Qualifying and routing inbound leads before they go cold

Workflow 1: enrich and score

The first workflow starts when a form is submitted. With Typeform, use the Typeform trigger. With your own site form, post the submission to a generic Webhook trigger.

Enrich. The Apollo block enriches the person and the company: size, industry, region and role. If you already pay for another enrichment provider, call it with the API block, with the key stored as a workspace secret. Enrichment fills gaps; it does not overwrite what the person told you on the form.

Score against rules. Your ideal customer profile (ICP) is written in Rules as plain-language checks, for example:

• Company has between 100 and 5,000 employees.
• Industry is logistics, retail, financial services or software.
• Region is one we sell in.
• Role is in operations, support, finance or engineering, at manager level or above.
• The form answer describes a use case we support.

Each rule returns pass or fail with a reason. The score is a weighted sum, and the tier (A, B or C) follows from it. The workflow writes everything to a leads table in Tables: the answers, the enrichment, each rule result, the score and the tier.

Before any of that is stored, a Guardrails block can detect and mask personal data you do not need, such as a phone number typed into a free-text field.

Workflow 2: route in HubSpot

The second workflow reads the new row and decides the owner. Tier and region pick a queue; a round robin inside the queue picks the rep, skipping anyone marked out of office in a small reps table.

The HubSpot block then creates or updates the contact and company, sets the owner, and adds a note with the score and the reasons. Tier A leads also post to the rep in Slack. Tier C leads go to a nurture list rather than a rep, and nobody is surprised by a lead they did not want.

Keeping routing in its own workflow means you can change the queues without touching the scoring, and each runs from a deployed version that someone reviewed.

The Relay: a first reply worth sending

A Relay, an AI coworker with its own instructions and access, watches the leads table for new tier A and B rows, checking as often as every minute. For each one it reads the form answer and enrichment, searches the messaging guide in the Knowledge Base, and drafts a first email from the owning rep’s point of view.

Relays never send email without an approval, under any autonomy setting. The rep sees the exact message, edits it if needed and approves; the approval covers that text only and expires after 24 hours by default. Because first emails go to prospects, the Relay’s recipient policy has to allow them: an allowlist of domains is too narrow here, so most teams set it to anyone and rely on the approval step. Start with internal only for a week, so drafts go to reps, before switching.

The Relay does not phone leads. If you want inbound calls answered and warm leads passed to sales, that is Agent Inbound & Outbound, a separate Valnora & Elric product available by invitation.

Watching the funnel

A dashboard in Dashboards on the leads table shows leads by tier, time from form to owner, time from form to first approved email, and how many tier A leads booked a meeting. The figures are yours to measure; we would not promise any in advance.

Two things are worth watching from the first week:

1. Rules that fail too often. If most leads fail on company size, either your form attracts the wrong people or the rule is wrong. Both are worth knowing.
2. Drafts that reps rewrite. If reps rewrite most first emails, improve the messaging guide rather than the prompt.

Logs record every run block by block, so when a rep asks why a lead went to someone else, the answer is one click away.

Rolling it out

Run the workflows in parallel with the current process for two weeks, writing scores and owners to the table without touching HubSpot. Compare the routing with what the coordinator would have done. When they agree on most leads, switch HubSpot writes on and introduce the Relay.