Your analytics calls it “direct traffic.” It was ChatGPT.
AI engines are already recommending brands like yours — and when the buyer clicks through, most tools file the lead under “direct” and move on. Leads catches the handoff: every lead in one inbox, stamped with the engine that actually sent it. Proven, never guessed.
Stamps the engine on every lead it can prove — and says so when it can't
Leads is the cash-register end of AI search: one inbox for every lead you get, each stamped with the AI engine that sent it — ChatGPT, Perplexity, Gemini, and six more — proven from a real signal the lead itself carries, never a model's guess. When there's no signal, it says “Direct / unknown.” The rest of the platform earns you the recommendation; Leads shows you the customers it produced.
The last mile of AI search is a blind spot.
You can win the citation, get recommended, and still have no idea it happened — because the moment a buyer moves from an AI answer to your site, most tooling loses the thread.
The junk drawer called “direct”
Buyers arriving from AI answers land in analytics as “direct” or generic referral traffic. The biggest channel shift in a decade, filed under miscellaneous.
Recommendations don't leave receipts
An engine says your name, the buyer opens a new tab, and the trail goes cold — unless something on your own site is built to catch the handoff as it happens.
Lead vendors sell strangers
Bought lists mean contacting people who never asked to hear from you — a compliance risk and the exact opposite of a buyer an AI engine just warmed up for you.
One inbox. Nine engines. The whole journey.
Everything a lead does — from the AI answer that sent them to the email you send back — in one place.
One inbox for everything
A paste-in site tag, a webhook your CRM or form builder pushes to, leads your AI Chatbot captures, inbound Gmail — every channel lands in one inbox, normalized and deduplicated.
The engine stamp
Nine AI engines recognized — ChatGPT, Perplexity, Google AI Overview, Gemini, Copilot, Claude, Meta AI, Grok, DeepSeek. The badge lands only when a real signal proves it.
The full journey
Expand any lead: first touch, the pages they read, the form they filled. Set statuses, archive, export — and the consent trail stays with the lead.
The attribution board
Leads per engine, with trends — which AI engines actually send you customers, in real counts. Not modeled percentages, not “AI-influenced” estimates.
Your next ten prospects
Ten a day, every day for B2B brands: fresh decision-makers matched to your ideal customer profile, with verified business contacts. Email each from your own Gmail, mark them contacted.
ChatGPT Ads, beside the leads
Connect your own ChatGPT ads account and see real spend, clicks, and per-campaign performance next to the leads coming in. Your key is stored encrypted, never shown again.
Attribution tools model it. Leads proves it.
“AI-influenced revenue” usually means a probability model you can't inspect assigned credit you can't verify. Leads goes the other way: every badge on every lead is backed by a real signal — or it isn't there.
How attribution usually works
- Assign credit with a black-box model, then report it as fact.
- Publish “AI-influenced” percentages nobody can verify.
- Lump AI referrals into “direct” and call the channel unmeasurable.
- Pad the pipeline with names scraped or bought from somewhere else.
How Leads decides
- A real signal, or nothing. The engine badge is earned from evidence the lead itself carries — the exact signals we check stay ours.
- “Direct / unknown,” said plainly. A lead with no AI signal is never force-fitted to an engine. The honest label is what makes the stamped ones count.
- Samples say SAMPLE. An empty inbox shows labeled example rows and your real zero — never fake momentum.
- No invented outcomes. We don't track your closed deals, so we don't fabricate conversion rates. What you see is what was measured.
Five leads land. Watch the stamps.
This is the inbox catching a week's handoffs — each lead arriving, checked for a real AI signal, and badged only when the engine is proven.
What the inbox hands you.
- The handoff, caught. The moment an AI answer turns into a visitor with a name, the engine stamp is on the lead — not buried in a traffic report.
- Proof, per lead. Expand any row: the engine, the first touch, the pages they read, the form they filled. Evidence you can show whoever asks.
- A channel report you can defend. Real counts per engine over time — the number of customers each AI engine actually sent, with the honest “unknown” beside it.
- Included with every plan. Inbox, capture tag, journey, and attribution on every plan — the full board and prospecting at AI Radar and up.
Modeled attribution vs. proven attribution.
| What you get | Not tracking | Analytics & attribution suites | Leads |
|---|---|---|---|
| The core question | — | “Where did traffic come from?” | “Which AI engine sent this customer — and can you prove it?” |
| AI referrals | Invisible | Lumped into “direct” and referral buckets | Stamped per lead across 9 engines, only when proven |
| When there's no signal | — | A model assigns credit anyway | “Direct / unknown,” said plainly |
| The lead itself | — | An anonymous session | A name, a journey, a status, and one-click outreach from your own email |
| Outbound | — | — | Ten fresh ICP-matched B2B prospects every day, with verified business contacts — never bought consumer lists |
| Ad spend | — | A separate dashboard | Your ChatGPT ad spend beside the leads it produced |
| Cost | An invisible pipeline | $$$ attribution suite, still guessing | Included with every plan; full board + prospecting at Radar and up |
Catch. Stamp. Work. Prove.
Catch
Paste one tag on your site — or point your CRM, forms, chatbot, or Gmail at your inbox. Every lead lands in one place, normalized and deduplicated.
Stamp
Each arriving lead is checked for real evidence of the AI answer that sent it. The engine badge lands only when it's proven; no signal reads “Direct / unknown.”
Work
Open the journey — first touch, pages read, the form they filled. Set statuses, export, and email any lead or prospect from your own connected address.
Prove
The attribution board tallies real counts per engine over time — which AI engines actually send customers, in numbers you can defend.
The vocabulary of the handoff.
The terms your Leads inbox uses — defined plainly.
- AI-source attribution
- The engine badge on a lead — earned from a real signal the lead itself carries, never assigned by a probability model.
- First touch
- The first thing known about a lead's visit: where they came from and the first page they landed on. It's where the handoff from an AI answer gets caught.
- Direct / unknown
- The honest verdict when a lead carries no AI signal. Leads says it plainly instead of guessing an engine.
- Capture tag
- One paste-in snippet that catches arrivals from AI engines and turns a site's existing forms into inbox entries — no rebuild required.
- Lead journey
- The trail from first touch to form: the pages a lead read and the AI referrer if there was one — expandable on every lead.
- Hot lead
- A new, unworked lead in the inbox. Counted for real on the dashboard — never padded.
Questions about Leads.
What is Leads?
Leads is AI Syndicate's lead inbox and AI-attribution layer — the cash-register end of AI search. Every lead you get lands in one inbox and is stamped with the AI engine that sent it — ChatGPT, Perplexity, Gemini, and six more — when a real signal proves it. It also hands B2B brands ten fresh outbound prospects every day and puts your ChatGPT ad spend beside the leads it produced.
How does Leads know which AI engine sent a lead?
From evidence the lead itself carries — a real trace of the AI answer that sent them, caught at the moment of the handoff. When the evidence is there, the badge lands; when it isn't, the lead reads “Direct / unknown.” There's no probability model assigning credit, and the exact signals we check stay ours.
Which AI engines does it recognize?
Nine: ChatGPT, Perplexity, Google AI Overview, Gemini, Microsoft Copilot, Claude, Meta AI, Grok, and DeepSeek. Every other lead is honestly labeled “Direct / unknown” rather than force-fitted to an engine.
How do leads get into the inbox?
Four ways, all feeding one place: a paste-in capture tag that catches arrivals and turns your existing forms into inbox entries; a webhook your CRM, form builder, or automation tool can push to; leads captured by your AI Chatbot; and inbound email synced from a connected Gmail. Everything is normalized and deduplicated on the way in.
What happens when a lead has no AI signal?
It's labeled “Direct / unknown” — plainly. That honesty is what makes the stamped leads meaningful: when the board says four leads came from ChatGPT, that's four proven handoffs, not a modeled share.
Do you sell leads or buy contact lists?
Never — we don't buy or sell consumer contact lists, full stop. Every consumer lead in your inbox came to you on its own. Outbound prospecting is different and clearly separated: it surfaces business decision-makers matched to your ideal customer profile, with verified business contact information.
What does prospecting do?
It hands you ten fresh prospects every day: decision-makers matched to your ideal customer profile — role, industry, company size — with verified business contact details. You can email each one from your own connected Gmail and mark them contacted, so outbound and inbound live in the same place. It's built for B2B; consumer brands see it surface partners and referral sources instead, and it says so.
Can I see my ChatGPT ad spend next to my leads?
Yes. Connect your own ChatGPT ads account and Leads shows real spend, clicks, and per-campaign performance beside the leads coming in. Your key is stored encrypted, used only to read performance, and never displayed again.
Which plans include Leads?
Every plan includes the inbox, the capture tag, the journey view, and AI-source attribution — entry plans see the top three AI sources on the board. The full attribution board and outbound prospecting unlock at AI Radar and up.
How is this different from my analytics tool?
Analytics counts anonymous sessions, and it files most AI referrals under “direct.” Leads works at the level of a person: a name, the engine that sent them, the pages they read, and the outreach that follows — with attribution that's proven per lead rather than modeled per channel.
Find out which AI engine sends your next customer.
Somewhere tonight, an engine is recommending someone in your market. Put the inbox in place — and see the stamp when the buyer walks in.