Copilots are now coworkers: What happens next with AI

What happens after the thrill of AI experimentation fades? What comes after the pilot projects, prompt libraries, Copilot rollouts and enthusiastic early adopters?

That’s when the real work begins: proving whether, at the enterprise level, AI and agentic AI can reliably do meaningful work, orchestrate business processes and fundamentally change how your organization operates.

Agents are here to stay, but can you trust them?

AI agents are moving quickly from emerging technology to enterprise reality. More enterprise applications are becoming agent-enabled, and a growing share of everyday decisions will soon be supported — or made — by agentic AI.

In other words, agents are becoming part of the enterprise operating fabric. They help your employees complete work, and increasingly perform work on their behalf, across productivity tools, enterprise data, business applications, workflows and customer-facing processes.

That changes what you need to discuss in the boardroom. Just deploying AI tools isn’t the challenge anymore. You need to know how to make AI agents as dependable, secure and productive as your best employees.

Like people, agents need enterprise guardrails. They need identities, defined access, governance, policies, observability, accountability and lifecycle management.

That makes the next phase of AI transformation not only about intelligence but about a combination of intelligence and trust.

Evolving from assistants to digital coworkers

The first generation of AI helped people work faster. The next generation gets work done.

Today's AI agents can reason across applications, connect information from multiple sources, complete multistep tasks and keep working without constant human intervention.

Microsoft’s own ecosystem illustrates this evolution. Microsoft Copilot Chat is ideal for quick, focused tasks, while Microsoft 365 Copilot brings secure AI into everyday work. Microsoft Copilot Cowork extends that capability by coordinating complex, multistep work across Microsoft 365, while Microsoft Scout introduces a more proactive model that tracks progress, nudges users and advances work behind the scenes.

This represents far more than incremental productivity gains. Looking beyond time saved per user, the opportunity is to redesign work itself: from sales preparation and customer follow-up to knowledge discovery, project coordination, service operations, finance analysis, HR support and procurement workflows.

Instead of just helping your employees complete these activities faster, AI can increasingly become an active participant in delivering the outcome. But with that power comes a new responsibility.

An agentic workforce needs a new trust model

When AI just summarizes information, the risks are relatively contained. Once AI agents begin taking action, those risks change dramatically. An agent might retrieve sensitive information, trigger a workflow, draft customer communications, update business systems or operate across multiple identity boundaries.

As we’ve explored in Marionettes and ghosts: Security considerations when AI agents join your workplace, the critical distinction is whether an agent operates as a marionette — borrowing a person’s identity and permissions — or as a ghost, with its own enterprise-managed identity and purpose-built access. The difference matters because it determines how your organization manages accountability, auditing, governance and risk.

That leaves CIOs, CISOs and other business leaders with important questions to answer. Who owns the agent? What data can it access? What actions can it take? How do we monitor what it has done? How do we govern its lifecycle? How do we stop or modify its behavior when outcomes drift?

The answer isn’t to build an entirely new governance model. To manage these agents, extend the trusted infrastructure that’s already used to manage your users.

Agents should therefore become part of the same enterprise control fabric that already governs those users. That means bringing them under the same identity, security and compliance framework using, among others, Microsoft Entra for identity and access, Microsoft Defender for threat protection, Microsoft Purview for governance and compliance, and the Microsoft 365 admin center for management, with Microsoft Agent 365 acting as the control plane for observing, governing and securing agents.

This is where trust becomes a business enabler. When you apply the right controls, you can move faster, making security, identity, governance and observability the foundation that will allow innovation to scale.

What should you do now?

There’s a balance to strike. On the one hand, you don’t want so much control of AI that innovation is stifled. On the other hand, you also don’t want too little control, which can result in shadow agents, unmanaged data exposure, unclear ownership and fragmented business outcomes.

The goal is disciplined acceleration:

1. Assess your AI and agent readiness. Look beyond technology. Evaluate your licensing estate, business opportunities, data readiness, Copilot readiness, security posture, shadow agents, operating model, skills and organizational readiness for change.
2. Prioritize use cases that deliver measurable value. Not every process needs an agent. Focus on workflows with repetitive effort, fragmented information, high coordination costs, decision bottlenecks or significant business impact.
3. Build a secure foundation. Strengthen identity and access controls, apply least-privilege principles, implement conditional access and sensitivity labels, protect enterprise data, establish agent identities and registries, define onboarding processes, create governance structures and embed responsible AI guardrails from the outset.
4. Scale through an agent factory. Create a repeatable process for discovering, designing, building, testing, publishing, monitoring and continuously improving agents across your business. Technologies such as Microsoft Copilot Studio, Microsoft Foundry and Microsoft Graph, combined with line-of-business integrations, telemetry, model evaluation and AgentOps, become key building blocks.
5. Treat AI as an ongoing transformation. Managing an agentic workforce requires continuous governance, portfolio management, ROI measurement, policy refinement, adoption programs and regular optimization.

Get there faster with NTT DATA

Having the right partner can significantly shorten the path from experimentation to enterprise value. Our structured Workplace Frontier Transformation with Microsoft helps you move beyond ambition to deliver measurable outcomes.

1. We start with Frontier Immersion & Value Discovery, helping you understand the evolution from Copilot as an assistant to agents as coworkers — and ultimately toward a human-led, agent-operated enterprise.
2. Next comes the Frontier Readiness Assessment, where we evaluate your organizational maturity, identify high-value use cases, assess your data and security readiness, and develop practical 30-, 60- and 90-day roadmaps.
3. From there, our Frontier Foundation Accelerator establishes the secure landing zone, governance model, tenant readiness, identity controls, Microsoft Purview and Defender policies, Agent 365 controls, adoption playbooks and pilot operating model.
4. Our Frontier Scale Accelerator then focuses on building and deploying agents across business functions, integrating enterprise platforms such as SAP, ServiceNow, Salesforce and Jira while establishing an agent portfolio supported by ROI dashboards.
5. Finally, Embrace and Lead Frontier provides the long-term operating model you need to govern, optimize and scale your enterprise agentic workforce.

The value of this approach is moving forward with purpose and control. We help you build a structured path that assesses your readiness, establishes the right foundations, scales high-value agents, supports adoption and embeds governance throughout.

Only by redesigning how you work and building trust into your operating model will you turn AI into an intelligence multiplier that amplifies human expertise, contextual intelligence and enterprise execution.

• ALSO READ: Future of Work 2030: Making AI work in the workplace

WHAT TO DO NEXT

Read more about NTT DATA’s Workplace Frontier Transformation with Microsoft to see how we help you design, build and scale AI, from ideas to impact.