Check out our upcoming events | Learn More
Articles

Metis Strategy Summit Highlights: Architecting the Enterprise for the AI Era

  • May 6, 2026
  • 9 Minute Read
Back to All Insights

Insights from the May 2026 Metis Strategy Summit in the San Francisco Bay Area. In the coming weeks, check out our YouTube channel and the Technovation podcast for full panel recordings.

The conversations at Metis Strategy’s May 2026 Summit reflected a noticeable shift in where senior technology and digital leaders are focusing their attention. A year ago, the prevailing question was whether AI investments could pay off. This year, executives moved past that debate and into a deeper one: what the enterprise itself needs to look like to scale AI responsibly, generate revenue from it, and build the trust and resilience that scale demands.

In his opening remarks, Metis Strategy founder and president Peter High framed eight forces shaping the current moment for companies responding to AI. Among them were AI agent autonomy doubling roughly every seven months, intelligence becoming abundant while judgment grows scarce, and architecture becoming a board-level political decision. Across the day’s panels, those forces translated into concrete enterprise choices: how operating models are being redesigned, how value is measured, where governance poses challenges, and what kind of leadership the moment requires.

From Productivity to P&L: Measuring What Matters

If 2025 was the year leaders rolled out copilots and counted hours saved, 2026 is the year they tied those investments to enterprise value. Shekar Pannala, CIO at The Hartford, described how his team set a hard target: cut average feature lead time from 150 days to 100 by year-end. For engineering teams, he said, the question is no longer whether they can build something, but how fast and how it changes the bottom line. “Software development is no longer the constraint.”

Sears Merritt, Head of Enterprise Technology and Experience at MassMutual, noted that for every AI initiative, there has to be a measure of success that translates back to a dollar value. That discipline is what allows organizations to navigate the SaaS-to-token economic shift without getting whipsawed by it, he said. Merritt was candid about MassMutual’s mainframe legacy, where switching costs had long made modernization untenable; new tooling, including Anthropic’s COBOL playbook, has changed that math.

Richard Jardim, SEVP and Chief Technology and Information Officer at CIBC, described a hub-and-spoke model with eight CIOs embedded in the businesses, each with a head of data and AI strategy. Jardim coined the labels “hunters” and “gatherers”: hunters in the business lines pursue use cases, while gatherers in the center build reusable capabilities.

Value itself takes different forms across organizations. Alexandra Guenther, SVP and CIO at Leidos, described moving beyond conventional ROI tabs to what she calls a “value calculus,” productivity, risk and opportunity, cost control, and differentiated capacity to drive net-new revenue. Kelly Fletcher, CIO at the U.S. Department of State, framed her two goals differently still: reducing administrative toil for foreign service officers, who move every one to four years, and creating decision advantage for diplomats. During the recent Middle East crisis, her team stood up an AI-driven common operating picture that gave senior leaders across the White House and State Department a single, consistently updated view of the data.

Architecture as Competitive Advantage

A theme that ran through nearly every panel was that scaling AI is fundamentally an architecture problem. At Hensel Phelps, CIO Dolores Mears and her team are building HP IQ, an orchestration layer that brings together schedules, drawings, contracts, financials, and trade communications behind one user experience, regardless of which underlying agentic solution is doing the work. Mears estimated the company is “probably 24 months away from completely digitizing the end-to-end experience.”

Fiona Tan, CTO at Wayfair, brought that architectural challenge into the context of agentic commerce. Even as some categories move toward autonomous, machine-to-machine purchasing, “humans still want agency,” she said, particularly in the home, a category where customers want more involvement in purchase decisions. The platform answer, in Tan’s view, is to build capabilities that can be exposed to external agents while doubling down on the human side of the brand: physical fulfillment, in-store experiences, and the assembly work that AI can’t do for the customer.

Prakash Kota, SVP and CIO at UKG, raised what may be one of the most pressing architecture questions of the year: identity and lifecycle controls for digital workers. With agents being spun up across the enterprise faster than governance can keep up, “every agent has to be tied to a human, because you can’t blame an agent when something goes wrong,” he said. UKG is working to extend the same lifecycle discipline that governs human employees and SaaS applications to agents themselves.

Vinod Bidarkoppa, until recently CTO at Walmart International, described how the company moved from operating distinct platforms across 19 countries to what he called a chassis model, one core platform with regional variants, much like an automaker. The result was striking. In the most recent year, inventory grew 4.3% while sales grew 4.7%. He summarized Walmart’s underlying mantra in a phrase: “Change is inevitable, but transformation is intentional.”

Scaling Trust, Identity, and Governance

As AI moves into customer-facing and clinical workflows, governance has ceased to be a back-office function. It is now either a strategic enabler or a constraint. Michael Pfeffer, Chief Information and Digital Officer at Stanford Health Care and Stanford School of Medicine, walked through Stanford’s FURM framework, which stands for Fair, Useful, Reliable Models. The framework evaluates a model not only on clinical performance, but also on the operational reality of acting on its outputs. A model that detects which patients need an echocardiogram is useless, Pfeffer noted, if there are not enough technicians to perform them.

Pfeffer also pushed back on the popular “human in the loop” framing. “It looks good three times and then it never gets checked again,” he said. The better model, in his view, is for the physician to receive AI-generated context at the point of care, with more information to make better decisions for the patient in front of them. The physician is not simply a backstop checking the AI’s homework.

Chad Wasserman, CIO at HCA Healthcare, which oversees roughly 47 million patient encounters a year, spoke about the prerequisite for innovation in a 24/7 operational environment: what he called “operational quiet.” HCA has deliberately tightened operational and cyber discipline so the team can stay focused on the generational opportunity in front of healthcare.

In financial services, governance is increasingly something to be engineered, not bolted on. CIBC’s Jardim described proactively educating regulators on how the bank uses and monetizes data, a playbook the team has run before, including with public cloud a decade ago. On the cybersecurity side, a fireside between Chris Davis, Partner and West Coast Lead at Metis Strategy, and Mandy Lu, who leads the firm’s cybersecurity practice, surfaced a parallel shift: AI is amplifying both good and bad actors, and defense is moving from rule-based to context-rich, behavioral, and predictive.

Human Change Is the Harder Problem

Across panels, executives circled back to the same point: technology is moving faster than people in many cases, making change management a critical priority. Leidos’s Guenther offered a memorable framing: change triggers a fight-or-flight response.  When navigating significant change, leaders should offer a clear, conviction-driven path forward and the assurance that team members have a place in the organization’s future.

At HCA Healthcare, that pathway has taken the form of a 12-year-old program called Healthcare Connection, in which cohorts of 10 to 12 technologists each month shadow nurses, ER teams, and supply chain operations. Wasserman, who shadowed a trauma center nurse, said the program connects technologists to the human stakes of the work, as well as the problems most worth solving with AI.

Mears from Hensel Phelps noted that strong adoption depends on whether frontline workers feel ownership of, not displacement by, the tools. “I’d love to see a construction foreman proud of an AI solution,” Mears said. Jaime Mineart, SVP and CTO at Caterpillar, made a related point about the limits of pure technology readiness: “It’s not just the technology readiness, it’s the customer readiness as well.”

From Data to Insights

Several panels pointed to concrete examples of AI moving past productivity and into top-line growth. John Gigerich, CIO at Keurig Dr Pepper, described a perfect-order tool that uses point-of-sale data, weather, seasonality, and historical patterns to recommend store-level orders to account managers. Instead of spending most of their time counting inventory, account managers now spend more time with store managers selling new products. The pilot drove a 3–5% percent improvement in on-shelf performance, which typically translates into a 1–3% percent sales lift. By year-end, the tool will be deployed across the company’s top 20 retailers.

Steve Bronson, CIO at Southern Glazer’s Wine and Spirits, described a similar effort with a platform called Proof and an AI module called Atlas, which gives sales reps next-best-action recommendations in the field. Bronson framed AI investment around three buckets: productivity, augmentation, and reinvention. He argued that the test of governance is whether it acts more like a steering wheel than a brake.

AI is also helping companies to build stronger supply chains and respond more quickly to disruptions. Sankar Chinnathambi, Global CIO at Driscoll’s, the world’s largest berry producer, described using AI to bring supply forecasting from a 25–35% margin of error down to 10–15%. For a fresh-fruit business operating against weather, geopolitics, and shelf-life pressure, the difference is material. He cited recent examples in which his team rerouted product in real time during a cartel-driven shutdown in Mexico and a weather-driven crossing closure in Morocco. Walmart’s Bidarkoppa described compressing product sourcing cycles from forty weeks to under eight, allowing the company to react to TikTok-driven trends in something close to real time. Ryan Kean, CIO at Total Quality Logistics, described AI helping his team detect a new form of carrier fraud, with identity theft replacing physical truck theft, by flagging anomalies in lane patterns and commodity types.

What AI Now Requires of Leaders

What does this shift demand of the leaders in the room? A panel of executive search leaders, Tarun Inuganti of Korn Ferry, Ryan Bulkoski of Heidrick & Struggles, and Mike Doonan of SPMB Executive Search, offered their view.

Bulkoski pointed to Heidrick’s recent board survey: 75% of board members named AI a top-three strategic agenda item, but fewer than 10% felt their boards were AI-ready. As a result, boards are playing a more active role in defining the AI leadership profile inside their companies. The bar for directors themselves has changed as well. AI expertise from five years ago is no longer enough, which has pushed clients to consider candidates without prior public-board experience, often paired with a coach for the first year.

Inuganti positioned AI literacy as an enterprise-level capability, not a CIO-only one. The skills that distinguish AI-era leaders, all three panelists agreed, are no longer primarily technical. Change management, storytelling, and influence now sit at the top. “Technical skills are table stakes,” Inuganti said. Doonan added that for the first time in years, CEOs are open to candidates from outside their industry. They are placing more value on the diversity of perspective that comes from a non-linear path, particularly as operating models themselves are being upended.

The day’s closing fireside with George Lee, Co-Head of the Goldman Sachs Global Institute, brought the conversation back to a theme that ran through the Summit: AI is now a physical infrastructure story, and architecture decisions sit at the intersection of geopolitics, energy, and trust. Marc Andreessen famously argued that software ate the world. Lee’s update is that AI ate the supply chain. From the scarcity of CPUs and industrial gases to the divergence in U.S. and Chinese public attitudes toward AI infrastructure, favorable views run roughly 30% in the U.S. and 80% in China, by Lee’s account, the leadership challenge is no longer purely technical. Lee closed on an optimistic note about the people who will define the next phase. He told the story of a 21-year-old new hire on a Goldman trading desk who, in her first 35 days, used an agentic harness to re-engineer a workflow that had existed for 35 years. “They know how to disrupt us,” he said.

What the Day Made Clear

When it comes to scaling AI, architecture, operating model, governance, and culture have all become first-class concerns, and the leaders best positioned for the next phase are those who can connect them. Productivity savings are table stakes; the real prize is in turning AI into a source of measurable value, defensible trust, and durable competitive advantage.

Insights from the May 2026 Metis Strategy Summit on how senior technology and digital leaders are redesigning enterprise architecture, operating models, governance, and culture to scale AI responsibly and turn it into measurable business value.