: Connect your googleDocsOAuth2Api credentials in Doc Conversion Request.

Step 5: Test and Activate Your Workflow

Validate the full flow from form submission to document creation, then enable it for production.

Click Execute Workflow and submit a test entry through Form Intake Trigger.
Confirm that LLM Scoring Agent returns a Markdown report with the ICP score sections and verdict.
Verify Doc Conversion Request completes successfully and a Google Doc is created with the company name in the filename.
When satisfied, toggle the workflow to Active to accept production form submissions.

🔒

Unlock Full Step-by-Step Guide

Get the complete implementation guide + downloadable template

Troubleshooting Tips

Explorium credentials can expire or need specific permissions. If things break, check your Explorium MCP access settings in your Explorium account first.
If you’re using Wait nodes or external rendering, processing times vary. Bump up the wait duration if downstream nodes fail on empty responses.
Default prompts in AI nodes are generic. Add your brand voice early or you’ll be editing outputs forever.

Quick Answers

What’s the setup time for this ICP scoring automation?

About 30 minutes if your Explorium and OpenAI credentials are ready.

Is coding required for this ICP scoring automation?

No. You’ll connect Explorium and your LLM provider, then adjust the scoring prompt and output fields.

Is n8n free to use for this ICP scoring automation workflow?

Yes. n8n has a free self-hosted option and a free trial on n8n Cloud. Cloud plans start at $20/month for higher volume. You’ll also need to factor in OpenAI usage costs (often a few cents per scored company, depending on your prompt and model).

Where can I host n8n to run this automation?

Two options: n8n Cloud (managed, easiest setup) or self-hosting on a VPS. For self-hosting, Hostinger VPS is affordable and handles n8n well. Self-hosting gives you unlimited executions but requires basic server management.

Can I modify this ICP scoring automation workflow for different use cases?

Yes, and you should. You can change the pillar weights and verdict thresholds inside the AI Agent prompt, then adjust the formatted output sent by the Doc Conversion Request. Common tweaks include adding “target geography” rules, enforcing a minimum company size, or changing the rubric to match product lines (SMB vs. enterprise). If you prefer a different model, swap the Anthropic Chat Engine for an OpenAI Chat Model node.

Why is my Explorium connection failing in this workflow?

Usually it’s expired credentials or missing MCP permissions in your Explorium account. Regenerate your token, update it in n8n, and retry the enrichment call. If you’re scoring lots of companies in a short window, rate limits can also look like “random” failures, so spacing requests out helps.

What volume can this ICP scoring automation workflow process?

If you self-host n8n, there’s no execution limit (it mainly depends on your server and API limits). On n8n Cloud, your monthly execution cap depends on plan, and the workflow uses one execution per company scored. In most setups, scoring a company takes long enough that you’ll want batching for big lists, but handling a few hundred a week is realistic.

Is this ICP scoring automation better than using Zapier or Make?

Often, yes. n8n is more comfortable when you want real logic (branching, retries, structured JSON), and it’s easier to self-host when executions grow. Zapier or Make can be fine for a lightweight “log the lead” flow, but LLM scoring plus enrichment tends to get messy fast in two-step automations. The bigger win is control: you can version the rubric, change weights, and keep the output consistent across teams. Talk to an automation expert if you want help deciding.

Once this is running, lead scoring stops being a recurring meeting topic. The workflow handles the repetitive evaluation, and you get back to the part that matters: talking to the right accounts.

Explorium + OpenAI: score leads with confidence

How This Automation Works

n8n Workflow Template: Explorium + OpenAI: score leads with confidence

Why This Matters: Lead scoring breaks when it’s subjective