AI models are getting more powerful — and the differences between them are getting smaller. What increasingly determines the value companies get from AI is whether the model understands how their organization actually works.

Here's the deal: Enterprises that have built institutional intelligence — a working record of how their organization actually operates — can make AI more productive at the organizational level. Without this intelligence, even the most advanced AI models can struggle to deliver meaningful business value.

Key numbers: Companies are rapidly scaling AI, but enterprise-level financial impact remains largely unchanged. According to McKinsey's 2026 State of AI report:

• 37% of respondents attribute at least some earnings before interest and taxes (EBIT) to AI use, about the same as last year.
• Just 6% qualify as AI high performers: companies that attribute at least 5% of EBIT to AI and call the impact "significant." That share has also stayed flat.

What this means: More companies are putting AI to work, but scaling the technology doesn't automatically translate into greater value across the business. What actually matters and gets results is the context organizations give AI.

That includes knowledge that often isn't captured in a manual or database — which exceptions are routine, which issues get escalated, who can override a process and why, and how teams actually move work forward.

Without that record, organizations can experience what we call execution drift, the gap between what leadership intends and what actually happens as work moves between people, teams and systems.

Worth a mention: Whether an AI agent is useful on day one or starts from scratch depends on the context it can access.

• A new employee can build knowledge over months or years.
• A new vendor may never gain access to all of it.
• An AI agent can use that knowledge immediately — but only if the organization has captured it in a form the agent can access.

When that record is codified, every agent an enterprise deploys can draw from it and potentially add to it. When it isn't, every new deployment risks starting from zero, regardless of how capable the model is.

The takeaway: The next gain in AI returns will come from giving AI better access to institutional intelligence — putting how an organization actually works into a form software can understand and act on.

Work management platforms can help create that foundation by bringing the people, processes, workflows and data behind day-to-day execution into one environment.

Smartsheet connects teams, workflows and AI on a centralized work management platform used by more than 85% of Fortune 100 companies.

The platform is designed to help organizations:

• Align work, people and AI: Connect work across teams while maintaining alignment with organizational goals.
• Unify workflows and data: Connect systems of record, everyday applications and AI tools rather than leaving intelligence siloed in individual systems.
• Adapt quickly: Give teams flexible tools that can evolve as projects, processes and priorities change.
• Scale securely: Apply enterprise controls and visibility while making AI actions traceable.
• Build on existing knowledge: As more work happens within the platform, teams can create a richer record of how work gets done.

The idea: AI can provide intelligence. Enterprises still have to provide the organizational context that makes that intelligence useful.

Codifying that context can help turn AI from a tool that knows what to do into one that understands how work actually gets done.

Put institutional intelligence to work with Smartsheet.