Claim
The bigger swing for go-to-market in the AI era is structural: build a cross-functional pod (product + engineering + data + marketing tech + operations) that owns "AI in GTM" as a single mandate, rather than asking each existing GTM function to bolt AI onto its current playbook.
Mechanism
Bolting AI onto each function preserves the seams between functions, which is exactly where the leverage is highest. A cross-functional pod can collapse those seams: product builds the agent, marketing tests the prompts, ops integrates the data, sales feeds back the win signal — all on one cadence, with one P&L for the bet. The pod's job is to find the swings, ship them, and pass the playbook back to the rest of the org and to the company's own customers.
Conditions
Works when:
- The org has executive cover for a pod that crosses traditional functional lines.
- The pod can ship to production and measure outcomes (not just write internal reports).
- Customers' product surfaces overlap with what the pod builds (so the playbook can be productized).
Fails when:
- The pod is staffed only with marketers; missing product/eng/data sinks the bets.
- The pod is positioned as "innovation theater" without P&L authority. The bets stay below the bar that would matter.
- The org's existing GTM functions resist the cross-cut and starve the pod of context.
Evidence
"Instead of integrating AI across the GTM, you need to rebuild your GTM around AI."
Numbers reported by HubSpot's Flywheel AI group, the precursor to the Agentic GTM & Systems group Kieran now leads:
- 345,000 net new accounts added to TAM in a year
- 82% of inbound chats handled with zero humans
- 1,850% growth in qualified leads from ChatGPT and Perplexity
- 10,000+ meetings booked per quarter from AI-personalised outreach
- 13% lift in win rate on deals using AI guidance
- 60% of internal support inquiries resolved without a human
- 7-point lift in customer save rate
— Kieran Flanagan, LinkedIn, 2026-05-01. (Yamini Rangan, HubSpot CEO, reported the same numbers in a separate post — see raw/articles/agent-first-gtm-hubspot-2026-04-28.md.)
Signals
- A named pod with members from at least three traditional functions (product, eng, marketing, ops).
- Bets framed as "rebuild this GTM motion end-to-end with AI," not "use AI tool X in our existing motion."
- Outcomes reported as customer-product wins (TAM expansion, conversion lift, deflection), not as internal productivity metrics.
- The pod's playbooks are productized for the company's own customers within 1-2 quarters of internal proof.
Counter-evidence
For smaller GTM teams without the capital for a cross-functional pod, "integrate AI across existing functions" may be the only realistic path. The rebuild claim is a frontier-org pattern; not every org has the org-design slack to execute it. There is also early evidence that pods staffed without strong product/engineering depth produce theater rather than wins.
Cross-references
- Agents work when treated as a team, not a single super-tool — the per-role agent shape the pod's bets often take
- The unlock for AI agent productivity is management skill, not technical skill — how non-engineer marketers contribute to a cross-functional pod