From MQLs to AQLs.
A stat that's been rattling around my head for three weeks: 94% of B2B buyers finalize their vendor preferences before ever talking to a salesperson. Nearly 9 out of 10 decisions are already made by the time your SDR gets the lead.
That number explains a lot of what I've been seeing in B2B growth teams lately. MQLs are flooding in, sales teams are calling them "bad leads," the marketing-vs-sales blame war is back, and no one can figure out why demos are converting at half the rate they did in 2023.
The answer isn't that the leads got worse. It's that the qualification happened somewhere we can't see.
The MQL model quietly broke
The Marketing Qualified Lead concept was invented in an era where buyers started their research by filling out forms. They downloaded a whitepaper → they got nurtured → they raised their hand → sales called them. The whole funnel was downstream of "buyer submits their email."
That model still works if you're selling to buyers who behave the way buyers did in 2015. But the B2B buyer of 2026 does not behave like that.
Here's what actually happens now: a VP at a mid-market fintech has a problem. They open Claude or ChatGPT. They ask "what's the best demand generation platform for a fintech with 50 reps?" The AI synthesizes across 3-5 vendors, summarizes pros/cons, recommends a top pick. They ask two follow-up questions. They read three reviews on G2. They talk to a peer on LinkedIn. They visit two vendor websites briefly — not to download a whitepaper, but to confirm what they already read.
Then — and only then — they fill out a "request a demo" form. By this point their mind is 90% made up. They know pricing ranges, they have opinions on feature depth, and they've usually already eliminated two vendors without those vendors knowing they were ever in consideration.
The MQL model assumes the buyer starts their research with you. In 2026, they usually don't start their research with humans at all.
Enter the AQL
The B2B demand gen world started throwing around a new term this year: Agent-Qualified Lead. It's the logical evolution of the MQL, adapted for the new buyer journey. Instead of a lead becoming "qualified" by filling out enough forms and reading enough emails, they become qualified by having meaningful conversations with an AI agent — yours or someone else's.
Two flavors of AQL are emerging:
Externally qualified. The buyer had a conversation with ChatGPT, Claude, or Perplexity, received a recommendation that included your brand, and arrived at your site already pre-sold on the category and narrowed to 2-3 vendors. Marketing didn't "generate" this lead in the traditional sense. AI engines did. Your job was getting cited in the first place.
Internally qualified. The buyer interacted with an AI agent on your own site — a product advisor chatbot, a deep technical explainer that responds to questions, a pricing calculator that asks them 5 contextual questions and produces a tailored quote. The agent qualified them in a way a form never could, and handed the sales team a pre-briefed prospect who already knows the basics.
Both flavors produce fundamentally better leads than MQLs. Docket (one of the companies pushing the AQL model) is seeing their AI-pre-qualified leads convert to sales-accepted at 3-4x the rate of traditional MQLs.
What this breaks
If you accept that the funnel is shifting this direction — and the data is pretty hard to argue with at this point — it breaks a lot of existing infrastructure:
Attribution. The conversation that qualified the buyer happened inside ChatGPT. You have no cookie, no tracking pixel, no UTM, nothing. The lead appears in your CRM as "direct" or "branded search" and your multi-touch attribution model is missing the single most important touch in the journey. Most marketing dashboards in 2026 are under-counting their true impact by 30-50%.
Lead scoring. Traditional scoring weighs form fills, content downloads, email opens. An AQL skips all of that. They arrive ready to buy. If your scoring model requires five engagements before a lead hits "qualified," your best prospects are sitting in your CRM marked as cold leads.
Nurture sequences. The 8-email drip that was supposed to "warm up" your lead is insulting to a buyer who already read a full AI summary of your category and made a shortlist. Sending them "Top 5 Things to Know About [Your Category]" makes you look unsophisticated.
SDR playbooks. Discovery calls designed to teach the prospect about the space don't fit a buyer who's already done the research. The SDR role shifts from "qualify and educate" to "confirm fit and accelerate decision." Different skill, different script.
What this fixes
Here's the good news that rarely gets the headline: the AQL shift fixes some of the most expensive problems in B2B marketing.
Sales cycles are getting shorter. Industry data shows B2B buying cycles compressing from 11.3 months to 10.1 months — and still trending down. Pre-qualified buyers don't need 6 calls and 4 emails to get to pricing. They need one call, a proposal, and a contract.
Win rates are up. Teams using AI-assisted qualification are reporting 50% higher win rates. You're not talking to tire-kickers anymore. The AI filtered them out before they ever reached you.
CAC is dropping for the right accounts. If you show up in AI engine recommendations, your cost to acquire that customer is near zero. Someone else's content earned the citation, the AI did the qualification, the buyer arrived pre-sold. Your only "cost" was being mentionable enough to get recommended.
Three things to do this quarter
1. Instrument for invisible traffic
You can't fully track AI-driven leads, but you can get much better than you are today. Start measuring:
Branded search volume over time — this is your best proxy for AI mentions. If branded search is growing while direct traffic is also growing, AI engines are recommending you.
"How did you hear about us?" as a required field on your demo form. Awkward but it works. Expect 30-40% of answers to be "ChatGPT," "a friend mentioned you," or "just Googled around" — all AI-adjacent signals.
Direct traffic to high-intent pages (pricing, demo, case studies). Traditional SEO would see those as "lazy" traffic. In 2026 they're the gold standard — a user who types your URL directly is usually already 80% converted.
2. Earn citations where buyers actually research
Where does your buyer's research actually happen? The answer is rarely just "Google" anymore. Increasingly it's:
ChatGPT / Claude / Perplexity / Gemini conversations. G2 / Capterra / TrustRadius / Clutch (heavily scraped by LLMs). Niche subreddits in your vertical (AI training corpora weigh Reddit heavily). LinkedIn thought leadership posts from your ICP's peers. Industry newsletters and podcasts.
Your job is to exist, credibly, across all of these. Not necessarily ranking. Existing. Being cited. Being mentioned. Being the answer when a peer gets asked "who should we use for X?" — either a human peer, or an AI peer.
3. Build an AI agent on your own site
If your lead flow still looks like "form → SDR call → discovery call → demo call," you're fighting 2026 with 2019's playbook. The highest-leverage thing you can build right now is an on-site AI agent that does the top half of the discovery process.
Not a chatbot. Not a 2015 "can I help you?" widget. An actual conversational agent that can answer "does this work for a 200-person biotech?" in a real, specific way — pulling from your docs, case studies, pricing, and feature data. That agent qualifies the lead before they ever hit your SDR. Your SDR gets a transcript. The buyer gets answered in seconds instead of waiting for a Monday morning call.
Every serious B2B team I've talked to in the last six months is building this. The companies that don't will feel the gap by Q4.
The uncomfortable truth
MQLs aren't going to die overnight. You'll still generate them. Your board still wants to see them on the dashboard. But their predictive power is dropping fast, and the companies that notice this first — and redesign their funnels accordingly — are going to compound faster than the ones who don't.
The question isn't whether to replace MQLs with AQLs. It's how fast you can add AQLs to your funnel before your competitors do.
If you're trying to redesign your B2B funnel around AI-qualified leads — whether that's instrumenting for invisible AI-driven traffic, building an on-site agent, or just figuring out how to talk to your CEO about why MQLs aren't converting anymore — this is exactly the kind of work I do. Get in touch and let's talk through where you are.
