AI investment keeps climbing. Performance lags. Martech Futurist | May 27, 2026
Three years into the enterprise AI cycle, productivity gains are no longer enough to satisfy boards. The gap separating the firms building durable advantage from the firms still chasing tool adoption keeps widening, and the McKinsey, Forrester, and HBR research from this week makes the structural reasons for that gap unusually clear.
This week's intelligence scan across Harvard Business Review, McKinsey, Gartner, Forrester, MarketingProfs, and the AMA Journal of Marketing surfaces an uncomfortable through-line. Adoption metrics keep climbing. Performance metrics largely don't. Three signals dominate:
The AI value paradox is now a board-level issue. McKinsey's latest framing puts hard numbers behind what most CMOs are sensing — adoption is widespread, but durable competitive advantage from AI is concentrated in a small group of companies that redesigned operating models, data architecture, and decision rights ahead of deployment.
The visibility vacuum has arrived faster than the playbook to fix it. As buyers migrate to AI answer engines, marketing organizations are losing the ability to see their buyers at all. Seventy percent of CMOs flag AI visibility as a top priority. Only 30% have assigned an owner.
Marketing teams ship more than they learn. Dashboards glow and campaigns deploy on schedule, while the underlying capability — turning insight into the next decision — atrophies. That deficit explains why firms with similar tech stacks are pulling apart on outcomes.
The center of gravity has moved. CMOs spent 2024 and 2025 buying tools. The work for the rest of 2026 is organizational: architecture, ownership, governance, and the learning systems that compound advantage over time.
Featured Insights
Where AI Will Create Value — And Where It Won't
Source: McKinsey & Company | May 21, 2026
McKinsey's latest analysis confronts the AI performance paradox directly: investment accelerates, adoption grows, sustained performance impact remains elusive. The piece distinguishes between three tiers of AI deployment — task-level acceleration (JPMorgan fraud scanning, BMW vision inspection), process-level embedding (Siemens predictive maintenance), and business-model innovation (where new growth and competitive moats actually form). Most enterprises remain in tier one, where time savings rarely translate to financial impact.
For marketing, the implication is structural. CMOs running AI inside existing campaign workflows will continue to see efficiency gains that don't show up in growth numbers. Tier-three value requires complementary changes — redesigned processes, new product and service models, ecosystem decisions about data and partners — built around the AI core. McKinsey names AI architecture as the single most consequential strategic decision, and in most enterprises that decision currently sits with IT rather than the CMO. That alignment will need to change.
Stop Replacing Traffic. Start Replacing Visibility.
Source: Forrester | John Buten, Principal Analyst | May 19, 2026
Buten opened his Forrester B2B Summit keynote by turning off the lights in an exhibit hall of nearly 2,000 people. The metaphor: AI answer engines are cutting visibility into buyer activity in half, and most marketing teams are responding by chasing the lost traffic when the deeper problem is lost line of sight to the buyer. Forrester's own community data shows 70% of marketers report AI visibility is a top priority for the CMO or CEO, while only 30% of companies have assigned a discrete owner for answer-engine visibility. The disconnect between urgency and accountability is producing what Buten calls "random acts of AEO" — consensus on importance without coordinated execution.
Traffic is becoming a downstream metric of a discoverability system that no longer behaves the way the dashboards assume. The work now is to see what buyers are asking AI engines, how brands surface in generated answers, and how competitors get described inside those answers — then use that to influence preference at the source. Marketing leaders who claim ownership of this redefinition will set the next CMO agenda for their organizations. The rest will keep optimizing channels that buyers are leaving.
Most B2B Marketing Organizations Are Not Short On Activity — They're Short On Learning
Source: Forrester | Christina Schmitt, Principal Analyst | May 21, 2026
Schmitt's argument lands hard for B2B CMOs. Campaigns ship, dashboards glow, and most B2B marketing organizations spend the year running variants of last year's plan while markets, buyers, and channels move faster than the team can adapt. Under-learning, in Schmitt's framing, is the real exposure. Her research identifies the pattern in high-performing organizations: learning gets engineered into how marketing operates day to day, with explicit ownership and shared infrastructure, rather than added on through quarterly training or post-campaign reviews. The closed loop — strategy informs execution, execution generates insights, insights reshape strategy — stays open in the organizations that compound, and breaks in the ones that stall.
Schmitt's challenge to leadership is the right diagnostic: "If your key people left tomorrow, how much of what your organization has learned would actually remain?" Honest answer for most marketing organizations: very little. Learning lives in individual heads, scattered slide decks, and unstructured Slack threads. CMOs should treat learning capability as core infrastructure on par with the data layer, with named owners, governance, and a real budget line.
Agentic AI Enters Its Enterprise Execution Era
Source: Forrester | Charlie Dai, VP, Principal Analyst | May 21, 2026
Dai marks the inflection where agentic AI moves from chat-based interaction to enterprise-grade work execution. The piece uses OpenClaw — featured in Jensen Huang's Nvidia GTC keynote in March — as the lens, but the analysis applies to the broader category. The shift is from systems that suggest work to systems that complete it, with three accelerants: execution-focused expectations, channel-native embedding (agents working inside Slack, Teams, and existing surfaces), and demand for local control on sensitive workflows. Risk profile shifts in parallel: when agents act, errors stop being wrong outputs and start being real-world consequences — data loss, compliance violations, cascading automation errors.
For marketing, "agentic" has stopped being a useful vendor adjective. It now describes an architectural decision about who holds execution authority over the work — the agent, the marketer, or some defined handoff between them. The constraint on adoption is governance velocity. Agents are acting faster than enterprise controls and approval workflows can keep up. Marketing leaders deploying agents in customer-facing workflows — response, personalization, journey orchestration — without explicit blast-radius controls and inspectable execution logs are absorbing brand risk that most boards have not yet priced.
Key Takeaways
1. Adoption is no longer the strategic question. Architecture is. McKinsey's framing pushes CMOs to ask where AI sits inside the enterprise, what data feeds it, who controls execution, and how learning from one deployment compounds into the next. Organizations measuring AI ROI primarily in hours saved are capturing tier-one value while peers a step ahead are building tier-three moats around redesigned business models.
2. Visibility deserves a named owner. The Forrester data point — 70% of CMOs treat AI visibility as a top priority, 30% have assigned ownership — exposes a real gap between executive recognition and organizational accountability. The first wave of answer-engine optimization will accrue to the companies that designate an owner, fund a measurement layer, and treat AEO as a cross-functional discipline that touches PR, product, content, and developer relations.
3. Engineer learning into the workflow. Schmitt's closed-loop framework gives CMOs a concrete operating-model question to act on now. Do strategy, execution, and insight actually connect inside your organization, or do they live in three separate teams with three separate tools and no shared cadence? The organizations that build learning into how the work flows day to day — through shared infrastructure, named owners, and faster feedback cycles — adapt quicker and connect insight to outcomes that sales and product can verify.
4. Govern agents before scaling them. Dai's risk framing belongs in the next agentic AI deployment review on the CMO calendar. Customer-facing agents introduce a category of failure that internal copilots do not. The organizations that will move fastest on agentic marketing are the ones building governance, evaluation, and rollback infrastructure ahead of the adoption curve.