Organization size:
1,000-9,999 employees
Partner type
SI
Industries
IT Services
Business function
Information Technology
Building an AI-first enterprise at scale
Atea is one of the largest IT infrastructure and services companies in northern Europe, serving public and private sector organizations across the Nordic and Baltic regions. With more than 8,000 employees in seven countries, the organization provides hardware, software, and consulting services that help customers modernize their IT environments. Headquartered in Oslo, Norway, Atea operates at the intersection of enterprise IT and digital innovation.
In 2025, Atea set a bold strategic direction: to make AI central to how every employee works. The organization adopted Microsoft 365 Copilot as its foundation, licensing it for more than half of its workforce. But Atea’s ambition went beyond individual productivity. The company wanted to build a comprehensive agentic AI framework that could also serve as a blueprint for its thousands of customers, using Microsoft Copilot Studio. Kate is being deployed through Microsoft 365 Copilot and Microsoft Teams, giving employees access to the agent where they already work.
Christoffer Besler Hansen, Head of AI Workplace at Atea, leads the initiative. “We have an ambition to make Atea the best Copilot users in the world,” he explains. “We are both a customer and a partner, so everything we build internally should also make sense for our customers.” The strategy focuses on three pillars: ensuring every employee uses Copilot Chat, driving daily agent usage among licensed users, and establishing an agentic AI framework as a Customer Zero initiative.
“We have an ambition to make Atea the best Copilot users in the world. We are both a customer and a partner, so everything we build internally should also make sense for our customers.”
Christoffer Besler Hansen, Head of AI Workplace, Atea
Scattered information slows teams
Despite its strong technology portfolio, Atea’s internal landscape presented familiar challenges. Customer information was scattered across multiple systems. Teams across countries used different CRM tools per country, while service delivery teams tracked customer cases in ServiceNow. Contract details, sales pipelines, and project histories were stored in separate repositories.
Preparing for customer meetings often required employees to switch between applications, manually compile information, and fill knowledge gaps. “Customer information lived across many systems,” Hansen says. “We needed something that could integrate with what we already have and bring it together.” Consultants preparing tenders faced similar friction, pulling CVs from one database, pricing information from another, and customer history from a third.
At the same time, employees were independently creating personal agents and automations. Atea needed a structured approach that could channel this enthusiasm into secure, governed solutions that benefited the wider organization.
A multi-agent orchestrator and an enterprise agent marketplace
Atea’s response was twofold: build a flagship multi-agent system to address the most pressing productivity gaps and create an enterprise-wide platform for agent discovery and governance.
Kate: A multi-agent orchestrator for sales and customer intelligence
The centerpiece of Atea’s agentic strategy is Kate—short for Key Account Team Executer—a multi-agent orchestrator built in Copilot Studio. Rather than responding directly to requests, Kate routes them to one of more than 35 specialized sub-agents responsible for areas such as CRM data, contract details, sales processes, tender preparation, and customer history.
Kate connects to the organization’s CRM systems through Model Context Protocol (MCP) servers, enabling the agent to read from and write to those systems. It also integrates with ServiceNow, an ERP system, and a CV database used for tender responses. For account managers, Kate provides a single conversational interface for information previously spread across multiple tools. Before meeting a customer, users can ask Kate for a briefing that combines CRM opportunity data, contract details, support history, and previous engagements into a single response. Tasks that once required searching several systems can now be completed in minutes.
Atea also uses native Dynamics 365 Sales agents in countries where that platform is in place. In countries that use different CRM systems, Atea built MCP servers that connect Kate into the organization’s CRM environment. The result is a side-by-side approach: Dynamics 365 agents support teams where Dynamics 365 is used, while Kate orchestrates across the systems each country actually relies on.
“I work with large, complex customers involving multiple contracts and several solution specialists. To stay on top of the current landscape, Kate helps me identify opportunities, develop strong customer solutions, and follow up with the right stakeholders at the right time.”
Mette Juelsen, Key Account Manager, Atea
The solution began with a simple request. “One of our account managers, who is not technical, asked if we could create an agent for his customers,” Hansen recalls. “That was the start of Kate. We just kept adding more and more.” What began as a single customer-knowledge agent has grown into a fully orchestrated multi-agent solution with dedicated sub-agents for sales, tenders, and customer contracts.
For more advanced use cases, Atea extends Copilot Studio with Azure AI Foundry. The organization’s Bid Board Agent, built in Foundry with five distinct AI models, automatically evaluates incoming tenders. Each model plays a different role in assessing whether Atea should invest resources in a response—saving the bid review team hours of preliminary analysis.
The architecture also demonstrates the flexibility of Microsoft’s agent ecosystem. As Hansen explains, “Even though we don’t have Microsoft services in all places, we are still able to leverage Copilot into it.” By connecting Copilot Studio to third-party systems through MCP servers, Atea extends intelligent orchestration beyond the Microsoft ecosystem.
The Atea Agent Hotel: A governed marketplace for enterprise AI
As employees started creating their own agents, Atea saw both momentum and a new organizational challenge: good agents, prompts, skills, and self-made apps were becoming harder to discover and reuse across teams. To bring structure to this growing ecosystem, Atea built the Agent Hotel—an internal portal where employees can discover, publish, and request agents. Users sign in with Microsoft Entra ID, and the portal shows relevant agents based on country, role, and security group membership. It also hosts a prompt library, use cases shared by colleagues, a skills marketplace, and a place to share self-made apps and other reusable AI assets.
Governance is central to the model. Employees can freely create and share agents within small teams, while agents that handle sensitive data or support business-critical processes go through a review process before broader distribution. Atea uses lifecycle management to ensure business-critical agents meet security, compliance, and continuity standards before they reach the organization-wide marketplace.
Faster responses, better data, and a culture of agent builders
For Atea’s account managers, Kate has streamlined customer engagement. Before a meeting, users can ask Kate for a consolidated briefing—pulling together CRM system data, contract history, and recent service interactions in seconds. During follow-ups, Kate can update the organization’s CRM systems directly with meeting notes and action items, eliminating manual data entry.
Mette Juelsen, Key Account Manager at Atea, says, “I work with large, complex customers involving multiple contracts and several solution specialists. To stay on top of the current landscape, Kate helps me identify opportunities, develop strong customer solutions, and follow up with the right stakeholders at the right time.”
Hansen describes the shift in practical terms: account managers “are able to reply more quickly to the customers” and do so “with the correct information.” Some have uncovered contract details or customer information they were previously unaware of, simply because they asked Kate before sending an email. The result is faster response times and higher-quality customer interactions.
The initiative has also improved data quality. Because Kate relies on well-structured customer data, account managers were motivated to clean and update their records before onboarding Kate. “It helps people get things in order,” Hansen notes. “We are getting a better data structure on our customers after we started using Kate.”
Across the organization, employees self-report saving approximately 50 minutes per day using Copilot and agents—time redirected toward higher-value customer work. “They save about 50 minutes per day using Copilot and agents,” Hansen confirms. “And they’re also seeing that what they are delivering is improving in quality.” With more than 10,000 agents now created across Atea, the agentic culture has taken root far beyond the AI team. Non-technical employees are building their own agents, dashboards, and applications using tools such as GitHub Copilot and Copilot Studio.
“They save about 50 minutes per day using Copilot and agents. And they’re also seeing that what they are delivering is improving in quality.”
Expanding Kate and AI adoption
Kate is currently in a pilot with approximately 100 active users, and Atea plans to expand access to teams across Norway and up to 1,750 employees by the end of the year. Ultimately, every sales professional will have access to Kate as a daily tool for customer intelligence and engagement.
The multi-agent architecture is already inspiring new systems across the business. Teams in managed services have begun building their own agent networks that connect to Kate, enabling sales teams to query available services and customer configurations through a single interface. HR is exploring employee self-service agents, and additional MCP integrations are planned for new data sources.
Hansen also sees the orchestration layer evolving. “Kate will not be the orchestrator—Copilot Chat will be,” he says. “We’ll have all the sub-agents ready to work that way.” Meanwhile, the Agent Hotel is expanding beyond agents to include AI-generated applications, custom dashboards, and shared skills, positioning it as a central hub for Atea’s AI-powered workforce.
“Kate will not be the orchestrator—Copilot Chat will be. We’ll have all the sub-agents ready to work that way.”