Form fills are triaged by ops on Mondays, assigned by Wednesday, called on Friday. The buyer booked a competitor demo on Tuesday.
Point-based scoring models built two years ago still give demo buttons to free-mail-address students while rejecting VPs with no title field.
Form fields are shallow. Enrichment happens in a nightly batch. Reps pick up leads blind and call without knowing the company size.
Half of your inbound form fills are bots, contractors, or students. Reps lose hours separating real from fake.
An enterprise lead gets routed to an SMB rep because the round-robin pointer happened to land there. Good luck recovering the deal.
Sales says the leads are bad, marketing says the leads are great, nobody updates the scoring model and the argument resets every QBR.
Instead of points, a reasoning agent compares each lead to your ICP, your offer, and your open pipeline. Decisions come with rationale.
Apollo → Clearbit → Ocean → your warehouse. Most leads enrich in under 15 seconds, with credits only spent on misses.
Free-mail heuristics, honeypot checks, IP reputation, behavioral patterns — low-signal leads auto-discarded or low-prioritized.
Enterprise lands with the AE, mid-market with the SMB rep, self-serve triggers a product-led email sequence. All automatic.
Qualified, enriched, routed, and optionally first-touch drafted before the buyer closes the thank-you page.
Every lead’s trajectory — score, route, touch, outcome — feeds back into the qualification model automatically.
Point HubSpot forms, Marketo, native React forms, or any webhook source at Black Magic AI. OAuth into your CRM. Schema maps itself.
Tell the agent what "Enterprise," "Mid-Market," "SMB," and "Disqualified" look like — in sentences. Save the segment library. Change any time by chat.
Bot checks run, enrichment waterfall runs, the qualification agent scores and tier-assigns, all under a minute. The lead enters the CRM already decorated.
Enterprise → named AE + Slack ping. Mid-market → SMB rep round-robin. SMB self-serve → drip sequence. DQ → archive with reason.
For top-tier leads the agent drafts a personalized opener that references both the form context and enriched account context. Rep sends in one click.
Every lead’s outcome — replied, demo’d, closed, ignored — feeds back. The qualification model nudges itself; your ops team stops maintaining point tables.
Every lead an AE sees is enriched, tiered, and ranked. No more "who even is this person" moments.
The full path — campaign, form, enrichment, tier, rep, outcome — lives in one record. Defensible ROI numbers, on demand.
The qualification logic lives in chat and in a live agent, not in a 400-row scoring matrix nobody trusts.
LLM reasoning over firmographic, technographic, intent, and product-fit signals.
Priority-ordered provider calls with credit-aware fallbacks.
Match inbound against open deals and existing contact records.
Freemail, honeypot, IP reputation, and behavioral checks baked in.
Enterprise, mid-market, SMB, PLG — all routed with override logic.
Closed-loop tuning from reply and conversion data.
Yes, practically speaking. You can keep HubSpot scoring as a safety net, but most teams find the agent output more accurate and easier to tune. We write the tier back to HubSpot as a first-class field.
The waterfall continues to the next provider; if all miss, the lead is still qualified on form-field signal alone and flagged for human review. No lead is ever dropped silently.
Every scored lead has a rationale log — which fields were used, which providers returned data, what the agent concluded. Click any lead to replay the decision.
Median 40 seconds from form submit to routed + enriched. P95 under 90 seconds. We publish latency metrics in-app.
