LEAD INTELLIGENCE
A click tells you which ad was pressed. A conversation tells you why.
LeadLogiX reads every conversation your AI agents hold, on your site, in your ads or on the phone. It verifies who is behind each one, scores the lead’s quality against comparable real leads and prices it by that quality, then puts it in your CRM with a brief on who to call first and what to say. When the person comes back, it knows them.
AI reads, code decides. Works with Avatar Studio, or with the agent you already run.
A form hands you five clean fields. A conversation hands you twenty free-form minutes, two people deciding, and a budget named three times. A transcript is not a lead, so every field LeadLogiX fills points back to the words it came from.
01
Your agent streams the conversation as it happens, with the actions it took and the campaign that brought the visitor. When it ends, one call hands the session over.
02
AI reads the transcript in any language, separates each speaker and pulls out role, stage, budget, financing and what they asked about. Each signal is checked against allowed values before it is used.
03
Every phone and email is checked, from syntax to carrier records. Five integrity checks are judged by rules and by AI; fail one and the lead is suppressed, unsold and unbilled.
04
Versioned code scores the lead against comparable real leads, then prices it, and records every input, so the same inputs always give the same score.
05
Into your CRM exactly once, to Slack, Teams, email or text, through the API, or to the AI assistant your team already uses.
Models are good at reading a messy conversation and bad at being consistent with money. So LeadLogiX lets AI do the reading and leaves every score and price to code.
THE BASELINE
For each new lead, LeadLogiX gathers comparable recent leads and measures the new one against them. Your own closed deals tell you what a lead is worth to you, and they never move what you are charged.
EVERY NUMBER ON A LEAD IS IN ONE OF THREE STATES
01
Comparable recent leads from the same line of business, narrowed to the same role: buyers against buyers, sellers against sellers. Below a minimum group size, nothing is computed.
02
Closer matches count for more, and older leads count for less as they age, so the baseline follows the market as it is now.
03
Outliers are pulled in to a cap rather than thrown away, so one odd lead can’t drag the baseline and no data quietly disappears.
04
As a percentile and a distance from the norm, with the sample size beside it: stronger than 84% of comparable leads.
Each line of business is calibrated on its own, with its own facts, roles, comparison groups and price ladder. Scores are never mixed across lines, so a profile with a home purchase and a life policy shows a score for each.
A lead your salesperson can act on in a minute, without listening to the recording. Prices never appear in the narrative, so the person calling can focus on the customer.
Name, phone and email with their verification status, address placed on the map, and the consent they gave.
Their role, how close they are to deciding, budget and financing, what they liked and what they rejected, and who else decides.
Interest in related services, each with the sentence where they said it and a check that the sentence is really there.
A brief, not a transcript: what they explored, what they asked, what was scheduled, the best channel, how soon to call and numbered follow-up steps.
Each lead keeps the campaign that brought the visitor beside what they said and did. Attribution data informs your campaigns and never changes a lead’s price.
In their own words, with the quote checked against the conversation.
The objection and what they rejected, each with the sentence where they said it.
Campaign tags and ad click ids, when your agent passes them in.
The partner or family member the buyer named.
The searches, pages and bookings your agent’s tools recorded, rebuilt as a journey, with the stage they reached and how sure that is.
The competitors the buyer named in the conversation.
People come back on a different day, a different device or a different channel. LeadLogiX weighs the evidence that links two conversations and decides one of three ways.
CUSTOMER PROFILES
Every conversation your agents hold, every ticket your help desk closes and every record your CRM already has, in one profile that runs from the first hello through support and renewal.
Budget, stage and score are followed across visits. A budget that grew and a score that climbed from 52 to 88 is a buying signal no single lead shows.
Stable facts, preferences and recent context, each traced to something the customer said. Each sentence of the AI summary is checked against its conversation, and one that can’t be traced is left out.
Help desk tickets and CRM records join the conversations with their source named, matched under the same evidence rules.
Questions, purchase, onboarding, support and renewal land on one record, so whoever calls at renewal knows what support heard.
WHAT TEAMS USE PROFILES FOR
Which campaigns produce buyers, counted across every visit.
Segments built from what buyers said: stage, budget band, what they want.
Real journeys of ready buyers become the playbook for the team.
What customers in a segment keep asking for, and what keeps stopping them.
On pay per lead, a profile is built from the leads you bought. The others show as locked entries with their date and price, one click from buying, with no separate charge for the profile.
WORKING THE LEADS
Tags decide who hears about a lead and who may work it. Roles decide what each person sees. The server enforces both, so nobody sees past their scope.
Applied by AI from a description you write, by a rule on a field your agent sends, by hand or by an automation, each recorded with its source.
An agent sees only the leads that carry their tags, any or all of them. The server applies the rule on every request.
Agents get the brief, the best channel and the follow-up steps. Procurement gets the score, price and expected value. Agents never see a price.
Opening hours for the workspace, each desk and each person, from LeadLogiX or your CRM, so an agent offers a callback when nobody is in.
Connect your closed deals and each lead shows an expected value from your own close rates and deal sizes, kept apart from the score and the price.
The funnel and where leads drop off, the AI agent that has gone quiet, and the return on what you spent per channel, for each line of business.
Each team, brand or region gets its own workspace with its own people and leads, billed through you with a breakdown per team on every invoice.
A weekly digest, aging leads, returning customers and a daily heartbeat, sent to the people on each tag.
Bought leads arrive in your CRM with their evidence, exactly once. Leads you haven’t bought are never pushed there.
Native connections you approve in their own sign-in window. Leads arrive once, with the LeadLogiX fields on the record; your staff and their hours come across; and closed deals feed an expected value beside each score.
A webhook, with its key kept in your secret vault, carrying the same fields and evidence.
Ready-made automations send hot leads, by tag, to a channel, an inbox or a phone, as they arrive or on a schedule.
A hosted MCP server lets an assistant read leads, customer profiles, routing and desk hours, with each permission approved by you. A local one lets your own agent submit conversations and recognize a caller. Turned on for your workspace on request.
Submit sessions, recognize a returning caller mid-conversation and read profiles. Documented in OpenAPI, with a sandbox playground, and turned on for your workspace on request.
Reports and alerts go out through your own mail service, under your domain. Alerts carry a lead id, the score and tags, never names or phone numbers.
Apps built for your team sign people in with LeadLogiX over OAuth 2.1 and OpenID Connect, and read their workspace and role, with no second user list to keep.
Fields your agents send become signals on the lead, and the calls their tools make (a search, a page opened, a tour booked) become graded evidence.
Ask for an automation and it arrives ready to switch on. You choose the leads, the schedule and the recipients, and every run is recorded.
Each lead’s price is set by code from its quality, its role and its market, and sealed the moment it is scored. Leads that fail a quality gate, and sessions too short to become leads, are never billed.
MARKET
Base value
What a lead is worth in its country, before quality.
ROLE
Role multiplier
A buyer, seller, renter or landlord is worth a different amount.
QUALITY
Conversion curve
The quality score raised to a power above one, so price falls faster than quality.
PRICE
Sealed price
Recorded on the lead with an integrity signature and never recomputed.
Below the pricing floor a lead is never priced or sold. The rates are calibrated to each market and kept unpublished, the way a credit bureau publishes its factors but not its weights.
If you sell something considered and already run a CRM, the conversation before the first call is where the sale is won or lost.
Buyers, sellers, renters and landlords, each measured against their own role, with budgets, financing, viewings and the services they’ll need next.
Policy owner, insured, beneficiary and payer told apart, with coverage and term captured. Health details are never kept, scored or released.
A new line is built and calibrated with you, with its own facts, roles, comparison groups and price ladder. Tell us how you qualify a customer.
From your side: someone who knows what a good lead looks like, access to your CRM, and where the agent should appear. Nothing in the right-hand column needs us afterwards.
Previews hold no personal details. Contact details and the playbook open only when a lead is bought.
Each application asks for each scope in plain words, contact details are a separate permission, and access ends the moment you disconnect it.
CRM credentials and the keys automations use are stored encrypted, handed to an automation only when it runs, and never shown again.
Old leads are cleared on a retention window you set. Leads you bought, and customers seen more than once, are kept the longest.
Administrators, procurement and agents each see only what their role and tags allow, and agents never see a price.
Staff sign in through your identity provider over OpenID Connect, on a domain you prove you own.
Alert emails carry a lead id, the score and its tags, never names or phone numbers.
On life insurance, the health details a caller mentions are never recorded, scored or released.
Who changed a lead, what an application did and which version scored it are all recorded, with names and times.