Key Outcomes
Improved SLA compliance, reducing the risk of regulatory penalties
More consistent handovers between triage and case-handling teams
Enhanced reporting and governance through structured summaries and classifications
Get in touch
Book a consultation with one of our experts.
The Challenge
Manual Email Triage Slowing Customer Responses
UK Power Networks faced a growing challenge: keeping pace with thousands of customer requests for new connections, which arrive by email. The manual effort required to open, read, and triage each message before replying, was placing pressure on teams and risking delays in response.
Manual Email Routing
Frontline agents were responsible for reading and manually routing each email to the correct internal team. This process was time-consuming and often led to delays in addressing high-priority cases.
Operational Strain
During peak periods, teams were stretched thin, and agents spent significant time interpreting vague or misrouted emails instead of resolving customer issues.
Data Structure Issues
The unstructured nature of email data meant information had to be manually re-entered into downstream systems. This made it difficult to monitor trends, identify recurring problems, or generate meaningful reports.
Rising Customer Expectations
As customers increasingly expect fast, transparent service, the manual processes in place struggled to meet these demands, highlighting the need for a more efficient and consistent approach.
“By orchestrating the triage of our high-volume inboxes with energy-sector specific language, the Customer Agent has helped free up frontline teams, improve response times, and ensure key details are captured consistently from the start."
Martin Knoebel, Head of Digital Solutions and Analytics, UK Power Networks
The Solution
Introducing an AI Customer Agent Built to Drive Efficiency
UK Power Networks worked with CKDelta to develop Customer Agent – an enterprise-ready AI agent designed to augment human intelligence to improve efficiency.
Rather than replace human agents, the goal was to remove repetitive, low-value work so that teams could focus on complex or urgent cases. The Agent was trained on real historical email data and refined through iterative validation with UK Power Networks’ customer service experts.
Built using Large Language Models within the customer’s Databricks environment and integrated with existing systems such as Salesforce, the Customer Agent:
Reads incoming customer emails, identifies key information and classifies the message type with 98% accuracy.
Extracts and summarises content from email threads to support quicker triage and more accurate case creation.
Seamlessly hands off cases into Salesforce, adapting to existing workflows and cross-functional processes.
Integrates with Outlook and updates the folder structure and metadata to align with team workflows.
The Business Impact and Strategic Value
The introduction of the Customer Agent has significantly enhanced how UK Power Networks manages the receipt of Connections applications from customers, delivering measurable improvements in efficiency, accuracy, and service quality.
By automating the summarisation and classification of customer emails, the AI Agent has helped frontline teams focus on resolution rather than triage.
The AI Agent has also brought structure and standardisation to customer message handling. Summaries now follow a consistent format, and classification logic is applied uniformly, making it easier to track performance, identify recurring issues, and ensure accountability across the customer service journey. This consistency has strengthened internal governance and improved collaboration between triage and case-handling teams.
Following its early success, UK Power Networks is scaling the AI Agent further and integrating it directly with Salesforce to automate case creation. This expansion marks a strategic shift toward a more insight-driven, responsive service model—where automation handles the routine, and people focus on what matters most.
“In terms of AI, the deployment of CKDelta’s AI agent has demonstrated how Agentic AI technologies can be responsibly deployed in regulated environments with human oversight and without disrupting existing workflows.”
Martin Knoebel, Head of Digital Solutions and Analytics, UK Power Networks
Tech Stack
