Verified
Duke Energy
Experimenting- Utility
- 26,441 employees
- GenAI tool for ~10,000 staff (2025)
- 50+ AI products (2025)
- AI saved customers ~$500M (as of 2024)
Duke Energy deploys AI through a generative AI productivity tool it gave to about 10,000 of its 26,441 employees in 2025, more than 50 AI products in production by Aug 2025, AWS-managed AI agents in grid interconnection studies, and named uses such as wildfire risk modeling, scam detection and AI Bill Insights. Duke Energy says AI-driven grid optimization had saved customers about $500 million as of 2024, but has published no AI return to its own costs or revenue.
duke-energy.comTheir approach to AI
- The platform
- Duke Energy gave a personal productivity generative AI tool, which it has not named, to about 10,000 employees in 2025, had more than 50 AI products in production and an internal AI development and governance platform called AI Central by Aug 2025, and builds on AWS, where its engineers were using AWS-managed AI agents in interconnection studies by Sep 2026.
- The models
- Duke Energy has not said which AI models sit behind its employee tool, Duke Energy Explorer, scam detection or AI Bill Insights; a Duke Energy presentation in Aug 2025 named 2 AI assistants, Copilot and Genie, without vendors or models, and AWS has not named the models behind the interconnection study agents.
- The reach
- About 10,000 of Duke Energy's 26,441 employees, roughly 38%, had the generative AI tool in 2025, after AI assistants reached 4,000 employees by Aug 2025; Duke Energy has published no count of employees who use them.
- The catch
- Duke Energy's only recent AI savings figure, about $500 million as of 2024, is stated as a saving to customers, and Duke Energy has published no AI return to its own costs or revenue, while most of its public talk about AI concerns powering data centers.
Snapshot
- Employees
- 26,441
- Dec 31, 2025, full-time, part-time and temporary (2025 Form 10-K)
- Technology budget
- ~3% of capital plan
- Corporate, IT and cyber is about 3% of the $103B 2026 to 2030 capital plan (2025 Impact Report); AI spending is not broken out
- AI/ML headcount
- Not disclosed
- A Duke Energy presentation names its AI team, madlab (Aug 2025), without saying how many people it employs
- Who owns AI
- No named AI chief
- CIO Richard Donaldson focuses on digital and AI; CAO Bonnie Titone oversees IT, cybersecurity and supply chain
- Platform
- Unnamed generative AI tool
- Deployed in 2025; AI Central for AI development and governance (Aug 2025); Duke Energy Explorer; AWS-managed agents for grid studies (Sep 2026)
- Models
- Not disclosed
- Copilot and Genie assistants named in Aug 2025 without vendor or model; no model named for Explorer, Bill Insights or the AWS agents
- Reach
- ~10,000 employees
- Generative AI tool deployed in 2025, after AI assistants for 4,000 by Aug 2025; no active-user count published
- Use cases
- Productivity, grid operations, grid studies, wildfire risk, scam detection, bill insights
- Plus Duke Energy Explorer, relicensing documents and construction tracking
Key findings
Duke Energy deployed a personal productivity generative AI tool to about 10,000 employees in 2025, roughly 38% of the 26,441 people it employed at the end of 2025, said in its Feb 2026 annual report that it was assessing agentic AI tools, and by Sep 17, 2026 had transmission engineers using AWS-managed AI agents in parts of interconnection studies.
Duke Energy added 2 AI uses for customers in 2026: AI that scans websites, social media and paid ads for scams impersonating the utility (Feb 12, 2026), and AI Bill Insights, a plain-language explanation of each monthly bill in the Duke Energy app, for Duke Energy Carolinas residential customers from Jun 29, 2026.
Duke Energy applies AI inside its regulated work: CIO Richard Donaldson said in Sep 2025 that Duke Energy Explorer, a conversational AI tool, answers regulatory questions from hundreds of data sources and that generative AI authors relicensing documents, and CEO Harry Sideris said in May and Aug 2026 that AI tools track construction of Duke Energy's new generation projects.
Duke Energy told EPRI's Open Power AI Consortium in Aug 2025 that it first adopted AI in 2017, had more than 50 AI products in production, had rolled out AI assistants called Copilot and Genie to 4,000 employees, and runs AI Central, a platform for AI development and governance.
Duke Energy has no named AI chief: SVP and CIO Richard Donaldson leads digital strategy with a stated focus on AI, EVP and Chief Administrative Officer Bonnie Titone oversees information technology and cybersecurity, and a committee of the CEO, CFO, CAO, chief legal officer and chief generation officer oversees enterprise AI governance.
Duke Energy's only recent AI savings figure goes to customers: its 2025 Impact Report (Jul 2026) says AI-driven optimization of grid operations, forecasting and outage management had saved customers about $500 million as of 2024, with no period or method, and Duke Energy has published no AI spending figure or AI return to its own costs as of Oct 2026.
Where Duke Energy sits
- Stage
- Experimenting
Duke Energy has reported no business return to its own costs, revenue or headcount that it credits to AI since May 2025. Its 2025 Impact Report (Jul 2026) says AI-driven optimization of grid operations, forecasting and outage management had saved customers approximately $500 million as of 2024, a saving stated as going to customers, with no period or method; its 2025 Form 10-K, 2026 proxy statement and 2026 earnings calls give no AI figure; agentic AI cutting interconnection data preparation from about 30 days to hours (Jul 2026) is a task speedup; and its older figures, more than $74 million of cost cut with computer vision and drones (Jul 2021) and more than $60 million saved by data science from 2016 to 2018 (May 2022), are too old to count. Duke Energy has not reported a qualifying company-wide rollout either: its 2025 Form 10-K says it deployed a generative AI productivity tool to about 10,000 of its 26,441 employees in 2025, and a Duke Energy presentation to EPRI counted AI assistants rolled out to 4,000 employees (Aug 2025), both access counts with no figure for use, and Duke Energy names no executive who owns AI. Duke Energy is therefore at Experimenting, with more than 50 AI products in production by Aug 2025; a return it credits to AI in its own operations, or a count of employees using its AI tools under a named owner, would move it to Piloting.
How the stages are defined - Ownership
- No named AI chief; SVP and CIO Richard Donaldson leads digital strategy with a stated focus on AI; CAO Bonnie Titone oversees IT
- Platform
- Hybrid: Unnamed generative AI tool; AI Central platform; Duke Energy Explorer; AWS-managed agents for grid studies
- Models
- Not disclosed; Copilot and Genie assistants named (2025)
- Hosting
- AWS: primary cloud since 2022 and manager of the grid study agents (2026); Azure for a 2021 methane platform; employee tool hosting not disclosed
- Reach
- About 10,000 employees with a generative AI tool (2025)
How they deploy
AI across the business, described in general terms
Duke Energy's 2025 Form 10-K, filed Feb 26, 2026, says the company uses AI across the organization to improve reliability, optimize grid operations, enhance customer service and accelerate business transformation, alongside digital tools and data analytics. Duke Energy names few specific systems: the filing cites a personal productivity generative AI tool deployed to about 10,000 employees in 2025 and says the company continues to assess and prioritize high-impact investments, including agentic AI tools.
Foundation first, then regulatory and licensing work
Richard Donaldson, Duke Energy's SVP and CIO, said in a Sep 2025 podcast interview that Duke Energy spent 2024 on foundational work, including protecting data and governing its AI platforms, before moving faster. Donaldson described Duke Energy Explorer, a conversational AI tool that draws on hundreds of data sources to answer regulatory questions, and the use of generative AI to author relicensing documentation.
AI agents in grid planning on AWS
Duke Energy entered a multiyear strategic collaboration with Amazon Web Services in Nov 2022, committing to AWS as its primary cloud provider for software development and moving information technology and grid analytics workloads there, including its Intelligent Grid Services planning applications. AWS launched Agentic Grid Planning on AWS on Sep 17, 2026, saying it built the AI agents with Duke Energy engineers, whose transmission planners use them for data preparation, analysis execution and workflow coordination in interconnection studies; Duke Energy and AWS plan to extend the work to large-load planning.
AI in grid operations and wildfire risk
Duke Energy's 2025 Impact Report, published in Jul 2026, says AI is already embedded in grid operations, outage prediction, asset monitoring and system planning, and that Duke Energy uses AI for faster forecasting and more efficient deployment of capital as demand grows. Duke Energy also runs an internally developed AI Wildfire Risk Assessment Model, part of a wider wildfire analytics platform, that ranks high-risk areas from weather, vegetation, drought, infrastructure and population data to guide mitigation and investment.
More than 50 AI products and a central platform
Duke Energy described its AI history to EPRI's Open Power AI Consortium on Aug 21, 2025: a first AI product in production after adopting AI in 2017, an AI team called madlab, more than 50 AI products in production spanning machine learning, computer vision, language processing and generative AI, AI assistants called Copilot and Genie rolled out to 4,000 employees, and AI Central, a platform for AI development and governance. The same presentation showed an enterprise chatbot in pilot and agentic AI designs for inventory management, and named storm response and energy orchestration as targets for AI agents.
Customer AI in small, public steps
Duke Energy has launched its customer-facing AI as discrete, named features. Duke Energy began using AI to find scam websites, social media posts and paid ads impersonating the utility in Feb 2026, and launched AI Bill Insights in its mobile app in Jun 2026 as a first phase that Duke Energy plans to expand with more detailed insights.
AI in the generation build
Duke Energy applies AI to oversight of its largest construction program. Harry Sideris said in May 2026 that Duke Energy's project management and construction team uses AI technologies to track milestones on new generation projects, monitoring construction down to the cubic yard of dirt excavated and concrete poured, and repeated in Aug 2026 that AI tools monitor construction deadlines and progress.
Data science before generative AI
Duke Energy built a company-wide data science program well before generative AI. Duke Energy said in May 2022 that the program had about 30 projects, including models to find faulty meters and energy theft, predict maintenance needs at nuclear plants, recommend products to customers and plan storm restoration, and its data analytics leader said the team was just getting started with AI and machine learning.
AI as a source of demand
Duke Energy's public discussion of AI centers on serving it. Harry Sideris wrote in Jul 2025 that Duke Energy is building new generation to support AI energy demand, cited contracted demand from AI and advanced manufacturing when Duke Energy reported 2025 results in Feb 2026, and Duke Energy had signed 7.8 GW of electric service agreements with data center customers by Aug 4, 2026, under a Customer Protection Plus framework announced in Jul 2026.
- Deployment model
- Duke Energy runs AI through a central technology organization with delivery in the business functions. EVP and Chief Administrative Officer Bonnie Titone oversees information technology, infrastructure, telecommunications, cybersecurity, physical security, supply chain and enterprise strategy, and SVP and CIO Richard Donaldson leads the company's digital strategy with a stated focus on using AI to transform internal operations and modernize customer experiences. Named AI uses sit in regulatory affairs (Duke Energy Explorer), licensing (relicensing documents), transmission planning (AI agents in interconnection studies), grid operations and wildfire risk, customer services (scam detection), customer experience (Bill Insights) and generation construction (milestone tracking). A Duke Energy presentation in Aug 2025 named an AI team, madlab, and AI Central, a platform for AI development and governance, without saying who leads them; Duke Energy's 2025 Impact Report says its AI work is overseen across IT, legal and business teams.
- Model providers
- Duke Energy has not named the vendor or models behind the personal productivity generative AI tool it deployed to about 10,000 employees in 2025, nor the AI behind Duke Energy Explorer, scam detection, AI Bill Insights or construction monitoring. A Duke Energy presentation to EPRI in Aug 2025 named 2 AI assistants, Copilot and Genie, rolled out to 4,000 employees, without naming their vendors or models. AWS manages the AI agents Duke Energy engineers use in interconnection studies, built with Duke Energy and offered to other utilities as Agentic Grid Planning on AWS from Sep 17, 2026, without naming the models behind them. Duke Energy's earlier named AI collaborators date from 2021, when Accenture and its Microsoft joint venture Avanade applied analytics, AI and cloud computing to a methane emissions monitoring platform on Microsoft Azure. Duke Energy executives also sit on the advisory committee of EPRI's Open Power AI Consortium, launched with Nvidia in Mar 2025 to build open AI models for the power sector. Duke Energy's 2025 Form 10-K says it assesses the cybersecurity and data protection risks of third-party AI tools and services before they connect to its network.
- Data and infrastructure
- Duke Energy has used AWS as its primary cloud provider for software development since a Nov 2022 agreement and is migrating information technology and grid analytics workloads to AWS, including the Intelligent Grid Services planning work that needs hundreds of millions of power flow calculations. The AI agents Duke Energy engineers use in interconnection studies are managed by AWS and work with existing simulation software, grid models and engineering standards (Sep 2026). Duke Energy's 2021 methane monitoring platform was built on Microsoft Azure. Duke Energy's 2025 Form 10-K says its information technology systems depend on cloud service providers, but Duke Energy has not said where the models behind its generative AI tool or Duke Energy Explorer run, or whether its own data centers host any AI workloads.
Platforms and internal tools
Generative AI productivity tool2025
Duke Energy deployed a personal productivity generative AI tool to about 10,000 employees across the enterprise in 2025, according to its 2025 Form 10-K, and has not published the tool's name, vendor or usage.
- Users
- About 10,000 employees (2025)
- Models
- Not disclosed
AI CentralBy Aug 2025
AI Central is Duke Energy's platform for enterprise AI development and governance, as a Duke Energy presentation to EPRI's Open Power AI Consortium described it on Aug 21, 2025. Duke Energy has not published its owner, its technology or which of its more than 50 AI products run on it.
AI assistants: Copilot and GenieBy Aug 2025
Duke Energy rolled out 2 AI assistants, Copilot and Genie, to 4,000 employees by Aug 2025, according to its presentation to EPRI's Open Power AI Consortium. The presentation does not name the vendors behind them or say whether either is the productivity tool Duke Energy's 2025 Form 10-K counts at about 10,000 employees.
- Users
- 4,000 employees (Aug 2025)
- Models
- Not disclosed
AI Wildfire Risk Assessment ModelBy Jul 2026
Duke Energy's internally developed AI Wildfire Risk Assessment Model analyzes weather, vegetation, drought conditions, infrastructure exposure and population to rank high-risk areas across its service territory, within a wildfire analytics platform that adds a rolling 7-day risk dashboard and fire behavior simulation.
Duke Energy ExplorerBy Sep 2025
Duke Energy Explorer is a conversational AI tool that Duke Energy built to answer regulatory questions by drawing on data from hundreds of sources, as Richard Donaldson described it in a Sep 2025 podcast interview. Duke Energy has not published its users, vendor or models.
- Models
- Not disclosed
Agentic Grid Planning on AWSSep 17, 2026
Agentic Grid Planning on AWS gives utilities AWS-managed AI agents that run parts of transmission interconnection studies with their existing simulation software, grid models and standards. AWS developed the agents with Duke Energy engineers, who use them for data preparation, analysis execution and workflow coordination, and launched the program to other qualified utilities on Sep 17, 2026.
- Models
- Not disclosed
Intelligent Grid ServicesExpanded on AWS from Nov 2022
Intelligent Grid Services is Duke Energy's suite of custom-built grid-planning applications that forecast where electric demand will grow, estimate how much load the distribution system can accommodate, and recommend the most cost-effective grid upgrades. Duke Energy agreed in Nov 2022 to build new smart grid software on AWS and expand the suite there; in May 2022 Duke Energy's data science team described iGrid, a suite of grid-planning products with the same purpose. Duke Energy's descriptions of the suite do not call it AI.
AI Bill InsightsJun 29, 2026
AI Bill Insights is an AI-powered feature on the Billing tab of the Duke Energy mobile app that gives a personalized, plain-language breakdown of what drives each month's bill, including weather and how usage compares over time. Duke Energy launched AI Bill Insights for Duke Energy Carolinas residential customers on Jun 29, 2026, scheduled Duke Energy Progress customers for mid-Jul 2026, and calls it a first phase.
AI scam detectionFeb 12, 2026
Duke Energy uses AI to scan websites, social media and paid advertisements for social engineering threats such as fake ads and impostor helplines, and flags them for removal with the hosting platforms; one takedown was a paid search ad impersonating Duke Energy with a bogus customer service number.
Methane emissions monitoring platformPilot from Aug 2021
Duke Energy built a cloud platform on Microsoft Azure with Accenture, Avanade and Microsoft to track methane leaks on its natural gas system using satellites, aircraft and ground sensors, with Accenture applying analytics and AI. Duke Energy planned to implement the platform by Oct 2021 and has not published its status since.
Vendors and partners
Amazon Web ServicesNov 2022
Primary cloud for software development; hosts Intelligent Grid Services and grid analytics; manages the AI agents used in interconnection studies · Duke Energy and AWS entered a multiyear strategic collaboration in Nov 2022, and AWS launched Agentic Grid Planning on AWS, developed with Duke Energy engineers, on Sep 17, 2026. Amazon is also a Duke Energy customer: Duke Energy's 2025 Form 10-K cites Amazon's planned $10 billion cloud computing and AI campus in Richmond County, NC.
EPRI Open Power AI ConsortiumMar 2025
Industry consortium building open AI models for the power sector; Duke Energy executives sit on its advisory committee · EPRI launched the consortium with Nvidia on Mar 20, 2025; Nvidia listed Duke Energy among the energy companies on its executive advisory committee.
Microsoft2021
Azure cloud for the methane emissions monitoring platform · Microsoft was one of the parties to Duke Energy's Oct 2026 North Carolina settlement on data center costs, alongside Amazon, Google and Meta.
Accenture and Avanade2021
Designed the methane monitoring platform, applying analytics, AI and cloud computing
Leadership
AI, data and platform leaders
Executives
Board
AI org map
Central technology organization under the Chief Administrative Officer; AI uses delivered in regulatory, grid, customer and generation functions
Duke Energy has no chief AI officer. EVP and Chief Administrative Officer Bonnie Titone, a former CIO, oversees information technology, infrastructure, telecommunications and cybersecurity, and SVP and CIO Richard Donaldson leads digital strategy with a stated focus on AI. Enterprise AI governance sits with the Executive Cybersecurity Oversight Governance Committee of the CEO, CFO, Chief Administrative Officer, Chief Legal Officer and Chief Generation Officer, and the board's Audit Committee receives updates on AI governance.
Duke Energy has not disclosed an AI or machine learning headcount. Duke Energy's technology work sits in the organization of EVP and Chief Administrative Officer Bonnie Titone, which covers information technology, infrastructure, telecommunications and cybersecurity, with SVP and CIO Richard Donaldson leading digital strategy and AI. John Pressley, managing director of digital strategy and engineering, spoke for Duke Energy on the AWS-managed AI agents in Sep 2026. Duke Energy's company-wide data science program ran about 30 projects in May 2022, and a Duke Energy presentation in Aug 2025 named its AI team, madlab, and counted more than 50 AI products in production, without giving the team's size or leader.
Reporting lines
President and CEO
Harry Sideris
Chief Administrative Officer
Bonnie Titone
IT, infrastructure, telecom, cybersecurity, physical security, supply chain, facilities, real estate and enterprise strategy; Duke Energy said in Mar 2024 that the CAO reports to the CEO
Information Technology
Richard Donaldson, SVP and CIO
Digital strategy, technology initiatives and infrastructure operations; the line to the CAO is inferred from Bonnie Titone's remit
Chief Security and Information Security Office
Chief Security and Information Security Officer (not named)
Runs the Enterprise Security Team, which assesses controls on internal AI use; the incident response team sits under the Office of the CAO
Executive Cybersecurity Oversight Governance Committee
CEO, CFO, CAO, Chief Legal Officer and Chief Generation Officer
Oversees enterprise AI governance activities and AI risk within cybersecurity and technology risk (2025 Form 10-K); the chief legal officer is Alex Glenn and the chief generation officer Kelvin Henderson
Finance
Brian Savoy, EVP and CFO
Customer Services Operations
Jessica Bishop, SVP
Runs AI scam detection; the reporting line is inferred, not published
Board Audit Committee
Primary board oversight of cybersecurity and technology, including updates on AI governance; members include Annette Clayton and, from Sep 28, 2026, Joyce Mullen
Duke Energy's 2025 Form 10-K lists no AI, data or analytics executive among its executive officers. The 10-K says the Enterprise Security Team reports cybersecurity status to the CAO, the CIO and the Chief Security and Information Security Officer, but Duke Energy has not published who the CIO reports to or when Richard Donaldson took the role.
Named people
Use cases
Duke Energy counts more than 50 AI products in production (Aug 2025) and has named a smaller set: a generative AI productivity tool for about 10,000 employees, AI assistants for 4,000, AWS-managed AI agents in grid interconnection studies, AI in grid operations and outage prediction, an AI wildfire risk model, Duke Energy Explorer for regulatory questions, generative AI for relicensing documents, AI scam detection, AI Bill Insights and AI tracking of power plant construction. Duke Energy has published no return to its own costs or revenue for any of these use cases; its only recent figure, about $500 million saved for customers by AI-driven grid optimization as of 2024 (Jul 2026), is stated as a customer saving, and its other dollar figures date from 2021 and 2022.
| Use case | Area | Status | ROI |
|---|---|---|---|
| Generative AI productivity tool Duke Energy deployed a personal productivity generative AI tool to about 10,000 employees across the enterprise in 2025. | Enterprise | In production | |
| AI agents in interconnection studies Duke Energy transmission engineers use AWS-managed AI agents for data preparation, analysis execution and workflow coordination in interconnection studies. Duke Energy's 2025 Impact Report (Jul 2026) says the agents cut data preparation from about 30 days to hours, while AWS put the manual work at about 2 weeks (Sep 2026); Duke Energy and AWS plan to extend the agents to large-load planning. | Transmission planning | Pilot | |
| Duke Energy Explorer Duke Energy Explorer is a conversational AI tool that answers regulatory questions by drawing on data from hundreds of sources, as Richard Donaldson described it in Sep 2025. | Regulatory | In production | |
| Generative AI for relicensing documents Duke Energy uses generative AI to author relicensing documentation, Richard Donaldson said in Sep 2025. | Generation licensing | In production | |
| AI scam detection Duke Energy's AI scans websites, social media and paid ads for scams impersonating the utility, such as fake ads and impostor helplines, and flags them for removal, after Duke Energy saw thousands of scam reports in 2025. | Customer services | In production | |
| AI Bill Insights AI Bill Insights in the Duke Energy app explains in plain language what drives each month's bill, including weather, for Duke Energy Carolinas residential customers, with Duke Energy Progress customers scheduled from mid-Jul 2026. | Customer experience | In production | |
| AI construction monitoring Duke Energy's project management and construction team uses AI technologies to track milestones on new generation projects, down to the cubic yard of dirt excavated and concrete poured. | Generation construction | In production | |
| Grid operations, forecasting and outage management Duke Energy says AI is embedded in grid operations, outage prediction, asset monitoring and system planning, without naming the tools or where they run. Duke Energy's 2025 Impact Report (Jul 2026) says this AI-driven optimization had saved customers approximately $500 million as of 2024 through better forecasting, faster outage response and cheaper restoration, a saving stated as going to customers, with no period or method. | Power grid | In production | |
| AI Wildfire Risk Assessment Model Duke Energy's internally developed AI model ranks high-risk wildfire areas from weather, vegetation, drought, infrastructure and population data to guide mitigation and investment decisions, as described in Jul 2026. | Wildfire risk | In production | |
| Inventory management agents Duke Energy showed EPRI in Aug 2025 a design in which an orchestrator agent routes questions to inventory, contracts, forecasts and lead-time agents that predict shortages and draft purchase orders pending approval; the presentation does not say whether it runs in production. | Supply chain | Pilot | |
| Intelligent Grid Services planning Duke Energy's custom grid-planning applications on AWS forecast where demand will grow, estimate how much load the distribution system can take and recommend cost-effective upgrades, as last described in Nov 2022; Duke Energy's descriptions of the tools do not call them AI. | Power grid | In production | |
| Faulty meter and energy theft detection Duke Energy's data science team built models for a revenue protection team that finds faulty meters and energy theft, as described in May 2022. | Customer and revenue protection | In production | |
| Methane leak monitoring Duke Energy's Azure-based platform, built with Accenture, Avanade and Microsoft, combines satellite, aircraft and ground sensor data to find methane leaks; Duke Energy has not reported its status since 2021. | Natural gas | Pilot |
Adoption
Duke Energy first adopted AI in 2017, built a company-wide data science program by 2022, moved software development and grid analytics onto AWS from Nov 2022, spent 2024 on data protection and AI platform governance, rolled out AI assistants called Copilot and Genie to 4,000 employees by Aug 2025, and deployed a personal productivity generative AI tool to about 10,000 employees in 2025, roughly 38% of the 26,441 people it employed at the end of 2025. Duke Energy has not published active-user counts, usage frequency or plans to give the tool to the rest of its workforce, and these are access counts, not counts of employees who use AI. Specialist AI reached regulatory staff through Duke Energy Explorer by Sep 2025 and transmission planners through AWS-managed agents by Sep 2026, and customer-facing AI reached Duke Energy Carolinas residential app users through AI Bill Insights on Jun 29, 2026.
By the numbers
Rollout, first pilot to today
2017
Duke Energy first adopted AI and put its first AI product into production, according to its Aug 2025 presentation to EPRI.
Jul 2021
CIO Bonnie Titone described computer vision and drones that inspect Duke Energy's solar farms and said the program had cut costs by more than $74 million and 385,000 labor hours.
Aug 2021
Duke Energy said it was building a methane emissions monitoring platform on Microsoft Azure with Accenture and Microsoft, with Accenture and Avanade applying analytics and AI, and began a satellite pilot in Greenville, SC.
May 2022
Duke Energy described a company-wide data science program with about 30 projects, including models for faulty meters and energy theft, nuclear maintenance and storm response, and a set of grid-planning tools called iGrid.
Nov 2022
Duke Energy and AWS began a multiyear strategic collaboration that made AWS Duke Energy's primary cloud provider for software development and moved its Intelligent Grid Services and grid analytics workloads to AWS.
2024
Duke Energy spent the year on foundational work for AI, including protecting data and governing its AI platforms, according to Richard Donaldson.
2025
Duke Energy deployed a personal productivity generative AI tool to about 10,000 employees across the enterprise.
Aug 2025
Duke Energy told EPRI's Open Power AI Consortium it had more than 50 AI products in production, AI assistants called Copilot and Genie rolled out to 4,000 employees, and AI Central, a platform for AI development and governance.
Sep 2025
Richard Donaldson described Duke Energy Explorer, a conversational AI tool for regulatory questions, and the use of generative AI to author relicensing documentation.
Feb 2026
Duke Energy began using AI to detect scams that target its customers, first announced company-wide on Feb 12 and for Ohio and Kentucky on Feb 26, 2026.
May 2026
Harry Sideris said Duke Energy uses AI technologies to track construction milestones on its new generation projects.
Jun 2026
Duke Energy launched AI Bill Insights in its mobile app for Duke Energy Carolinas residential customers, with Duke Energy Progress customers scheduled to follow in mid-Jul 2026.
Jul 2026
Duke Energy's 2025 Impact Report said AI is embedded in grid operations, outage prediction, asset monitoring and system planning, and described an internally developed AI Wildfire Risk Assessment Model.
Sep 2026
AWS launched Agentic Grid Planning on AWS, built with Duke Energy engineers, who use its AI agents in parts of transmission interconnection studies.
Economics & controls
Spend and ROI
Technology budget
Duke Energy does not publish a technology operating budget or an AI budget. Duke Energy's capital plan is about $103 billion for 2026 to 2030, mostly for generation and grid infrastructure, of which corporate, IT and cyber make up about 3%, according to its 2025 Impact Report (Jul 2026), and its 2025 Form 10-K warns that keeping pace with changing technologies, including AI, involves significant development and implementation costs.
What AI itself costs
Duke Energy has not disclosed what it spends on AI tools, licenses, cloud computing for AI or AI staff in its 2025 Form 10-K, its 2026 proxy statement or its 2026 earnings calls.
Reported ROI
Duke Energy has published no AI return to its own costs, revenue or headcount as of Oct 2026, either firm-wide or for a named use case. Duke Energy's 2025 Impact Report (Jul 2026) says that, as of 2024, AI-driven optimization of grid operations, forecasting and outage management had saved customers approximately $500 million, crediting better forecasting, faster outage response and shorter, cheaper restoration; the report states the figure as a saving to customers and gives no period, method or share of Duke Energy's own costs. Duke Energy's older figures are too old to count: Bonnie Titone said in Jul 2021 that computer vision and drones for solar farm inspection had cut more than $74 million of cost and 385,000 labor hours, and Duke Energy credited data science projects in May 2022 with more than $60 million saved on revenue protection from 2016 to 2018 and about $10 million a year on running energy-efficiency programs. The interconnection study agents cut data preparation from about 30 days to hours (Jul 2026), a task speedup with no cost or capacity figure attached.
Productivity and cost markers
Duke Energy said its cost control initiatives kept operations and maintenance costs flat despite inflation (Feb 2026) and that it ranked third among electric industry peers for non-generation operations and maintenance cost per customer in 2025 (Aug 2026). Duke Energy did not credit AI for either result. Duke Energy also reports that smart, self-healing grid technology avoided about 2.2 million customer outages in 2025, but does not describe that technology as AI.
Headcount
Duke Energy had 26,441 employees at Dec 31, 2025, including 5,027 represented by unions. Duke Energy's 2025 Form 10-K and 2026 earnings calls do not tie AI to headcount, hiring or job cuts.
Governance and risk
AI governance inside the cybersecurity framework
Duke Energy folds AI into its cybersecurity risk management program: the framework covers the risks of the company's development and use of AI, the Enterprise Security Team assesses controls on internal AI use in line with Duke Energy's internal AI governance policies, and the Executive Cybersecurity Oversight Governance Committee of the CEO, CFO, CAO, Chief Legal Officer and Chief Generation Officer oversees enterprise AI governance activities (2025 Form 10-K).
A governed data foundation and cross-functional oversight
Duke Energy's 2025 Impact Report (Jul 2026) says its AI approach rests on a secure, governed data foundation meant to make AI safe and reliable for critical operations, with cross-functional oversight built across IT, legal and business teams. Duke Energy described AI Central to EPRI in Aug 2025 as its platform for AI development and governance.
Foundation before rollout
Richard Donaldson said in Sep 2025 that Duke Energy spent 2024 on foundational work, including protecting data and governing its AI platforms, before speeding up its AI work.
Engineers decide, agents assist
John Pressley said in Sep 2026 that Duke Energy's engineers lead interconnection study work and make all final engineering decisions while AWS-managed AI agents handle data preparation, analysis execution and workflow coordination.
Third-party AI tools
Duke Energy requires any third party that handles its data, hosts an application or connects to its network to pass a cybersecurity assessment, and as part of that process evaluates the cybersecurity and data protection risks of third-party AI tools and services (2025 Form 10-K).
Board oversight
Duke Energy's Audit Committee has primary board oversight of cybersecurity and technology and receives periodic updates on digital transformation, AI governance practices and emerging AI-related technology and cybersecurity risks. Duke Energy's board considers AI risks within its enterprise risk framework, lists AI among the technology skills it looks for in directors, and added Joyce Mullen, a former Insight Enterprises CEO, to the Audit Committee in Sep 2026.
Disclosed risks
Duke Energy's 2025 Form 10-K warns that AI, including generative AI, can be used to carry out cyberattacks on its systems, that its own use of generative AI and use by its vendors and agents may create data privacy, legal, regulatory and security risks, and that failing to implement critical technology, including AI, could hurt results.
Shareholder engagement
Duke Energy's 2026 proxy statement lists the board's risk oversight role, including the use of artificial intelligence, among the topics on the agenda of its shareholder engagement meetings, alongside contingency plans for data centers coming onto its grid.
Recent signals
Sep 25, 2026
Duke Energy appointed Joyce Mullen, former CEO of Insight Enterprises, to its board and Audit Committee from Sep 28, citing her experience at the intersection of AI, digital innovation and customer-focused execution.
Sep 17, 2026
AWS launched Agentic Grid Planning on AWS, AI agents for interconnection studies built with Duke Energy engineers, who use them for data preparation, analysis and workflow coordination; the 2 plan to extend the agents to large-load planning.
Aug 4, 2026
Harry Sideris said Duke Energy is using AI tools to monitor construction deadlines and progress on its generation build, and Duke Energy reported 7.8 GW of electric service agreements with data center customers.
Jul 2026
Duke Energy's 2025 Impact Report said AI-driven optimization of grid operations, forecasting and outage management had saved customers approximately $500 million as of 2024, and that agentic AI built with AWS cut interconnection data preparation from about 30 days to hours.
Jun 29, 2026
Duke Energy launched AI Bill Insights in its mobile app for Duke Energy Carolinas residential customers, with Duke Energy Progress customers to follow in mid-Jul 2026.
May 5, 2026
Harry Sideris said Duke Energy uses AI technologies to track construction milestones down to the cubic yard of dirt excavated and concrete poured, and, citing growth driven by AI and advanced manufacturing, that data center agreements had reached about 7.6 GW.
Mar 20, 2026
Duke Energy's 2026 proxy statement said the board considers AI risks within enterprise risk oversight and listed the board's risk oversight, including the use of AI, among the topics of its shareholder engagement meetings.
Feb 26, 2026
Duke Energy's 2025 Form 10-K reported a generative AI productivity tool deployed to about 10,000 employees in 2025, said agentic AI tools are under assessment, and described AI governance inside its cybersecurity program.
Feb 26, 2026
Duke Energy Ohio and Kentucky announced they are using AI to detect scams that target their customers.
Feb 12, 2026
Duke Energy said it is using AI to scan websites, social media and paid ads for scams impersonating the utility and to flag them for removal.
Feb 10, 2026
Harry Sideris cited contracted demand from AI and advanced manufacturing as Duke Energy reported 2025 results and extended its 5% to 7% earnings growth target through 2030.
Quick answers
- How does Duke Energy deploy AI?
- Duke Energy deploys AI through a generative AI productivity tool it gave to about 10,000 of its 26,441 employees in 2025, more than 50 AI products in production by Aug 2025, AWS-managed AI agents in grid interconnection studies, and named uses such as wildfire risk modeling, scam detection and AI Bill Insights. Duke Energy says AI-driven grid optimization had saved customers about $500 million as of 2024, but has published no AI return to its own costs or revenue.
- What does Duke Energy use AI for?
- Duke Energy counts more than 50 AI products in production (Aug 2025) and has named a smaller set: a generative AI productivity tool for about 10,000 employees, AI assistants for 4,000, AWS-managed AI agents in grid interconnection studies, AI in grid operations and outage prediction, an AI wildfire risk model, Duke Energy Explorer for regulatory questions, generative AI for relicensing documents, AI scam detection, AI Bill Insights and AI tracking of power plant construction. Duke Energy has published no return to its own costs or revenue for any of these use cases; its only recent figure, about $500 million saved for customers by AI-driven grid optimization as of 2024 (Jul 2026), is stated as a customer saving, and its other dollar figures date from 2021 and 2022.
- Who leads AI at Duke Energy?
- AI at Duke Energy is led by Bonnie Titone (Executive Vice President and Chief Administrative Officer); Richard Donaldson (Senior Vice President and Chief Information Officer). Ownership of AI on the org chart: No named AI chief; SVP and CIO Richard Donaldson leads digital strategy with a stated focus on AI; CAO Bonnie Titone oversees IT.
- What AI platform does Duke Energy use?
- Duke Energy gave a personal productivity generative AI tool, which it has not named, to about 10,000 employees in 2025, had more than 50 AI products in production and an internal AI development and governance platform called AI Central by Aug 2025, and builds on AWS, where its engineers were using AWS-managed AI agents in interconnection studies by Sep 2026.
- Which AI models does Duke Energy use?
- Duke Energy has not said which AI models sit behind its employee tool, Duke Energy Explorer, scam detection or AI Bill Insights; a Duke Energy presentation in Aug 2025 named 2 AI assistants, Copilot and Genie, without vendors or models, and AWS has not named the models behind the interconnection study agents.
- How many Duke Energy employees use AI?
- About 10,000 of Duke Energy's 26,441 employees, roughly 38%, had the generative AI tool in 2025, after AI assistants reached 4,000 employees by Aug 2025; Duke Energy has published no count of employees who use them.
- What return has Duke Energy reported from AI?
- Duke Energy has reported no business return to its own costs, revenue or headcount that it credits to AI since May 2025. Its 2025 Impact Report (Jul 2026) says AI-driven optimization of grid operations, forecasting and outage management had saved customers approximately $500 million as of 2024, a saving stated as going to customers, with no period or method; its 2025 Form 10-K, 2026 proxy statement and 2026 earnings calls give no AI figure; agentic AI cutting interconnection data preparation from about 30 days to hours (Jul 2026) is a task speedup; and its older figures, more than $74 million of cost cut with computer vision and drones (Jul 2021) and more than $60 million saved by data science from 2016 to 2018 (May 2022), are too old to count. Duke Energy has not reported a qualifying company-wide rollout either: its 2025 Form 10-K says it deployed a generative AI productivity tool to about 10,000 of its 26,441 employees in 2025, and a Duke Energy presentation to EPRI counted AI assistants rolled out to 4,000 employees (Aug 2025), both access counts with no figure for use, and Duke Energy names no executive who owns AI. Duke Energy is therefore at Experimenting, with more than 50 AI products in production by Aug 2025; a return it credits to AI in its own operations, or a count of employees using its AI tools under a named owner, would move it to Piloting.
Open questions
- Which generative AI tool and which models did Duke Energy deploy to about 10,000 employees in 2025, is it one of the Copilot and Genie assistants rolled out to 4,000 employees by Aug 2025, how many employees use it, where do its models run, and will it reach the rest of the 26,441-person workforce?
- How much of the approximately $500 million that AI-driven grid optimization saved customers as of 2024 was a reduction in Duke Energy's own restoration and operating costs, over what period, and measured how?
- Who owns AI at Duke Energy: who leads the madlab AI team and the AI Central platform, and does Richard Donaldson report to Bonnie Titone?
- What does Duke Energy spend on AI within the roughly 3% of its 2026 to 2030 capital plan set aside for corporate, IT and cyber, and does any of it appear in the rate cases it files with state regulators?
- What return do Duke Energy's deployed AI uses deliver to its own costs or revenue, such as scam takedowns, call volumes avoided by AI Bill Insights, planner capacity gained from the interconnection study agents, wildfire mitigation spending targeted by the risk model, or schedule and cost effects of AI construction monitoring?
- Will Duke Energy give an AI figure on its third-quarter 2026 earnings call, scheduled for Nov 5, 2026?
- How much has generative AI saved on relicensing documents? An AI-assisted transcript of Richard Donaldson's Sep 2025 interview quotes 200,000 hours saved on nuclear relicensing documents, a figure Duke Energy has not published itself.
- Which other agentic AI tools is Duke Energy developing beyond interconnection studies, and which models run behind the AWS-managed agents?
- Which AI tools does Duke Energy use to improve reliability and optimize grid operations, as its 2025 Form 10-K states without naming them?
- What became of Duke Energy's data science program, the iGrid tools and the 2021 methane monitoring platform, and do they now use machine learning or generative AI?
- Does Duke Energy run any AI workloads in its own data centers, and does it use Microsoft Azure or another cloud alongside AWS for AI today?
Sources
Duke Energy
- Corporate leadership
- Bonnie Titone, EVP and Chief Administrative Officer, biography
- Richard Donaldson, SVP and Chief Information Officer, biography
- Harry Sideris, President and CEO, biography
- Brian Savoy, EVP and CFO, biography
- Sasha Weintraub, EVP and Chief Customer Officer, biography
- Duke Energy uses AI to protect customers from scams (Feb 12, 2026)
- Duke Energy Ohio and Kentucky use AI to protect customers from scams (Feb 26, 2026)
- New this summer: an easier way to understand your energy bill, AI Bill Insights (Jun 29, 2026)
- Duke Energy board appoints Joyce Mullen (Sep 25, 2026)
- Duke Energy protects customers from data center costs (Oct 7, 2026)
- Data center growth will deliver customer savings, Customer Protection Plus (Jul 23, 2026)
- Duke Energy appoints Harry Sideris president and CEO (Jan 13, 2025)
- Duke Energy appoints new chief information officer, Bonnie Titone (May 23, 2019)
- Duke Energy collaborates with AWS on smart grid solutions (Nov 17, 2022)
- Duke Energy teams with Accenture and Microsoft on a methane monitoring platform (Aug 23, 2021)
- illumination: Duke Energy's data science team and the grid (May 9, 2022)
- illumination: Harry Sideris on AI and energy demand (Jul 23, 2025)
Filings and investor materials
- Duke Energy 2025 Form 10-K (filed Feb 26, 2026)
- Duke Energy 2026 proxy statement (Mar 20, 2026)
- Duke Energy 2025 Impact Report (Jul 2026)
- Form 8-K, appointment of Joyce Mullen to the board (Sep 25, 2026)
- Form 8-K exhibit, executive leadership appointments (Mar 15, 2024)
- Q4 and full-year 2025 earnings release (Feb 10, 2026)
- Q4 2025 earnings call transcript, The Motley Fool (Feb 10, 2026)
- Q1 2026 earnings call transcript, The Motley Fool (May 5, 2026)
- Q2 2026 earnings call transcript, The Motley Fool (Aug 4, 2026)
Partners
- Amazon: AWS launches Agentic Grid Planning program to accelerate interconnection studies (Sep 17, 2026)
- AWS: Agentic Grid Planning on AWS
- Nvidia: EPRI, Nvidia and collaborators launch Open Power AI Consortium (Mar 20, 2025)
- EPRI: Open Power AI Consortium committee meeting deck, including AI at Duke Energy (Aug 21, 2025)
News and interviews
- T&D World: AWS, Duke Energy use AI agents to accelerate interconnection studies (Sep 18, 2026)
- VentureBeat: Duke Energy used computer vision and robots to cut costs by $74M (Jul 18, 2021)
- Enterprise AI Innovators: When AI meets infrastructure with Duke Energy CIO Richard Donaldson (Sep 10, 2025)
- CEOInterviews.AI: AI-assisted transcript excerpt of the Richard Donaldson interview (Sep 10, 2025)
- The Conference Board: Richard Donaldson biography
Data confidence
- Confirmed
- Stated in a primary or reputable public source.
- Estimated
- Derived or approximated from public reporting.
- Unconfirmed
- Reported, but not yet corroborated.
Duke Energy profile · Enterprise AI diffusion report by Imagine · facts checked against their sources on Nov 1, 2026