Your Portco's Next SDR Hire Is Probably a Mistake

Lenny Rachitsky publishes a biannual state of the tech job market using data from TrueUp, which tracks openings at over 9,000 tech companies globally. The most recent edition, from March 2026, has a finding that every operating partner reviewing a portco headcount plan should sit with.

AI roles are hockey-sticking. Not growing — vertically accelerating. The category includes all roles at AI-native companies (OpenAI, Anthropic, Cursor) plus AI-specific roles at non-AI companies (an AI PM at Figma, an AI engineer at a legacy SaaS company). Both are going up sharply.

Meanwhile, design roles have flatlined since early 2023 — the same period AI started becoming genuinely useful for creative work. PM and engineering demand is at three-year highs. And a third of all open AI roles are concentrated in the Bay Area.

None of this is surprising. What is surprising is how slowly this shift is showing up in how portfolio companies staff their GTM functions.

---

What the Hiring Data Is Actually Saying

The design plateau is the canary.

When AI tools make it trivially easy to generate interfaces, images, and design assets, companies stop waiting for designers. They move faster without them, or they hire fewer. The market is reflecting that in real time — demand for design talent has gone flat while PM and engineering demand accelerated.

The same dynamic is coming for SDRs. It’s just a few years behind.

An SDR’s core job is research, personalization, outreach at volume, and follow-up. That is exactly what a well-built Clay table running chained Claygents, fed into HeyReach and Instantly, produces — at the output level of three to four humans, running continuously, without ramp time, sick days, or commission structure.

The SDR role isn’t disappearing tomorrow. But the SDR-heavy GTM model — ten reps booking meetings for five AEs — is being disrupted the same way design is being disrupted. The signal is already in the hiring data. Most portcos just aren’t reading it.

---

The Role That’s Actually in Demand

If you look at what’s being hired inside the AI role explosion, a pattern emerges: companies want people who can build and operate AI systems, not people who execute manual processes.

In GTM terms, that maps to a role that doesn’t have a widely agreed-upon title yet. GTM Engineer. Revenue Architect. AI GTM Lead. The job description varies but the core is consistent: someone who understands the commercial objectives of a sales team, can build automated enrichment and outreach pipelines, instrument them for performance, and iterate based on what the data shows.

This person is not a sales rep who learned to use Clay. They understand system design. They can write a prompt that produces specific, researched outreach rather than generic personalization. They know when a high bounce rate means a data problem versus a deliverability problem. They treat outbound as an engineering discipline, not a numbers game.

One person in this role, running a mature system, can generate the qualified meeting volume of a four-person SDR team — with better data, better targeting, and a feedback loop that compounds over time.

---

What Most Portfolio Companies Are Actually Doing

The standard portco GTM build at Series A or B still looks roughly like this: a VP of Sales, a handful of AEs, and an SDR team that’s supposed to generate pipeline. The SDR team is managed on activity metrics — dials, emails, meetings booked. The assumption is that more SDRs means more pipeline.

That assumption held when outbound was a purely human-paced activity. It’s breaking down now for a few reasons.

First, the productivity ceiling of a human SDR is real. Eight to ten cold emails a day with genuine personalization is about the limit before quality degrades. An AI system with proper agent chaining runs that across hundreds of accounts simultaneously.

Second, SDR ramp time is expensive. Three to six months before a new hire is contributing real pipeline. An AI system is productive immediately — and unlike a human, it doesn’t leave for a competitor after twelve months.

Third, the cost profile is structurally different. An SDR costs $60-80k in salary plus commission, management overhead, and the inevitable churn and rehire cycle. A well-built outbound system running Clay, HeyReach, and Instantly costs a fraction of that — and the marginal cost of reaching more accounts approaches zero.

None of this means AEs are going away. Closing is still a human skill. But the pipeline generation layer — the SDR function — is being automated in the same way design work is being automated. The companies that figure this out first will have a structural cost and velocity advantage over those still running the 2018 playbook.

---

What an Operating Partner Should Be Asking

When you’re reviewing a portco’s GTM headcount plan, the question isn’t “how many SDRs do they have?” It’s “do they have anyone who can build the system that replaces manual SDR work?”

Most portfolios don’t. They have sales reps and sales managers. They might have a RevOps person managing the CRM. They don’t have someone who wakes up thinking about Clay waterfalls, Claygent prompt architecture, and contact data decay rates.

That’s the hire that compounds. And it’s the hire that most portcos aren’t making because it doesn’t fit neatly into the sales org chart they inherited from the last decade.

A few questions worth adding to the next portfolio GTM review:

• What percentage of outbound is AI-automated vs. manually executed by SDRs?
• Who owns the outbound system architecture? Is there a single person accountable for how it’s built and how it performs?
• What does it cost to reach 1,000 new qualified accounts? How does that compare to the cost of an SDR headcount doing the same work?
• When did the outbound system last get iterated? What changed and what did the data show?

The portcos that can answer those questions are ahead. The ones whose sales leader responds with “we’re ramping two new SDRs this quarter” are building toward a headcount model that’s already being disrupted.

---

The Operating Partner’s Leverage Point

The AI hiring explosion Lenny’s data shows isn’t just a tech industry curiosity. It’s a leading indicator of where competitive advantage is shifting.

Companies are racing to get AI-fluent people into every function. GTM is one of the last functions to wake up to this — partly because the SDR playbook is so deeply embedded in how B2B companies are built, and partly because the ROI on AI GTM investment is harder to model than the ROI on another sales hire.

But the math is changing fast. Operating partners who push portcos to build the system — and to hire for it — before competitors do are giving them a structural advantage that compounds. The portcos still running the 2018 SDR model are building technical debt in their GTM motion, the same way a company that ignored cloud infrastructure in 2012 paid for it for the next decade.

The data is already showing where the market is going. The question is whether your portfolio is reading it.

---

The Dark Funnel covers AI GTM infrastructure for operators who build, not advise. New posts when there’s something worth reading.