Duke Energy
Self-service chatbot in the mobile app
The self-service chatbot in Duke Energy's mobile app logged more than 280,000 interactions from 55,000 unique users in three months and cut manual feedback form submissions by 90%.
Objective
Handle customers' most frequent pain points in the mobile app to deflect requests toward self-service and ease the other channels.
The deployment
Duke Energy launched a chatbot in its customer mobile app in April 2023, built with its Mobile App and Chatbot teams. It answers the main pain points identified through the app's feedback feature, alongside a redesign of the More screen. In its first three months it logged more than 280,000 interactions and 104,000 chat sessions from 55,000 unique users, and cut manual feedback form submissions by 90%. The project received a Gold Digital Experience Award.
Results Proof D
Figures presented by Duke Energy at a conference (a Chartwell writeup quoting the product leads), repeated with the same values by a sector analysis. A real, quantified deployment, but with conference-declared sourcing.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- outil Chatbot in-house Duke Energy NLP chatbot developed by the internal Mobile App and Chatbot teams; no external platform is publicly named.
- infra Application mobile Duke Energy The chatbot's single channel, with access to customer account data and a redesign of the More screen carried out in parallel.
- outil Fonction feedback in-app Source of the pain points that prioritize the cases handled by the bot; manual form submissions dropped by 90%.
How it runs, concretely
For ops teams-
1Detecting pain points data team
The app's feedback feature surfaces the main customer pains.
-
2Answering in the app AI
The chatbot answers common requests and guides toward self-service.
-
3Measured deflection AI
Feedback form submissions drop as answers are resolved upstream.
The pain points surfaced by the app's feedback feature. Without this feedback flow, the chatbot targets the wrong cases and deflection drops.
How your customers perceive this type of use
Sourced studiesLes consommateurs n'acceptent pas les chatbots par defaut : 64% prefereraient que les entreprises n'utilisent pas d'IA dans leur service client (Gartner, 2024) et pres d'un utilisateur sur cinq du service client par IA n'en retire aucun benefice (Qualtrics, 2025). L'acceptation se construit sur trois conditions mesurees par Salesforce : savoir qu'on parle a une IA, pouvoir escalader vers un humain, comprendre la logique de l'agent.
Acceptance conditions
- Etre informe qu'on parle a une IA et non a un humain (pres de 75% le demandent, Salesforce 2024)
- Un chemin d'escalade clair vers un agent humain (45% plus enclins a utiliser l'agent IA, Salesforce 2024)
- Une logique de l'agent clairement expliquee (44% plus enclins, Salesforce 2024)
Red lines
- Rendre l'humain injoignable : c'est la premiere inquietude des consommateurs sur l'IA dans le service client (Gartner 2024) et 50% craignent que l'IA les coupe du contact humain (Qualtrics 2025)
- Remplacer le service client par l'IA sans alternative : 53% envisageraient de partir chez un concurrent (Gartner 2024)
Sources: Salesforce 2024 · Gartner 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- structured surfacing of customer pain points
- access to account data
- a catalog of frequent requests
Org prerequisites
- a mobile product team
- an active feedback loop in the app
Possible stack
- an NLP chatbot integrated into the mobile app
- connectors to the account systems
- a feedback analysis tool
The plan, step by step
- Step 1Analyze in-app feedback and list the five most frequent pain points.Deliverable: A prioritized list of cases to handle, with volumes
- Step 2Design the response journeys and read access to account data.Deliverable: Dialogue scripts and account connectors validated
- Step 3Develop and test the chatbot in the app, with QA on the five priority cases.Deliverable: A working bot in pre-production
- Step 4Roll out gradually to a fraction of the app's users.Deliverable: Chatbot in production with session tracking
- Step 5Measure deflection (forms, contacts) and extend the covered cases.Deliverable: A quantified deflection readout and a backlog of the next cases
First step: Analyze the app's feedback to list the five most frequent pains and answer them with a chatbot in the app.
The same pattern, proven in other industries
ADT
genAI customer agent
Air India
genAI customer agent (automated customer service)
AT&T
Governed enterprise genAI assistant: an LLM connected to internal knowledge, open to tens of thousands of employees and called at scale by automated systems
Sources
- S1 Duke Energy Leverages Mobile App Chatbot to Address Top Customer Pain Points Secondary archive pending
- S2 How AI is transforming customer service and outage management for utilities Secondary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.