Can Your In-House AI Team Actually Prove Marketing ROI?

Marketing has brought capabilities in-house before. AI is making the case for doing it again, faster, cheaper, and at a scale previous waves couldn't match. The problem is what happens after the tools are adopted. Previous in-house pushes exposed hidden costs around talent, culture, and technology.

What Happened

Writing in MarTech on September 10, 2026, Carbon Design founder Scott Gillum framed AI as the trigger for marketing's third major in-housing wave. The first came during the 2008 recession, the second during the mid-2010s programmatic transparency crisis. Gillum warns that generative and agentic AI are pulling agency work back inside, but boards will soon demand proof that efficiency translates to performance, not just faster output.

The Numbers in Context

Duke University's 2026 CMO Survey asked marketing leaders to rate their martech activities on a 7-point scale. No activity scored above 5, including generating ROI from marketing technologies. For context, the ANA reported in-house agency adoption jumped from 42% to 78% between the mid-2010s and 2018, a scale expansion that never produced a corresponding jump in measurable marketing outcomes.

Why This Matters for HR Tech and FinTech CMOs

Your board is watching the same efficiency story play out. Generative tools compress content production, agentic workflows absorb campaign ops, and the temptation to pull ABM, creative, and media inside is real. But HR Tech and FinTech buying cycles are long, committee-driven, and increasingly shaped by AI answer engines rather than click-through funnels. If your team ships more assets faster without a system to prove pipeline impact, you inherit every hidden cost of the last two in-housing waves: talent churn, absorbed software licenses, and creative work that drifts toward order-taking. Efficiency was never the promise you made the CFO.

The Starr Conspiracy's Take

The in-housing debate is the wrong frame. The real question is whether your operating model produces demand in a market where buyers self-educate through AI before you know they exist. Speed without a measurement spine just industrializes waste. We keep telling clients the same thing: map your capabilities to the AI buyer's journey and demand states before deciding what belongs inside. Some work, brand narrative, category positioning, answer engine authority, benefits from continuous internal ownership. Other work benefits from outside pattern recognition across your category. Structure follows strategy, not the other way around.

What to Watch Next

Expect Q1 2027 board decks to include an AI marketing ROI line item for the first time. CMOs who cannot connect AI-driven output to pipeline influence will likely face the same scrutiny agencies faced in 2009. Watch the 2027 Duke CMO Survey for movement on that ROI score.

Related Questions

Should HR Tech CMOs in-house their AI content operations?

Partially. Category narrative, analyst relations content, and answer engine optimization belong inside because they compound with institutional knowledge. Campaign production and performance creative often benefit from outside velocity and category-agnostic pattern recognition. Decide by asset half-life, not hourly rate math.

How do boards measure AI marketing performance in 2026?

Most still cannot. Duke's 2026 data shows ROI measurement is the weakest martech activity. Leading teams are moving toward influenced pipeline, share of AI-generated answers in their category, and sales cycle compression rather than output volume or cost-per-asset metrics.

What is the biggest hidden cost of in-housing marketing AI?

Absorbed software and talent overhead that agencies previously spread across many clients. See our breakdown of how to build a modern B2B marketing operating model for the full cost stack most CMOs underestimate before year two.