How to automate your entire sales pipeline without writing a single line of code

If you've ever watched a sales rep spend their morning copying lead data from a form into a CRM, enriching it manually on LinkedIn, scoring it in a spreadsheet and then pasting a Slack message to the team — you've watched about three hours of work that should take three seconds. This guide is about getting those three hours back.

The workflow we're building has four steps. It's not a toy example — it's the exact agent we run internally at Flowpath for every inbound demo request, and it's been processing leads without manual intervention for over six months.

Step 1 — The trigger: Typeform form submission

Every automation needs a starting point. Ours is a Typeform webhook — every time someone submits your demo request form, Flowpath receives the payload instantly. The data you'll capture from the form should include: first name, last name, work email, company name, job title, team size and use case. These fields feed into the enrichment step, so the more complete your form, the better your enrichment results will be.

In Flowpath, you add a Trigger node to the canvas, select Typeform from the integration library, authenticate your account, and choose the specific form. Setup takes under two minutes. From this point on, every submission fires the agent automatically — there's no polling, no delay and no manual check required.

Step 2 — Enrichment: Clearbit

A form submission gives you what the prospect chose to share. Clearbit gives you everything else. With a single API call using the prospect's work email, Clearbit returns: company size, industry, annual revenue estimate, funding stage, tech stack, Alexa rank, employee count by department and the LinkedIn profile of the contact.

In Flowpath, you add an Action node, select Clearbit, choose the Enrich Person endpoint and map the email field from the previous step. The enriched data is now available to every subsequent step as named variables — {{company_size}}, {{industry}}, {{funding_stage}} and so on.

One important note: Clearbit enrichment fails on roughly 8% of contacts — typically personal email addresses, very small companies or newly founded startups. Add a conditional branch after the enrichment step that checks whether the company_name field is empty. If it is, route those contacts directly to a manual review queue in HubSpot rather than through the scoring step.

Step 3 — Scoring: GPT-4o

This is where the intelligence lives. You add an AI step node and write a prompt that evaluates ICP fit using the enriched data. Here is the exact prompt we use:

You are a lead scoring assistant. Score the following lead on a scale of 0 to 100 based on fit with our ideal customer profile (ICP). Our ICP is: B2B SaaS or tech companies with 50–500 employees, Series A or B funded, with operations, sales or marketing teams of at least 5 people. Score higher for VP-level or above job titles. Score lower for agencies, non-profits, students and individual freelancers. Return only a single integer between 0 and 100. No explanation, no punctuation, nothing else. Lead data: Company: {{company_name}}. Industry: {{industry}}. Employees: {{employee_count}}. Funding: {{funding_stage}}. Job title: {{job_title}}.

The "return only a single integer" instruction is critical. Without it, GPT-4o will sometimes return "Score: 74" or "74/100" or a sentence — all of which will break the downstream branch condition. Be explicit about output format in every AI step prompt you write.

Step 4 — Routing: Branch logic

With the score in hand, the workflow splits into two paths. Score ≥ 70 triggers the qualified path: a Slack message to #sales-alerts with the lead's name, company, score and a link to the HubSpot record, followed by a HubSpot contact creation step that assigns the lead to the next SDR in round-robin rotation and adds them to an active sequence. Score < 70 triggers the nurture path: a HubSpot contact creation that adds the lead to a 30-day email nurture sequence tagged for re-scoring at day 30.

The result

The workflow processes a lead in an average of 1.4 seconds from form submission to Slack notification. Your SDRs wake up each morning to a #sales-alerts channel full of pre-scored, pre-enriched leads with HubSpot records already created. No copy-pasting, no manual enrichment, no spreadsheet scoring.

At Flowpath, this workflow alone saves our sales team an estimated 11 hours per week. At your company, the number will depend on your lead volume — but the math is simple: if you're getting 50 inbound leads a week and each one takes 8 minutes to process manually, that's 6.5 hours. This workflow gets that to zero.