UK-based Colt Technology Services has recently concluded a proof of concept for an agentic AI pricing engine jointly developed with Microsoft.
This innovative tool significantly cuts down the time needed to produce complex deal pricing, reducing what used to take several days into about ten minutes. It was trained across most of Colt’s markets within three days and boasts 99% accuracy. Plans are in place for broader implementation later this year.
Addressing a common hurdle in enterprise infrastructure procurement, the pricing engine aims to streamline lengthy and opaque global project pricing, particularly when dealing with multiple markets. While the AI autonomously generates quotes, Colt’s teams ensure these are reviewed before being presented to customers.
Technical Implementation and Deployment Context
The agentic AI engine leverages Microsoft’s cloud-based AI technology integrated with Colt’s telecommunications expertise. This solution is part of the company’s overall strategy, which emphasizes secure, scalable, and responsible AI ecosystems for both employees and customers.
Frank Miller, Chief AI and Platforms Officer at Colt Technology Services, highlighted that large-scale infrastructure projects can be extremely time-consuming to price accurately. The objective was to harness AI to break this cycle, enabling customers to concentrate on their objectives while simplifying the procurement process.
Rick Lievano, Worldwide CTO for Telco, Media and Gaming at Microsoft, noted that agentic AI has significant potential to revolutionize complex enterprise workflows in telecommunications, where speed, accuracy, and scale are crucial. He also emphasized that the Colt initiative aims to automate intricate processes and facilitate faster access to decision-making information for customers.
Wider Application
The pricing agent is viewed as one of the first proof-of-concepts within Colt’s broader agentic AI program. The company is investigating the use of agentic AI throughout various stages of the customer journey, including from initial pricing to onboarding, with human review still in place for certain steps to ensure responsible deployment.
