AI agents can automate change management tasks. But automating tasks is not the same as managing change across an enterprise.

The rise of agentic AI is creating a compelling proposition for change teams. Why spend hours analyzing documents, drafting communications, building stakeholder plans, or summarizing feedback when an AI agent can do much of that work in minutes?

Major consulting firms are already taking agentic approaches to market. Accenture is applying AI and agentic systems to transformation and change, while Deloitte is actively promoting agentic AI, multi-agent systems, and AI-enabled change solutions. Deloitte describes AI agents as digital workers capable of identifying, planning, and executing work with increasing autonomy.

We think this direction is right.

The question isn't whether change teams should use AI agents. They absolutely should.

The more important question is: what happens when you have dozens of agents doing change management work across dozens, or hundreds of initiatives?

That is where the distinction between AI automation and an enterprise change management platform becomes critical.

An agent can do the task. But does it understand the enterprise?

Imagine an organization deploying separate agents to create communications, conduct impact assessments, develop stakeholder plans, analyze sentiment, and prepare reports.

Individually, they may be excellent.

But each agent needs context. Which projects are affecting this employee? Which business unit do they belong to? What other changes are landing at the same time? What was their previous change impact? Who is authorized to see their information? What benefits was the initiative supposed to deliver? Are those benefits actually being realized?

Without a common data and governance layer, each agent is working with only part of the picture.

And this isn't simply a Matae argument. Deloitte's own research identifies lack of a unified and accessible data foundation, inability to trust and govern agents, and integration complexity among the biggest barriers to scaling agentic AI.

The risk is that organizations replace one form of fragmentation—spreadsheets, PowerPoints, and disconnected project tools—with another: a collection of intelligent but disconnected agents.

‍

Four Things Disparate Agents Struggle to Provide

‍

1. Controlled Access to People Data

Change management depends heavily on people data: organizational structures, roles, stakeholders, impacts, readiness, engagement, and potentially sensitive employee information.

Giving multiple independently deployed agents access to that information creates an immediate governance question: who can access what, for what purpose, and under whose authority?

A platform approach allows AI to operate within an established environment where permissions, organizational data, and change information can be managed consistently.

The goal isn't to prevent AI from accessing context. It is to give AI the right context, under the right controls.

‍

2. A Genuine Portfolio View

‍

An agent working on Project A may be very good at Project A.

But employees don't experience change one project at a time.

A team might simultaneously be affected by an ERP implementation, restructuring, AI rollout, new operating model, and regulatory program. Understanding that cumulative impact requires information across the portfolio—not simply intelligence within individual projects.

This is where a common platform becomes fundamentally different from a collection of agents.

Matae creates a connected data environment across change initiatives. That allows organizations to move beyond asking:

What's happening on this project?

to asking:

What's happening across our organization?

That enterprise view makes it possible to understand change saturation, competing impacts, sequencing, portfolio risk, and where leadership attention is most needed.

‍

‍

3. Benefits Measurement

‍

Generating an impact assessment or communications plan is useful.

But ultimately, organizations aren't investing millions in transformation to produce better change artifacts. They're investing to achieve business outcomes.

That means change activity needs to connect to measurement and benefits realization.

What were we trying to achieve? What changed? Did people adopt the new way of working? Did the expected benefit materialize? And where it didn't, what should we do differently?

An individual agent can analyze a metric. A connected platform can help maintain the thread between change activity, adoption, performance, and benefits across the transformation portfolio.

That's a much bigger proposition.

‍

4. Overarching Governance

‍

As agents proliferate, governance becomes more important—not less.

Deloitte itself argues that increasing agent adoption creates new governance challenges and that orchestration, proactive management, and planning are essential for secure enterprise deployment.

A change platform provides the operating environment around the AI: common methodologies, structured data, permissions, portfolio visibility, reporting, and governance.

Instead of every team creating its own AI-enabled version of change management, the organization can establish one governed system for how change is understood, delivered, and measured.

‍

‍

So Why Sherpa Rather Than Dozens of Separate Agents?

‍

This is an important distinction in how we've designed Matae.

Sherpa isn't intended to be a single chatbot that performs one task at a time.

It is the AI layer operating within the context of the Matae platform.

That means Sherpa can assist with different change activities while working from the connected context of your change environment. As the capability evolves, AI can work across multiple activities and workflows without forcing the user to manage a growing collection of disconnected agents.

The intelligence matters.

But the context surrounding the intelligence matters just as much.

‍

The Future Isn't Agents Versus Platforms

‍

This isn't an argument against AI agents.

Quite the opposite.

AI agents are likely to become extraordinarily capable, and change teams should embrace the opportunity to automate administrative work that currently consumes valuable capacity. Accenture argues that enterprise AI requires organizations to redesign work and measure outcomes against value, while Deloitte says the organizations that create the most value won't necessarily be those deploying the most agents fastest, but those building the governance and human capabilities to use them effectively.

The strategic question is therefore not:

"How many change management agents can we deploy?"

It is:

"What environment will allow AI to understand and manage change across our organization?"

That's the role Matae is designed to play.

AI to do more of the work. A common data foundation to give it context. Controlled access to people data. Portfolio visibility to see the whole organization. Benefits measurement to prove outcomes. And governance to ensure it all operates as one connected change system.

Because deploying more agents can give you more automation.

Connecting AI across your change portfolio gives you something much more valuable: enterprise intelligence.

‍

‍

‍