Why Traditional Segmentation Fails in AI SDR Campaigns

Many teams still rely on job titles or geography when building AI SDR campaigns, but this approach consistently underperforms. According to SaaStr's real-world deployment, segmenting by demographics like location or role fails because it ignores behavioral signals that indicate actual buying intent. Instead, their data shows that campaigns grouped by context and intent — such as website visitors, event attendees, or lapsed contacts — drive 4–7x higher conversion than broad or cold targeting. This isn’t just theoretical; SaaStr’s hyper-segmented framework, which caps campaigns at 100–500 contacts per segment, generated $4.8 million in additional pipeline over eight months by focusing exclusively on warm, engaged audiences.

When AI SDRs are pointed at cold lists or broad firmographic segments, they lack the contextual triggers needed to personalize outreach effectively. Jason Lemkin explicitly advises against the “old-school way” of segmentation, noting that geography and title play no meaningful role in his outbound funnel. Instead, he prioritizes segments rooted in demonstrated interest: abandoned trials, marketing leads, former customers, and warm referrals — all of which reflect prior engagement or intent. These segments allow AI systems to tailor messaging with relevance, increasing the likelihood of response and qualification. In contrast, demographic segmentation treats all contacts as interchangeable, wasting the AI’s ability to adapt and learn from behavioral patterns.

This misalignment creates a critical bottleneck in the response process — one that Worqd helps clients uncover during the “find the bottleneck” phase of their growth workflow. By shifting focus from who a lead is to what they’ve done, companies unlock far greater efficiency in their AI SDR efforts. The result isn’t just more conversations — it’s more qualified conversations, faster. Teams that adopt intent-based segmentation see better lead quality, higher booking rates, and lower cost per qualified opportunity — outcomes that align directly with Worqd’s promise of turning interest into booked calls in under 60 seconds, 24/7.

The SaaStr Hyper-Segmentation Framework: A Proven Market Segmentation Example

If you want one example of market segmentation that actually produced results, look at what SaaStr did with its own go-to-market. Over 8 months, its AI agents helped generate $4.8 million in additional pipeline — and the segmentation strategy behind it was nothing like the textbook approach.

Jason Lemkin, SaaStr's founder, rejects the old-school way of slicing a market. "Don't segment the old-school way — by geography, title, or role. None of those exist in my outbound funnel. Instead, segment by context and intent," he writes, based on eight months of hands-on deployment across 20 AI agents.

His segments look like this:

• Website visitors who showed interest but never filled out a form
• Inbound leads and abandoned trials
• Event leads needing timely follow-up
• Former and current customers, plus lapsed contacts
• Low-scoring leads, alumni, and warm referrals

The discipline matters as much as the segments themselves. Each campaign is capped at 100–500 contacts, with customized training per segment. "I see people running one campaign for 10,000 leads. That's insane," Lemkin says. Small, focused campaigns let the messaging match the context — which is exactly what AI SDR workflows need to perform.

The results back it up. SaaStr's deployment data shows deal volume more than doubled and win rates nearly doubled, alongside $2.4 million in closed-won revenue first-touch sourced by AI agents. Separate performance research ranks which segments work best: reactivation, event follow-up, and website visitor follow-up all outperform cold outbound, because AI SDRs work best when leveraging existing relationships or demonstrated interest rather than purely cold prospecting.

That last finding is the quiet takeaway for any business sitting on a CRM full of contacts it never followed up. Marc Benioff admitted Salesforce accumulated more than 100 million uncontacted leads over 26 years — "we just have not had the people." Segmentation by intent turns that dormant database into a prioritized list of warm opportunities instead of a graveyard.

It also explains why this example maps so cleanly to how Worqd approaches growth. The first step in any engagement is finding the bottleneck — buyer, offer, channels, response process, and data — before touching anything. Segmentation is that diagnostic step: it shows you which slice of your market is warmest and where follow-up is breaking down. And because reactivation is a top-performing segment, reviving old leads often beats pouring more budget into cold outreach.

One caveat worth noting: experts agree this isn't a "set it and forget it" play. Expect to train AI systems the way you would a new hire, with human oversight and handoffs where a real conversation matters.

Book a Growth Call and find your bottleneck. Every inquiry gets qualified in under 60 seconds, 24/7.

Applying Intent-Based Segmentation to Pipeline Recovery and Lead Conversion

Most companies sitting on a "cold" lead database are actually sitting on their cheapest source of pipeline — they just haven't segmented it by intent yet. When Jason Lemkin's team at SaaStr deployed 20 AI agents across their go-to-market over eight months, the highest-performing segments weren't cold prospects at all. They were reactivation campaigns, event follow-up, and website visitor follow-up — outreach to people who had already shown interest.

That finding reframes what "segmentation" means in practice. Instead of splitting your CRM by industry or job title, you split it by engagement history: who visited pricing pages, who abandoned a trial, who went dark after a sales call, who attended your webinar and never replied. Lemkin's rule is blunt — segment by context and intent, not demographics, and cap each campaign at 100–500 contacts so the messaging actually fits the segment. His team credits this approach with $4.8 million in additional pipeline over eight months.

The economics reinforce the point. A cost comparison of human versus AI SDRs puts a human SDR at roughly $98,000 per year for 15–20 qualified opportunities a month, while an AI SDR setup runs about $28,000 annually for 40–60 opportunities. When the highest-converting segments are the ones already inside your CRM, that cost gap makes reactivation the obvious first move rather than the last resort.

This is exactly where pipeline recovery work begins, and why Worqd starts every engagement by finding the bottleneck before touching anything. The typical segmentation for a recovery campaign looks like this:

• Lapsed contacts — leads who engaged once, then went quiet
• Abandoned trials or quotes — people who got 80% of the way to buying
• Uncontacted inquiries — leads that came in after hours or during busy weeks and never got a reply
• Former customers — warm relationships that already trust you

Salesforce illustrates the scale of that last bucket of missed demand. Marc Benioff has said the company accumulated more than 100 million contacts over 26 years it never had the people to call back. Most businesses have a smaller but proportionally identical problem: a CRM full of people who raised a hand and never heard back.

One caveat from practitioners: this is not a "set it and forget it" play. SaaStr's team reports that AI outreach takes real training and human oversight, with calls handed to a real person when the conversation gets serious. Done that way — small intent-based segments, fast qualification, human handoffs — your existing database becomes your highest-intent, lowest-cost source of booked calls.