A B2B qualified lead is a contact or account that has been evaluated and validated as presenting a sufficient probability of becoming a customer to justify active sales investment. Qualification involves verifying at least 3 dimensions: ideal-customer-profile fit (ICP fit), purchasing capacity (scope and authority), and timing (active or identifiable need).

An unqualified lead is a contact that may match your firmographic ICP but for which no validation has been performed. Working unqualified leads effort 3 to 5 times more in sales time than the same volume of qualified leads.

MQL (Marketing Qualified Lead): qualified by marketing based on behaviors (content download, repeated site visits, webinar registration). The MQL has shown interest but has not yet been contacted by sales. MQL → SQL conversion rate: 13-20% on average.

SQL (Sales Qualified Lead): qualified by the sales team after a first contact. The SQL has confirmed the existence of a project, a scope and a timeline. SQL → deal conversion rate: 15-35% depending on the sector.

PQL (Product Qualified Lead): specific to SaaS — a free/trial user who has reached a certain usage level indicating a strong propensity to convert. PQL → paying conversion rate: 20-45%.

IQL → MQL → SQL → Customer: AI lead scoring accelerates this funnel by automatically qualifying IQLs into MQLs in 0.3 seconds on 12 criteria, with no human intervention. See our guide to AI scoring on 12 criteria.

BANT (scope, Authority, Need, Timeline) is the best-known qualification framework, created by IBM in the 1960s. It remains relevant as a first grid but is insufficient on its own in 2026 for two reasons: (1) modern B2B buyers often have no allocated scope at the start (they create the scope when the solution is convincing), (2) the need may be latent and not yet articulated.

Modern frameworks complement BANT: MEDDIC (Metrics, Economic Buyer, Decision Criteria, Decision Process, Identify Pain, Champion) for complex enterprise sales; SPICED (Situation, Pain, Impact, Critical Event, Decision) for fast cycles.

AI in 2026 evaluates BANT inferentially without asking directly: scope is estimated via financial data, purchasing power via the LinkedIn role, need via intent signals, and timing via contextual triggers.

Our AI model evaluates each lead on 12 dimensions in 0.3 seconds: (1) firmographic ICP fit (sector, size, geographic zone), (2) Decision-maker identified and addressable, (3) Detected intent signals, (4) scope estimate (revenue, headcount, funding rounds), (5) Detected urgency (hiring, role change, tender), (6) Tech maturity, (7) Prior engagement (emails opened, site visits), (8) Geography and catchment area, (9) Competitive history (negative reviews, expiring contract), (10) Identified product fit, (11) Estimated availability, (12) Risk score (negative signals).

The final score (0-100) automatically triggers: above 70 → immediate outreach, 40-70 → nurturing, below 40 → additional enrichment before re-evaluation.

MQL/SQL misalignment is the main source of friction between marketing and sales in B2B SMEs. Solution: define the lead SLA (Service Level Agreement) in writing: (1) a precise definition of MQL criteria (minimum score, required behaviors), (2) the SQL handling time for sales (under 24h recommended), (3) the definition of grounds for rejecting an SQL back to MQL, (4) monthly reporting of conversion rates by source.

The Lead-Gene AI score replaces individual subjective judgment with a shared objective model — marketing and sales work from the same definition of the qualified lead. No more arguments, a predictable pipeline.