Becoming a Frontier Firm: Our IT playbook for the AI era

Microsoft Digital, the company’s IT team, is rapidly transforming into a Frontier IT Firm—an organization fundamentally restructured for the AI era, where AI-agents are digital colleagues rather than peripheral tools.

“Agents are the most significant technological change we’ve seen since the shift to the cloud, and arguably since the advent of the Internet. They represent a generational opportunity to rethink our entire portfolio of IT services. But that innovation requires a careful and thoughtful strategy to maximize impact and minimize time to value.”

Brian Fielder, vice president, Microsoft Digital

This emerging model emphasizes hybrid human-agent teams, dynamic organizational structures, and a culture of continuous innovation. Employees evolve into “agent bosses,” orchestrating outcomes by delegating tasks to autonomous agents that plan, reason, and adapt.

IT teams are essential to enabling this transformation, from shaping and governing the data that informs AI, to deploying and managing the systems that power agentic workflows, to driving the culture and learning programs that are essential to unlocking a frontier mindset.

“Agents are the most significant technological change we’ve seen since the shift to the cloud, and arguably since the advent of the Internet,” says Brian Fielder, vice president of Microsoft Digital. “They represent a generational opportunity to rethink our entire portfolio of IT services. But that innovation requires a careful and thoughtful strategy to maximize impact and minimize time to value.”

This guide is our blueprint for other IT organizations to follow, so you can ensure your own team is on the pathway to becoming a Frontier Firm.

Transforming business and IT with AI agents

The integration of agents in the enterprise presents transformative opportunities for every organization, but becoming a Frontier Firm requires a clear understanding of the journey ahead. IT leaders must recognize three patterns that mark the evolution toward frontier status. Each pattern reflects increasing organizational, technical, and cultural maturity, unlocking greater value from agents:

1. Human with assistant: In this foundational pattern, employees leverage intelligent AI assistants—such as Microsoft 365 Copilot—to enhance their productivity and effectiveness. This is the starting point where individuals develop an “AI habit,” reimagining legacy workflows and discovering new efficiencies with the support of AI.
2. Human-led agents: The second pattern introduces autonomous agents as “digital workers” who perform specific tasks under human direction. Here, agents join teams and execute actions independently, allowing employees to delegate routine work and focus on higher-value activities.
3. Human-led, agent operated: In the final pattern, the synergy between humans and agents is fully realized. Human-led, agent operated teams set strategic direction, while agents autonomously run entire business processes and workflows. Agents periodically report their progress to support scalable and resilient operations, while human oversight is maintained through evaluation and audits to verify accuracy and quality of outcomes.

Becoming a Frontier Firm

AI maturity starts at simple AI assistance, then progresses to more complex patterns between humans and agents.

As organizations experience these patterns, agents evolve from basic information retrieval to sophisticated, autonomous systems capable of executing complex, end-to-end workflows. These mature agents drive transformation through self-directed operations, continuous adaptation, and improvement powered by data and feedback. Multiagent systems further enable scalable, personalized services, automate intricate tasks, and optimize resource allocation.

Microsoft categorizes agents into three tiers, reflecting the different uses we’ve observed across the enterprise:

• Personal Agents: Designed for individual users, these agents automate personal tasks like meeting summaries, email drafting, and task reminders. They operate within a single user’s scope and permissions. Personal agents can be enabled at scale with no friction due to low risk.
• Team (Business) Agents Designed for departmental or team workflows that are built in Copilot Studio or with Copilot Studio Lite. They handle tasks like tracking project status, routing expense approvals, and answering FAQs within a shared team context. Teams can share these low-risk agents, each limited to the data the team requires, while operating in secure, policy-governed environments.
• Enterprise Agents: Operating at an organizational level, these agents integrate with authoritative systems to deliver large-scale services such as compliance checks, enterprise search, and policy enforcement. They are centrally governed, leverage enterprise data platforms and differ from personal and team agents due to their scale, complexity and impact.

Within Microsoft Digital, our strategic investments across all three classifications have already generated meaningful business benefits. The following sections detail how these investments are accelerating our path to becoming a Frontier IT organization—and how they can propel your organization too.

The evolving role of enterprise IT

Our journey to becoming a Frontier Firm started with AI agents, which are the next major step in our evolution of IT. Agents provide us with a foundation that we’re using to reimagine our IT services with potential to lower our costs, improve our outcomes, and accelerate our innovation.

IT plays an essential role for businesses that aspire to become Frontier Firms. In Microsoft Digital, we’ve identified eight core services that IT teams must provide to enable organizations to become Frontier Firms.

1. Driving strategic business objectives

To succeed in this transition to the frontier, we’ve found that enterprise IT must become a highly valued consultative partner to our internal business leaders by helping them to understand where agents can drive strategic goals and then jointly developing those solutions.

For example, if a strategic goal is to expand into a new market without a large support staff, our IT team might propose deploying multilingual customer service agents to handle inquiries in that market. This kind of proactive solutioning demonstrates how our team is thinking beyond just keeping systems running—we’re contributing to growth and innovation. In effect, our role shifts from technology enablement towards business strategy execution via technology architecting.

2. Enabling agentic capabilities

To accelerate the adoption and impact of agents, we must take a proactive role in building the foundational platforms, tools, and infrastructure that empower our larger organization. This means developing systems specifically designed for agent memory and orchestrating seamless integration between agents and core enterprise systems. IT teams like ours should also curate and expose certified data sources, ensuring agents have reliable information to draw upon.

The IT function must evolve from managing technology to offering AI capabilities as a strategic service. For instance, for us establishing an “agent development sandbox” is enabling our business groups to safely prototype and test agents using secure access to scoped corporate data—empowering innovation while maintaining necessary oversight and governance. We also need to define and shepherd processes so agents can safely “graduate” from personal or team-based tenants to becoming available enterprise-wide.

By leading this charge, we ensure that our business units see us as a trusted partner rather than a gatekeeper. Without this support, organizations may try to implement agents independently, resulting in unmanaged “shadow IT” and increased organizational risk. Providing intuitive AI platforms, clear guidance, support structures and effective training encourages collaboration and positions us as the enabler of safe, scalable agentic transformation.  

3. Managing AI-enabled enterprise architecture

The rise of agents is transforming enterprise architecture into a more dynamic, adaptive, and intelligent system. Traditional systems were largely deterministic, with predictable inputs and outputs. In contrast, agentic systems introduce probabilistic behaviors, autonomous decision-making, and continuous learning, which require new architectural thinking and governance models. Our IT architects must evolve their design principles to address this shift:

• Ensuring reliability in AI-driven systems: AI components, including agents, operate with varying degrees of confidence and uncertainty. We in IT must implement mechanisms such as confidence thresholds, fallback logic, and human-in-the-loop controls to ensure that agentic decisions are reliable and aligned with business goals.
• Auditing and accountability of agents: Agents must be discoverable, observable and auditable. Our organization must design systems that log agent decisions, track their reasoning paths, and enable post-hoc analysis. This supports compliance, risk management, and continuous improvement of agent behavior.
• Integrating agents into service-oriented architectures: Agents should be treated as first-class citizens within the enterprise architecture. This means defining clear APIs, service contracts, and interaction patterns that allow agents to interoperate with existing services, data platforms, and business processes.
• Determine your agent creation enablement strategy: It was important to clarify which employees are empowered to build, and how and when they engage with us in IT throughout the process. Start by allowing employees to create agents in self-serve environments for personal productivity. As agents mature, establish clear checkpoints for them to engage with you, such as review and publishing for compliance, security, and governance before agents are scaled to broader use. This staged approach ensures your employees can assist and find value quickly, while you provide oversight and support to maximize impact and maintain organizational standards.
• Agent governance platform: As organizations deploy multiple agents across domains—customer service, finance, operations, legal, HR, etc.—we are finding there is a growing need for centralized agent governance. This layer is not just a technical construct; it is a strategic capability that we must design, implement, and operate. Its responsibilities include managing how agents interact with each other and with human users, policy definition and enforcement, monitoring and observability of agent activity, agent lifecycle management, security and access controls, and agent interoperability and integration.

By owning this layer, we in IT have become the steward of our agentic governance, ensuring that autonomous systems remain aligned with enterprise objectives. This also positions us as a strategic enabler of innovation, allowing business units to deploy agents confidently within a controlled and scalable framework.

4. Driving data and integration strategy

For agents to operate effectively, they require reliable access to accurate, comprehensive data and seamless connectivity across enterprise systems. This places us at the heart of enabling agent-driven workflows. We proactively identify and catalog data sources our agents need, we break down data silos, and build “AI Ready” data access layers—such as APIs, data lakes, productivity data enablement with strong hygiene that are easily discoverable, well managed and inventoried for use by agents and other AI-powered solutions. By leading these efforts, we are strengthening our strategic role, with data architecture and governance becoming foundational to AI success.

Additionally, we must oversee the integration of agents with legacy systems to ensure they can securely initiate transactions within core systems. Ultimately, we in IT move from simply maintaining operations to actively designing and supporting intelligent, future-ready workflows that position the enterprise for ongoing innovation and event driven and real time systems.

5. Stewardship of governance and ethical AI

Robust governance is foundational to successful agent adoption. We in IT must partner closely with risk and compliance teams to lead the implementation of a comprehensive AI governance framework. We find that agreeing and aligning to the right levels of governance are the key to enabling AI successfully. To get there, you need the following:

• Senior leaders who are responsible for driving enterprise agentic prioritization and initiatives across disciplines and organizations to realize business value.
• A cross-discipline team to provide technical and strategic guidance serving to help accelerate, learn and land delivery of AI initiatives to drive adoption, impact and build an “AI forward” culture in IT.
• Human stewards augmented by AI who enable AI transformation through enterprise grade data strategy, governance, and architecture.
• Agentic collaboration and productivity through strong data hygiene standards to help you make sure only the right data is available to agentic systems.
• Adherence to global compliance regimes and privacy frameworks, as well as responsible and ethical applications of AI, like our Microsoft Responsible AI principles.

By setting clear standards for transparency, fairness, and security, we are driving innovation and ensuring alignment with our company’s values and regulatory responsibilities.

Our stewardship extends across the entire lifecycle of agents, including versioning their models, updating their knowledge bases, and retiring or replacing agents as needed. It also includes mirroring software version controls for intelligent, learning systems.

While this increases the complexity of what we manage, it expands our strategic value. With a holistic approach, we can break down silos and build a unified, effective ecosystem of agents that work collaboratively across the enterprise.

6. Training and education

Our role is evolving to provide high quality and authoritative education and training content for the larger company. By providing a centralized repository of training on agent development, we are maintaining our relevance with our business partners. We are providing a visible way to gather input, suggestions, and maintain insight to the use cases that our employees and business partners require—this is enabling us to provide feedback on current solutions and to identify data and APIs required for us to deliver desired results.

As agents become integral to enterprise workflows, we must take a proactive role in cataloguing and delivering high-quality, authoritative training and educational resources to our business partners. Establishing a centralized hub for agent development training not only strengthens our partnership with our business stakeholders but it also ensures that the technical solutions we build are informed and validated by our colleague’s real-world needs.

IT teams should implement visible mechanisms—like feedback channels and suggestion platforms—to collect input, track emerging use cases, and maintain ongoing insight into evolving business requirements. This continuous dialogue enables us to refine current solutions and identify the critical data sources and APIs necessary to support future agent-driven initiatives, fostering a culture of innovation and responsiveness across our organization.

7. Change management and adoption

Agents mark a transformational shift in human-computer interaction. This profound change requires us in IT to play a central role in supporting our employees as they adapt and rethink their daily work with agents. Effective change management is therefore a critical responsibility for us—it’s enabling us to move beyond technology deployment to actively guiding our employees, business units, and leaders through this transition to agent-driven workflows.

We must develop and execute comprehensive strategies that include clear communication plans, targeted training and enablement, and robust feedback mechanisms to identify challenges and accelerate adoption. By embedding change management into our core mandate, we ensure that agent-driven innovation is not only implemented but also embraced across the organization—driving tangible business value, resilience, and positioning us as a strategic partner in enterprise transformation.

8. Measuring agent impact and outcomes

Measuring the success of the journey to becoming an AI frontier firm requires that we take a comprehensive approach to evaluating business outcomes for the company. Enterprises that aspire to become Frontier Firms must establish a unified dashboard and a recurring cadence to review key performance indicators (KPIs) that capture the full spectrum of adoption, impact, maturity, and responsibility. These metrics are providing us with clear and actionable insights across several critical domains. Examples include:

• Usage and adoption: We’re tracking the percentage of our employees leveraging AI-agents and the number of AI use cases deployed in production.
• Efficiency gains: We’re monitoring hours saved, automation rates, and cost reductions achieved through AI-driven processes.
• Business performance: We’re evaluating improvements in operational metrics, output quality, and the speed of decision-making cycles.
• Employee and customer satisfaction: We’re assessing our employees’ survey results regarding AI adoption and we’re monitoring customer satisfaction scores.
• AI maturity and coverage: We’re gauging the maturity level of our AI capabilities and the extent of AI projects that our employees have implemented across all departments.
• Responsible AI and governance: We’re tracking the percentage of models audited, monitoring bias and fairness metrics, and aiming for zero compliance incidents.
• Cross-functional engagement: We’re measuring our executive sponsorship of our AI initiatives and counting the number of joint projects spanning multiple business units.

“A thoughtful, structured approach is key to accelerating your journey to the frontier. Making the time at the start of the journey to ensure you have the right processes and frameworks in place as well as an empowered leadership team to make quick decisions will ensure your IT organization is poised to accelerate your company’s frontier journey.”

Monika Gupta, partner group engineering manager, Microsoft Digital

By defining clear outcomes and tracking progress against targeted metrics, we’re objectively demonstrating where agents are creating business value and identifying areas where we can improve. This approach is reinforcing that agent adoption is a strategic priority, ensuring that our victories are recognized and that we’re proactively addressing our risks, which is driving responsible innovation and lasting impact throughout our enterprise.

Enabling the agentic revolution at Microsoft

In the era of agents, we and IT organizations like yours have an opportunity to be a strategic cornerstone for companies aspiring to become Frontier Firms. By managing the foundational infrastructure, data, and governance frameworks that support agents, the IT function is pivotal in enabling agent-driven digital transformation. The evolution of IT—from a traditional service provider to an orchestrator of AI capabilities, a champion of responsible technology practices, and a true partner in business innovation—will determine whether we and organizations like yours can fully realize the benefits of agents.

“A thoughtful, structured approach is key to accelerating your journey to the frontier,” says Monika Gupta, a partner group engineering manager in Microsoft Digital. “Making the time at the start of the journey to ensure you have the right processes and frameworks in place as well as an empowered leadership team to make quick decisions will ensure your IT organization is poised to accelerate your company’s frontier journey.”

IT teams that proactively embrace this shift will enhance their strategic value, guiding their organizations through the complexities of intelligent automation and positioning them for competitive advantage. Conversely, passive IT risk being sidelined, which can lead to fragmented user experiences and increased security vulnerabilities.

The charge for CIOs is clear: Take the lead in adopting agents and redefine IT’s role at the core of your business strategy. This is a generational opportunity for IT to move beyond routine operations to instead become a driver of lasting innovation and impact.