If 2026 Is the Year of Execution, the Missing Skill Is Commercial Fluency in Governance

The AI sandbox is officially closed. Welcome to the year of execution. Boards are spending less time on abstract promises and demanding hard data on how AI actually performs in their environment, closing the gap between AI's promise and its practical reality. The money has not dried up. It is just moving to different owners inside the business. Centralized R&D labs are seeing their exploration budgets flatten, with those dollars now sitting in operational P&L lines like customer service, security, finance operations, and supply chain.

Lawrence Roberts has spent 27 years watching these exact kinds of procurement cycles play out in the IT channel and distribution industry. Currently serving as Senior Global Manager for Strategic Go-to-Market Enablement at TD SYNNEX, following global and EMEA marketing leadership roles at Tech Data and Avnet, he oversees enablement tools across five regions for high-growth technologies, including AI, cloud, and cybersecurity. From his vantage point, sales conversations are fundamentally changing. Buyers now focus more on governance and long-term operation than on transformation promises.

"The dollars once labeled 'AI exploration' are being relabeled 'AI operations' and pushed into the P&L lines that own the workflow, such as customer service, security, finance operations and supply chain. This means the buyer is no longer the CIO's innovation lead; it is the line of business owner with a margin target." Roberts says.

Capital continues to flow into the market. The largest cloud providers forecast up to $665 billion in AI infrastructure spend in 2026, supporting projections that worldwide AI spending will reach into the trillions. But Roberts notes the more important transition is happening inside the enterprise, where line-of-business owners have become the primary economic buyers with margin targets and expectations for defensible, multi-year returns.

Governance eats transformation for breakfast

The operational focus highlights a gap Roberts sees in many organizations between how fast teams deploy AI and how fast governance processes adapt. That friction can trap deals in legal review. To unstick them, many sellers now lead by proving operational safety before they talk about transformation. Winning enterprise deals more often depends on a deep understanding of the customer's business and a clear path to solving measurable challenges.

"Recent IDC work shows AI adoption is scaling faster than the controls around it, and that gap is the single biggest reason deals are stuck in legal review," Roberts says.

The ratio has shifted decisively. "The honest answer is that transformation is now the warm-up, not the pitch," he says. "I would put it at around 60% governance, compliance and auditability, 40% feature set, and that gap widens every quarter."

For sales teams, the repositioning is tactical. "Sales teams who win will lead with three things: how the data is sourced and traced, how the model is bounded and observable, and how the customer can produce evidence for an internal audit, an external regulator, or a board risk committee," Roberts says. "The repositioning is not about abandoning transformation language. It is about earning the right to use it by leading with control around real use-case applications."

The infrastructure is splitting

These strict new demands are physically rewriting enterprise tech stacks for many buyers, pushing them toward specialized AI deployments and pragmatic hybrid infrastructure. Customers are starting to focus on the management plane: who owns the control policies, how they are enforced across environments, and where those policies live.

Small language models are gaining traction, but not for the reasons most demos would suggest. "Small language models are an architectural choice, not a demo-driven one, which means platform engineering and security teams are involved from the outset," Roberts says. "The design phase becomes more deliberate, as organizations are sizing models to fit specific workloads rather than forcing workloads to conform to a model."

At the same time, the economics of where inference lives are hardening. "Recent licensing changes have effectively multiplied virtualization costs that had remained stable for over a decade," he says. "At the same time, hyperscaler egress fees and the broader silicon constraint, what we are calling the 'foundry wall,' is putting hard limits on where AI inference can economically live."

Roberts observes that leaders who handle the "cloud-appropriate" conversation well come in with a workload-by-workload portability story, not a platform fight. "Cloud is still right for large-scale training and elastic bursts. Edge is where most production inference belongs. On-prem is where the regulated, the latency-sensitive, and the heavily data-gravity workloads sit."

Robots keep the receipts

Meanwhile, the physical world of robotics offers a highly instructive reference point for software teams. The rise of the Robotics-as-a-Service model, where fleets are leased as an operating expense, proves how making a technology financeable through standard procurement accelerates adoption.

Roberts points out that this hardware-centric model establishes an uncompromising standard for ROI measurement. "The physical world forces honesty about ROI," he says. "A robot either does the work, or it does not. There is nowhere for a vague 'productivity gain' to hide. Secondly, robotics teams have to design for failure visibly. That forces a level of resilience engineering and observability that software GTM has historically been able to skip."

Software GTM teams can borrow from that discipline by building observable resilience into their pitches. Some vendors are finding they must clearly articulate what buyers gain by choosing a managed product over a lightweight, internally built application. "Seat-based licensing models may need to adapt as AI becomes the new interface for employees," Roberts says. "Equally, it is going to be important for software GTM teams to explain why someone should be investing in their application, versus vibe coding one themselves."

The missing skill

As the enterprise technology stack matures, many organizations are discovering an opportunity to evolve their sales motions. The technology is now working reliably enough in production that the human element is catching up to a new baseline. GTM leaders are beginning to retrain their teams to navigate layered procurement cycles and align marketing, sales, and enablement around governance-heavy conversations. Bridging that gap remains a major enablement priority in the channel.

"A great seller in 2026 needs to be able to walk into a room with a CIO, a CISO, a Chief Risk Officer, and a procurement lead, and have a credible conversation about data lineage, model observability, audit evidence, and total cost of operation, all in one sitting," Roberts says.

That fluency is what separates the teams landing budgets from the ones stuck in pipeline. "If 2026 is the year of execution, the missing skill is what I would call commercial fluency in governance," he says. "This skill is about being able to balance the perspectives and needs of multiple stakeholders, which is genuinely rare. Most enterprise sales teams were built for a feature-and-benefit world. The buying committee they are now selling into is a risk-and-resilience committee."