Cisco Deploys Personalized AI Agents to All 90,000 Employees

Cisco's rollout of personalized AI agents to its approximately 90,000 employees is an internal enterprise initiative for workforce augmentation through task automation and intelligent model routing.

Executive Summary

Cisco, a major player in the networking sector, is implementing personalized AI agents for its entire workforce of approximately 90,000 employees. This deployment, spearheaded by Chief Financial Officer Mark Patterson, aims to enhance efficiency by providing tools that can manage routine tasks and direct more complex queries appropriately. The initiative begins with the start of the new fiscal year at the end of July 2026. The focus on personalized assistants means that each employee receives a tool tailored to their needs, which can assist in a wide range of daily activities without requiring extensive training or adjustment periods. This represents a significant step in how large corporations are beginning to incorporate AI into their standard operating procedures.

The core of the deployment involves giving each employee a personalized AI assistant. These assistants are capable of performing task automation, responding to employee questions, and routing requests to the AI model best suited for the job. This setup allows for workforce augmentation without the need to rely exclusively on the most advanced models available. Mark Patterson has already utilized his own AI agent to benchmark Cisco's performance against peers on key metrics including revenue growth, earnings per share, research and development expenditures, and how capital is allocated across the business. The system is designed to dynamically select the appropriate model for each use case rather than defaulting to the most advanced options.

A key quantified outcome is that 80 to 90 percent of the first draft for the Management's Discussion and Analysis section is now completed by AI. This represents a significant augmentation of the finance workforce and demonstrates the practical application of agentic AI in enterprise settings. The result shows how internal AI tools can shift employee focus from initial drafting to higher value review activities while maintaining accuracy in financial disclosures.

What background led Cisco to adopt company-wide AI agents?

The background for this rollout involves Cisco's position as a networking giant that is also advancing its own AI infrastructure. The company has chosen to focus on efficiency and maintaining control over data by utilizing on-premises systems. This approach contrasts with some other organizations that might lean more heavily on external frontier models without the same level of internal customization. The decision reflects a strategic choice to test enterprise-wide agentic AI in a controlled environment where both performance and security can be monitored closely.

Mark Patterson has played a central role in driving this effort from the finance perspective. He has personally tested the capabilities of the AI agent by using it to compare Cisco's metrics against those of peer companies. The metrics in question include revenue growth, earnings per share, research and development expenditures, and how capital is allocated across the business. This hands-on involvement allows leadership to identify practical benefits early and adjust the rollout based on real internal usage patterns.

This internal testing provides a foundation for understanding how the technology can be scaled across the organization. By starting with the CFO's office, the company demonstrates a top-down commitment to integrating AI into daily operations. The emphasis remains on practical applications that deliver measurable benefits in areas such as financial analysis and reporting. The broader context includes Cisco positioning itself as a test case for how large enterprises can adopt agentic systems while advancing their own infrastructure capabilities.

What specific capabilities do the personalized AI agents offer to employees?

The personalized AI agents are designed to serve as individual assistants for each of the 90,000 employees. Their primary functions include handling a variety of tasks that would otherwise require manual effort, providing answers to common questions that arise in the course of work, and making decisions about which AI model to engage for a given request. This routing feature is key to maintaining efficiency across different departments and job functions. The agents support workforce augmentation by taking on repetitive elements of daily work.

By routing requests intelligently, the system avoids unnecessary consumption of resources. As described by the leadership, the agent understands which tool is most effective and most efficient for the situation at hand. This prevents the overuse of high-capacity models on tasks that could be handled by simpler ones, leading to better overall resource management. Employees benefit from faster responses and reduced wait times for information or task completion.

The capability extends to supporting employees in their specific roles by tailoring responses and actions to individual needs. Since each assistant is personalized, it can learn from previous interactions to improve future performance. This personalization is intended to make the AI more useful in diverse departments across the company. The result is a more integrated tool that aligns with the varied workflows present in a large organization like Cisco.

How is Cisco configuring its AI infrastructure for this deployment?

Cisco is constructing its AI infrastructure primarily on-premises. This choice allows the company to maintain greater control over costs and data security. Rather than depending on external services for all AI processing, the on-premises setup enables the organization to manage its own environment while still accessing a range of models as needed. The configuration supports the goal of keeping sensitive corporate information within internal systems throughout the rollout.

The strategy involves building proprietary AI stacks that can query various models depending on the particular use case. This method is viewed as the most efficient way to incorporate AI into operations. It provides flexibility without the drawbacks associated with always using the largest and most resource-intensive models. The on-premises focus aligns with broader goals of data protection in a large enterprise.

By keeping much of the processing internal, Cisco can ensure that sensitive information remains within its controlled systems. This configuration supports the rollout to all employees by providing a stable and secure foundation for the AI agents. The approach also positions the company to scale the system as more use cases are identified without increasing external dependencies.

What quantified outcomes have emerged from the AI agent use?

The most prominent quantified outcome relates to the preparation of financial disclosures. According to available reports, 80 to 90 percent of the initial draft work for the Management's Discussion and Analysis is now completed through AI assistance. This represents a substantial shift in how such documents are produced within the finance team. The automation allows team members to dedicate more time to analysis and verification rather than initial compilation.

This level of automation suggests that the AI agents are effectively augmenting the capabilities of the existing workforce. Employees can focus on higher-level review and analysis while the initial compilation and drafting are handled by the system. The result is a more streamlined process that could lead to faster turnaround times for important reports. The outcome provides a concrete example of how agentic AI can impact specific business functions in measurable ways.

How does Cisco's approach differ from other AI strategies in the market?

Cisco's method emphasizes building custom AI stacks on-premises and using dynamic model selection. This differs from strategies that might default to frontier models for every interaction. The focus on efficiency means that the system selects the tool based on what is most appropriate rather than what is most advanced. The distinction helps control operational expenses while still delivering capable assistance to users.

Comparison of AI Deployment Approaches

Aspect	Cisco Approach	Alternative Approach
Infrastructure	On-premises for cost and data control	Cloud-based frontier models
Model Usage	Dynamic routing to efficient models	Default to frontier models
Resource Management	Minimizes token usage on advanced models	Higher token consumption possible
Customization	Build own stacks for specific use cases	Reliance on external providers

The table above illustrates the distinctions in how different organizations might implement AI agents. Cisco's choice reflects a preference for control and optimization tailored to internal needs. This could serve as a benchmark for other companies evaluating their own AI investments. The comparison highlights trade-offs between convenience and oversight that executives must weigh when planning similar initiatives.

What implications does this have for stakeholders in the enterprise sector?

For stakeholders in the enterprise sector, Cisco's rollout offers a case study in scaling AI agents across a large organization. The emphasis on on-premises solutions may appeal to companies concerned with data privacy and operational costs. It shows that AI can be integrated without necessarily increasing dependency on external frontier model providers. Stakeholders can observe how internal leadership involvement accelerates adoption and identifies high-impact use cases early.

Executives in similar positions may look at the quantified improvements in document preparation as an indicator of potential returns on investment. The ability to benchmark performance using the AI tool itself provides an additional layer of value for financial leadership. This dual use for both operational support and analytical tasks demonstrates versatility that can be applied in other sectors.

The initiative also highlights the role of senior leadership in championing AI adoption. When the CFO takes an active part in testing and promoting the technology, it signals organizational priority. Other companies might consider similar executive involvement to accelerate their own deployments. The overall implication is that measured, controlled rollouts can deliver tangible efficiency gains while addressing common enterprise concerns around cost and security.

What reactions and commentary have been offered by Cisco leadership?

Mark Patterson has commented on the broader significance of AI in the current technological landscape. He has described it as one of the most significant transitions seen in a lifetime, with Cisco positioned centrally within it. This perspective underscores the company's view of AI as a core element of its future operations and its readiness to lead in practical applications.

AI is the most significant technology transition that we’ve seen in probably our lifetime, and I think Cisco was right at the heart of it.Mark Patterson, CFO, Cisco

Additional comments from Patterson emphasize the practical aspects of the implementation. He has noted that the system does not waste resources on frontier models when they are not necessary. Instead, it selects the most effective and efficient option for each situation. This approach is seen as central to achieving the desired efficiency gains across the employee base.

It’s not going to burn a whole bunch of tokens with frontier models. It knows which tool is most effective and most efficient.Mark Patterson, CFO, Cisco

Patterson has also explained the rationale for developing internal AI stacks. He states that building their own systems allows the company to query different models based on the specific requirements of each use case. This method is considered the most efficient path forward for integrating AI into the business and supports the goal of controlled, scalable deployment.

We feel like that’s the most efficient way is to build our own AI stacks, which will go out and query the different models based on the particular use case.Mark Patterson, CFO, Cisco

What should peer executives take away from this deployment?

Peer executives can draw several lessons from Cisco's experience with AI agents. The first is the value of starting with internal applications that address specific pain points, such as financial reporting. The second involves the benefits of on-premises infrastructure for maintaining control. The third relates to the use of model routing for cost and efficiency optimization. These elements together provide a framework for evaluating similar projects in other organizations.

1. Begin with executive-level testing to validate use cases before full rollout.
2. Prioritize on-premises solutions when data control is a primary concern.
3. Implement dynamic model selection to avoid unnecessary resource expenditure.
4. Track specific metrics like document preparation time to measure impact.
5. Use AI tools for internal benchmarking to gain additional insights into company performance.

These steps can help other organizations replicate some of the successes observed at Cisco. By focusing on measurable outcomes and maintaining flexibility in model usage, companies can position themselves to benefit from agentic AI without overcommitting to any single approach. The experience at Cisco illustrates that leadership engagement and careful infrastructure choices contribute to successful enterprise-wide implementations.