Lori Schafer in Forbes Tech Council: Streamlining Isn't Enough: Turning Static Data Into Autonomous Action

Reach the article in Forbes Tech Council here.

Enterprise companies have spent years modernizing their data stacks, but are they getting enough out of them?

Over the last decade, data engineers and officers have been migrating to the cloud, cleaning data lakes and investing in analytics. However, many still struggle with insights arriving too late, manual processes and execution falling behind market realities.

The issue isn’t access to data or the amount of data. The pain point is that most data systems and processes have been designed to inform decisions, not carry them out. It’s the common scenario: Don’t just deliver the news, tell me what to do with it.

This is why IT and data leaders are exploring agentic AI and autonomous workflows.

Agentic AI is the buzz of 2026. From a data perspective, in particular, agentic AI can analyze datasets, make guided decisions and act across the business. That said, its effectiveness depends on proper implementation and alignment with organizational processes.

Implementing Agentic AI

A study by MIT Sloan Management Review found that many companies are not only exploring agentic AI but also adopting it before they even have a strategy in place. The research found that 35% of the global executives surveyed have adopted agentic AI in just two years, with 44% deploying it soon. The study also noted that there are tensions underway with adoption, as companies attempt to scale too quickly or shoehorn the technology into unsupportive frameworks.

This can be a big problem. Agentic AI introduces a new kind of operational risk because autonomous systems surface weaknesses in data quality, decision ownership and organizational readiness rather than flaws in the technology itself.

When agents are deployed without clear boundaries, shared context or mechanisms to evaluate outcomes, failures don’t always announce themselves. They can manifest as small, compounding inefficiencies or misaligned decisions that only become visible over time.

For autonomous intelligence to deliver value, it must be designed as part of a governed system, not a stand-alone capability. That means defining decision authority, establishing feedback loops that measure impact and maintaining human oversight where judgment and accountability still matter.

The video player is currently playing an ad. You can skip the ad in 5 sec with a mouse or keyboard

When these foundations are in place, scalability becomes possible not because the system moves faster, but because it moves with control.

Moving Beyond Automation To Autonomous Intelligence

Another important step when implementing an autonomous intelligent workflow is moving away from an automation mindset.

Companies have been using automation for a long time, but it is fundamentally rules-based. Automated workflows excel at executing predefined tasks and working under stable conditions. For example, in retail, automated workflows might be advancing a replenishment order or syndicating a product attribute.

This approach is efficient, but it can be limited. By following a predefined task, automation can assume a predictable outcome. In other words, it executes instructions but doesn’t evaluate an outcome or adjust fluctuations in conditions. It also can’t reason across domains.

On the other hand, autonomous action empowers AI agents and intelligent workflows to continually learn from the data ingested, understand context and adjust decisions in real time. The critical difference is authority. Automation executes a task. Autonomous AI makes decisions with contextual awareness, assesses the total impact and initiates action across workflows.

When designed well, agentic systems can reduce data silos. Static data, even when perfectly modeled, doesn’t move the needle for businesses on its own. Adaptive, autonomous intelligence makes decisions in one domain that trigger action in another. It’s a shift from isolated automation to intelligent workflows in motion.

Unifying Data And Company Culture

When data and workflows are unified, teams stop operating in functional silos. Retail merchants, pricing analysts, supply chain planners and digital teams work from a shared system of intelligence.

Healthcare organizations also sit on vast amounts of static data, such as clinical data, operational data and patient engagement data, which can benefit from having knowledge orchestrated and shared autonomously across the organization. An adaptive, connected model can enable agents to coordinate follow-up appointments, adjust staffing forecasts and deliver personalized patient outreach.

In these environments, autonomous intelligence forges a cultural change. Each team across the organization, whether it’s a hospital or a retailer, can feel more connected. They see what’s happening across functions, and they spend less time reconciling data and more time shaping strategy.

To build a silo-free environment, companies need more than new tooling. They need to prioritize organizational alignment around how decisions are made and shared.

It’s best to start with establishing a common operating model for data: shared definitions, governed access and a clear “source of truth” that every function trusts. From there, leaders should define cross-functional workflows with explicit ownership, such as who approves pricing changes, who validates inventory signals and who is accountable for patient outreach outcomes, so that autonomous systems aren’t operating in ambiguity.

I would also suggest incentives that reinforce collaboration, like shared KPIs across departments, joint planning cadences and transparency into how decisions are triggered.

Pushing Toward Autonomous Intelligence

Enterprise businesses have invested in analytics, cleaned their data lakes and staffed up on data scientists. Yet they still need help turning insights into action. Agentic AI and autonomous, intelligent workflows can unify data and make real-time, reasoned business decisions.

Companies don’t need more dashboards or disconnected AI pilots. The next step is implementing AI as a living system, learning, adapting and acting across workflows without sacrificing trust or control.

I believe the key is to design autonomy with guardrails. Define what agents can decide, when human approval is required and which outcomes they’re allowed to optimize for.

Trust comes from transparency, including clear audit trails that show what data informed an action and why it was taken. When decisions are observable, measurable and easy to override or escalate, organizations increase speed while still maintaining accountability and control.

This is the next phase of enterprise intelligence, and it’s becoming a priority for the year ahead. Ultimately, companies aren’t defined by how much data they collect. They’re defined by how they make that data move.