Each Monday, a short list of the sharpest sales + GTM thinking from the past week. Not summaries — read the originals, they're better. Just what's worth your time.
Four pieces this week.
1. Trish Bertuzzi — "The hybrid pod wins, but the volume-quality inversion is the real story"
Trish Bertuzzi (Founder, The Bridge Group) has been the most rigorous benchmarker of inside sales teams for two decades. The Bridge Group's 2026 SDR Metrics & Comp report puts hard numbers on the hybrid AI+human pod experiment most RevOps teams have been running without data. The headline: hybrid configurations — one human SDR per two AI SDR seats — book 1.9x more meetings per dollar than pure-AI pods and 2.4x more than human-only. Cost per qualified opportunity fell from $487 (human-only) to $224 (hybrid). On its face, a clean AI-wins story.
The second number is the one worth sitting with: per-rep monthly outbound volume rose from roughly 1,150 to 7,400 with AI augmentation, while raw reply rates fell from 4.7% to 2.9%. Volume is up 6x. Conversion is down a third. The efficiency gain is real — but it's coming almost entirely from denominator expansion, not from signal quality improvement. Bertuzzi's benchmarks make visible what most AI SDR deployments are obscuring: you can automate volume indefinitely without moving the conversation quality needle at all.
→ The Bridge Group's SDR Metrics & Comp report is the primary data source.
2. Sangram Vajre — "AI doesn't fix broken GTM; it scales it"
Sangram Vajre (Co-founder, GTM Partners; WSJ bestselling author of MOVE) runs GTMonday, a weekly research note at 175K+ subscribers. His piece "7 GTM Truths AI Won't Change" is the clearest short-form articulation I've seen of why AI feature adoption keeps climbing while GTM performance stays flat. The argument: growth is a system, not a capability — and AI doesn't change the system, it just runs faster through it. If ICP targeting is wrong, AI targets wrong faster. If value messaging is misaligned to buyer outcomes, AI misaligns at scale. If pipeline governance is wishful, AI generates confident-sounding wishful projections.
The research anchor is High Alpha's portfolio benchmark: companies with AI-native GTM features saw strong adoption, but monetization and retention didn't follow. AI fixed the build rate. It hadn't fixed alignment. GTM Partners' own research found 60% of companies they work with can't articulate or quantify their GTM ROI at all — not a tooling problem, a system problem.
→ Vajre's GTM Operating System framework is the conceptual architecture behind the research notes.
3. Pete Kazanjy — "When AI does the work, the rep doesn't build the skill"
Pete Kazanjy (founder of Modern Sales Pro, the largest sales ops and enablement community; author of Founding Sales; former co-founder of Atrium, acquired by Fullcast) has been thinking more carefully than most about the second-order costs of AI automation in sales. His GTMnow episode GTM 120 makes a structural argument: the skill-building value of sales tasks isn't just in completing them — it's in the pattern recognition, edge-case exposure, and judgment that accumulates from doing them. When AI prospecting replaces manual prospecting, the rep stops building the intuition about what a real buying signal looks like versus a polite reply. When AI call analysis replaces rep-led self-review, the rep stops developing self-awareness about their own talk patterns and objection gaps.
Kazanjy names these "skill byproducts." They're not the primary output of the task — they're the residue that accumulates in the rep as a side effect of working. Automate the task, you also automate away the byproduct. At scale and over time, you get high-volume execution from reps with an increasingly thin skill base underneath it. The pipeline metrics look fine until they don't.
→ GTMnow also ran a companion essay "The AI Tradeoff: Preserving Human Skills in an AI World" that extends the argument into org design.
4. Kevin "KD" Dorsey — "The first operating layer is inspection — and no AI surface replaces it"
Kevin "KD" Dorsey (CRO, LeanScaper; founder of Sales Leadership Accelerator) has built his reputation on an unfashionable but empirically defensible position: the most important skill a sales manager has isn't coaching — it's diagnosis. Coaching without diagnosis is theater. You can run playbooks, script calls, and review recordings all day; if you can't first identify the actual constraint in the rep's conversion funnel, you're solving the wrong problem well. His SaaSiest talk on the five-step roadmap to high-performing teams is the practical framework: start with structured data, diagnose the break point in the rep's funnel, then coach to the specific constraint — not to a generic "improve" directive.
The diagnostic layer he describes is an operating discipline, not a technology. It requires knowing what baseline conversion looks like at each stage for each rep type, what a deviation signal means versus noise, and how to distinguish a skill problem from a process problem from a pipeline problem. That's judgment built from consistent, structured inspection over time — not something you can shortcut with a dashboard summary.
→ Dorsey's LinkedIn is where he writes most frequently; Sales Leadership Accelerator is the program hub.
Common thread this week. All four — Bertuzzi, Vajre, Kazanjy, Dorsey — are pointing at the same structural gap: AI raises execution volume without automatically improving system quality. The insight threading through all of them is that volume without governance is a more expensive noise machine. Bertuzzi has the data on where quality goes when volume scales. Vajre has the research on what misalignment looks like at speed. Kazanjy has the theory of why skill bases erode beneath the activity layer. Dorsey has the diagnostic discipline that holds it together.
Next Monday: a new four.
If you've read something this week worth flagging in next week's list — hello@mallin.io.