CIO 100 Leadership Live New York: CIOs push past AI pilots for measurable returns

Technology executives from across the New York metropolitan area gathered July 16 at Convene, One Liberty Plaza, for CIO 100 Leadership Live New York, a full day of roundtables and panel discussions on enterprise AI investment, governance, and organizational change.

Several key areas of consensus emerged throughout this highly interactive event. Infrastructure fragmentation continues to block the path to securing returns on AI investments prompting leaders to understand rising cloud spend attributed to large language model utilization. This has caused a growing number of organizations to refocus on on-premises and hybrid options in C-suite and board-level capital planning conversations. Speakers, along with comments from the audience, described a shift from project thinking to product thinking, with smaller multidisciplinary teams moving faster than legacy structures.

Several participants repeatedly warned that automating broken processes just amplifies dysfunction. Governance and measurement remain unresolved, with usage metrics still getting mistaken for business value. One of the panels explored how CIOs may benefit from applying venture capital-style scrutiny to enterprise bets, weighing team execution as heavily as the technology itself. The throughline was a redefinition of the CIO role, from technology executor to business strategist fluent in revenue, board engagement, and transformation ownership.

Morning roundtable tackles AI infrastructure

The day opened with an invitation-only executive breakfast roundtable, “Beyond the Pilot, Building the Infrastructure for Real AI Returns,” co-hosted by Unisys and Dell Technologies. Over a dozen executives representing major public and private sector organizations across the New York metropolitan area joined Steve Hollander, senior director of Americas global alliances at Dell Technologies, and Matt Marshall, CIO at Unisys for a workshop-style discussion.

The session explored the strategic, operational, financial, and technological issues that must be mastered to optimize infrastructure decisions and separate organizations that are experimenting with AI from those competing on it. Discussion questions probed how CIOs measure whether AI investment is translating into business results, how they can break the cycle of fragmented and siloed AI deployments, how boards are beginning to scrutinize seven-figure token spend and whether on-premises or hybrid infrastructure can rein in costs.

The take-home point: the organizations pulling ahead are the ones that stopped treating AI as four separate problems, strategic, operational, financial, technological, owned by four separate functions, and started running it as one coordinated decision. Fragmentation is the actual cost center here, not the token spend itself. A CIO who solves the infrastructure question in isolation from the governance question, or the cost question in isolation from the talent question, ends up optimizing one silo while the other three keep bleeding value. Competing on AI, instead of just experimenting with it, means the finance, operations, technology and business sides are reasoning from the same picture of what’s being built and why, so the tradeoffs get made once, together, instead of getting re-litigated at every handoff.

Forum sessions open with a mandate for growth

Following breakfast, the main forum program began with “The New CIO Mandate, Delivering Growth, Not Just Technology.” In a moderated conversation, Laksh Nathan, chief information officer at Paramount Skydance, drew on his experience with mergers, enterprise transformation and AI-enabled development to describe a shift from project and application management toward a product-centric operating model. Nathan addressed how smaller, multidisciplinary teams are changing expectations on both the business and technology sides of the enterprise, and what mindset changes CIOs must lead to turn AI into an engine of growth rather than a cost center.

PwC followed with a session on “Designing the Intelligent Enterprise, From AI Investment to Evolving Operations.” Darren O’Meara, principal and chief technology officer for managed services, and Meghna Shah, principal for engineering and AI, examined why fragmented outcomes persist even after heavy investment in technology and transformation.

The intelligent enterprise, they posited, is less about working toward achieving specific technology outcomes and more about creating operating models that integrate strategy, technology, operations, and governance into one system. This, they explained, requires linking AI, data, and decisions across the business and will leave an indelible mark on how decision rights are redesigned, funding models are developed, and accountability is enforced to accommodate the speed of the agentic economy.

Talent, tradeoffs, and the cost of getting it wrong

The session “Return on Transformation: Time, Talent, and Tradeoffs” — with Prashant Hinge, chief information and transformation officer at MSIG USA; Joseph Gimigliano, chief technology officer at Northwell Health; and Eduard de Vries Sands, AI executive advisor at PatientPoint — examined why transformation initiatives so often lose their way.

The main culprit, even today in 2026, continues to revolve around a persistent instinct for technology implementations to become the objective rather than the means to a measurable business outcome. The panelists made the case for doing the incredibly difficult work of re-engineering (if not entirely re-imagining) existing processes before automating them and then placing smaller bets inside that bigger vision.

Ricky Thakrar, head of sales and account management at Zoho, took the stage to present “Smaller, Smarter, Safer, The Enterprise AI Architecture Most Leaders Get Backwards,” arguing that constrained, context-rich architectures consistently outperform expensive models bolted onto fragmented systems.

A round of Hot Topic Discussion Groups and a networking lunch followed, including the Next CIO Luncheon featuring Robert Half Regional Director Jason Deneu.

Afternoon sessions turn to security, scale, and investment signals

CSO and CIO Contributor Joan Goodchild moderated “Securing Trust in the Agentic Economy,” a discussion with Marlowe Cochran, CISO at the New York State Education Department, and Gee Rittenhouse, vice president of security services at AWS, on how organizations are balancing speed, innovation and security as AI agents move from experimentation into productization at scale.

Rittenhouse framed agentic risk as closer to human risk than traditional software risk, describing how an independent agent acting in a non-deterministic way really does look like a potential insider threat, pushing CISOs toward behavioral monitoring over static workload protection. He tied this to a structural shift in defense, noting it’s hard to do agentic security if you’re not observing it, putting observability at the center of agentic risk management.

Cochran concurred, adding that many of the key tools that are needed to move into the agentic economy already exist, but must be implemented more aggressively, comprehensively and even more creatively. CISOs don’t need to invent an entirely new security discipline for the agentic era so much as extend identity management, access control and monitoring frameworks they already run to cover a new class of non-human actor — agents.

A session on “AI, From Experimentation to Enterprise Impact” brought together Meagan Gentry, national AI practice manager and distinguished technologist at Insight and Yuri Gubin, chief technology officer at DataArt, for a candid look at why pilots stall before reaching scaled production and what operating capabilities, governance, cost visibility, continuous education, must be in place to sustain AI once a proof of concept works.

During the session’s Q&A segment, a discussion emerged around how proof-of-concept success can result in a false signal, raising questions about whether pilots should be considered successful before the intended outcomes have had time to materialize, and drawing a distinction between measuring usage and adoption versus measuring business value.

The panelists explored how CIOs can identify the small number of transformational AI opportunities worth pursuing rather than managing hundreds of incremental use cases, and even challenged whether prioritization is the CIO’s job at all. The discussion closed on a sequencing question with real strategic weight, whether AI-first strategies are putting the technology ahead of the business problem CIOs are trying to solve, and what role CIOs should play with boards in defining the outcomes AI is expected to support.

A shift in perspectives

The “Think Like a VC, Investment Shifts Towards Focused AI Applications” session featured three venture investors, Aaron Darr, partner at Lead Edge; Isabelle Phelps, partner at Lerer Hippeau; and Marshall Porter, general partner at AlleyCorp. The panel explored how investors evaluate risk and talent in a market where products and competitive positions can shift within months, and what separates a focused AI application with durable enterprise value from an AI wrapper built to chase a trend.

The panel challenged the enterprise instinct to seek certainty in a market moving this fast, questioning whether CIOs should stop looking for technologies that will future-proof the enterprise and instead grow more comfortable continuously reassessing their bets. Investors framed this as a deliberate departure from the traditional low-tolerance-for-failure posture that has long governed enterprise technology purchasing, arguing that the search for certainty has itself become a risk in a market where products and business models can shift within months. The discussion pressed CIOs to weigh how they can adopt a more dynamic investment mindset without compromising the enterprise security, governance and accountability their organizations still depend on.

A Lightning Insights followed, featuring five-minute briefings from Insight, Platform9 and Console, followed by Keystone Senior Principal Ellora Sarkar’s talk on why most enterprise AI investment fails to produce measurable value and what separates the small share of firms capturing real return on investment from the majority still stuck in pilots.

Closing the day

The forum closed with “What’s Next for the CIO, Preparing for the Next 12 to 24 Months,” a fireside conversation with Leif Maiorini, CIO for corporate services at Omnicom. Maiorini discussed why business processes need to be redesigned for agentic speed rather than automated around existing human workflows, how organizational structures may shift as autonomous agents reshape visibility and decision support, and where sustainable differentiation will come from once AI capability itself becomes widely accessible.

Maiorini encouraged the industry to clearly distinguish between nondifferentiated services that should be made as efficient as possible and the differentiated capabilities that actually influence why customers choose to do business with an organization, once the major efficiency gains from optimization and AI have been captured.

He was candid about the governance gap agentic systems open up, noting that agents lack the professional reputation, personal accountability and inherent constraints that shape human behavior, which creates new risk when autonomous decisions occur at machine speed. That combination, reinvesting efficiency gains into genuine differentiation while building governance models suited to non-human decision-makers, framed his closing case for why human creativity and judgment remain the enterprise’s most durable asset even as the underlying technology becomes commoditized.

Join the CIO 100 Awards & Conference Aug 17–19, 2026 at Omni PGA Frisco Resort & Spa, Frisco, TX — where top IT leaders celebrate innovation and connect.  Learn more to attend or partner.