RB2B Intent Triage

Before you begin

Install Connecting RB2B's Profiles API to Claude first. This skill uses it to pull your identified visitors straight from RB2B, so you don't have to export a CSV each time you run it.

You'll also need a MoltSets account connected to Claude. MoltSets does the enrichment and the people search, and every lookup runs on your own MoltSets tokens. Sign up for a free MoltSets account.

Prefer working from a file? An RB2B CSV export still works. Upload or paste it into the chat and the skill uses it instead of the API.

What it does

The other four skills each solve one problem well. This one runs everything RB2B has identified and decides, visitor by visitor, which of those problems it's looking at.

A real week of RB2B data is a mixture: people who match your ICP, people who don't but work somewhere that does, company rows with nobody attached, and traffic not worth spending a credit on. Handled separately that's several passes and a merge. Handled here it's one pass that ends in a single ranked list.

When to use it

This is the one to run on a regular cadence — the weekly or fortnightly sweep of everything RB2B has identified since you last looked. It's the wrong tool for a narrow question; if you specifically need alternate emails, reach for that skill rather than processing everything.

Because it's an advanced skill with more moving parts, it rewards having your ICP written down properly before you start. Vague criteria produce a vague list.

How it works

1. Claude pulls your visitors from the RB2B API, or reads a full CSV export, and you define your ICP and any company fit criteria.
2. Every visitor is triaged into a bucket according to what it is.
3. ICP visitors are enriched directly with email and phone.
4. Company-only rows and non-ICP visitors at fitting companies are searched for the right contacts.
5. Everything is merged and ranked by intent.

How the ranking works

Priority is built from the engagement RB2B already records — all-time page views, recent page count, whether the profile is new, and any tags you've applied. The effect is that depth and recency of reading outrank job title.

That's the right instinct. A director who has read one page is a colder prospect than a manager who has read nine across three sessions this week, and a list sorted by seniority alone hides that entirely.

What you get back

One table: priority, name, company, the signal that earned the priority, where the contact came from, title, business email with risk grade, and mobile. The summary covers rows processed, ICP visitors enriched, companies searched, total contacts produced, fill rates for email and mobile, and the risk distribution.

The source column is worth keeping an eye on over time. If most of your contacts are coming from company searches rather than direct enrichment, your identified traffic is skewing away from your buyers — which is a targeting problem upstream, not something more enrichment will fix.

Where it fits in the week

The output is a prioritised list, which makes it a natural input to the process rather than the end of it. Pair it with a clear response process so the top of the list gets worked while the intent is still fresh, and with owner routing in HubSpot so each contact reaches the rep who owns that relationship.

Start small

Run it on a narrow slice first: a day or two of visitors from the API, or twenty or thirty rows of an export. That's enough to see the hit rate and the risk grades you're getting for your particular audience, and to decide which grades are worth keeping before you point it at a month of traffic.