Who leads AI at Shell?

AI at Shell is led by Robbert van Rutten (Executive Vice President Information and Digital Technology and Chief Information Officer); Amy Challen (Global Head of AI); Gerben de Jong (CIO, Integrated Gas, Upstream and Projects & Technology (Jun 2025)). Ownership of AI on the org chart: Global Head of AI Amy Challen; CIO Robbert van Rutten reports to the CEO.

What AI platform does Shell use?

Shell runs its industrial AI on the Shell.ai Platform, built on C3 AI on Microsoft Azure since Sep 2018 with Databricks for data and extended under a multi-year C3 AI deal in Jun 2026; Shell also built Shell e, a platform on which teams build generative AI applications.

Which AI models does Shell use?

Shell builds its own machine learning models for equipment monitoring, says it uses GPT-4 through Microsoft's Azure OpenAI Service and Microsoft Copilot, and fine-tuned a chemistry language model with NVIDIA NeMo; Shell e runs open-source and other large language models.

How many Shell employees use AI?

Shell's AI community had more than 12,000 members in Jan 2025, against about 96,000 staff at the end of 2024; Shell has published no count of employees who use Microsoft Copilot or Shell e.

How far has AI spread at Shell?

Shell has reported no business return it credits to AI, for any use case or for the company: none of its $5.1 billion of structural cost cuts from 2022 to 2025, or the $700 million more in the first half of 2026, is credited to AI; CFO Sinead Gorman named AI as one of several levers for the $5 billion to $7 billion cost target and CEO Wael Sawan has put no figure on its part; and the hundreds of millions of dollars of value put on predictive maintenance in Jun 2026 is a claim by vendor C3 AI, not by Shell. Shell has not reported a company-wide AI rollout with a user count either: it says it uses GPT-4 through Azure OpenAI, Microsoft Copilot and its own Shell e platform without saying how many of its 85,000 employees use them, and the AI community of more than 12,000 members is a membership figure from Bernard Marr (Jan 2025), not a count of tool users. That leaves Shell at Experimenting although predictive maintenance has run in production since 2018 and covers more than 13,000 pieces of equipment; one cost, revenue or output figure Shell credits to AI, or a firmwide user count with a named owner such as Global Head of AI Amy Challen, would move it to Piloting.

Verified Nov 1, 2026

Shell

Experimenting

Oil and gas
85,000 employees
C3 AI on Azure since 2018
13,000+ pieces of equipment monitored
No AI return or user count

Shell deploys AI mainly in its operations: since 2018 it has run predictive maintenance on the C3 AI platform on Microsoft Azure, which covered more than 13,000 pieces of equipment when Shell extended the deal in Jun 2026 and is adding AI agents. For office and research work, Shell says it uses GPT-4 through Azure OpenAI, Microsoft Copilot and Shell e, a generative AI platform it built. Shell has published no usage count and no financial return from AI.

shell.com

Their approach to AI

The platform: Shell runs its industrial AI on the Shell.ai Platform, built on C3 AI on Microsoft Azure since Sep 2018 with Databricks for data and extended under a multi-year C3 AI deal in Jun 2026; Shell also built Shell e, a platform on which teams build generative AI applications.
The models: Shell builds its own machine learning models for equipment monitoring, says it uses GPT-4 through Microsoft's Azure OpenAI Service and Microsoft Copilot, and fine-tuned a chemistry language model with NVIDIA NeMo; Shell e runs open-source and other large language models.
The reach: Shell's AI community had more than 12,000 members in Jan 2025, against about 96,000 staff at the end of 2024; Shell has published no count of employees who use Microsoft Copilot or Shell e.
The catch: Shell reports no financial return from AI: none of its $5.1 billion of structural cost cuts by the end of 2025 is credited to AI, and the only value figure for predictive maintenance comes from vendor C3 AI.

Snapshot

Employees: 85,000; Dec 31, 2025 (2025 Annual Report); 96,000 at the end of 2024
Technology budget: Not disclosed; R&D spend was $1.17B in 2025; Shell publishes no IT or AI budget
AI/ML headcount: ~600est.; Mathematics, computational and data science staff supporting 100+ AI applications (Databricks case study, undated)
Who owns AI: Amy Challen, Global Head of AI; CIO Robbert van Rutten reports to the CEO (Dec 2025); Challen's reporting line not published
Platform: Shell.ai Platform and Shell e; C3 AI on Microsoft Azure since 2018 with Databricks for data; Shell e for generative AI applications
Models: In-house ML, GPT-4, Copilot, custom LLM; GPT-4 via Azure OpenAI and Microsoft Copilot (Shell website, undated); 10,000+ production ML models in Mar 2022
Reach: 12,000+ in the AI community; Jan 2025; no generative AI usage figures published
Use cases: 100+ AI applicationsest.; In development and deployment, per Shell's website (undated)

Key findings

Shell has run industrial AI on the C3 AI platform on Microsoft Azure since Sep 2018, and when Shell extended the deal under a multi-year agreement on Jun 4, 2026, C3 AI said the predictive maintenance program monitored more than 13,000 pieces of equipment and would add agent-based root cause analysis.
Shell says it uses GPT-4 through Microsoft's Azure OpenAI Service and Microsoft Copilot and has built Shell e, a platform on which teams build applications with open-source and other large language models, but Shell has published no count of employees using any of them.
Shell's AI community had more than 12,000 members in Jan 2025, several thousand of them building AI, and Shell sorts AI applications into green, amber and red risk zones that set how much oversight each one gets, according to Bernard Marr.
Shell reports no financial return from AI: CFO Sinead Gorman listed AI among the levers for a $5 billion to $7 billion structural cost reduction target by 2028 in Mar 2025, and CEO Wael Sawan said on Jul 30, 2026 that Shell was halfway through that range and was asking how AI could add more value, without tying a figure to AI.
Robbert van Rutten is the first CIO in Shell's history to report directly to the CEO, according to a Dec 2025 interview, while Dan Jeavons, who led Shell's core AI program for 13 years, left Shell in 2025 to become President of Applied Computing.

Where Shell sits

Stage: Experimenting
Shell has reported no business return it credits to AI, for any use case or for the company: none of its $5.1 billion of structural cost cuts from 2022 to 2025, or the $700 million more in the first half of 2026, is credited to AI; CFO Sinead Gorman named AI as one of several levers for the $5 billion to $7 billion cost target and CEO Wael Sawan has put no figure on its part; and the hundreds of millions of dollars of value put on predictive maintenance in Jun 2026 is a claim by vendor C3 AI, not by Shell. Shell has not reported a company-wide AI rollout with a user count either: it says it uses GPT-4 through Azure OpenAI, Microsoft Copilot and its own Shell e platform without saying how many of its 85,000 employees use them, and the AI community of more than 12,000 members is a membership figure from Bernard Marr (Jan 2025), not a count of tool users. That leaves Shell at Experimenting although predictive maintenance has run in production since 2018 and covers more than 13,000 pieces of equipment; one cost, revenue or output figure Shell credits to AI, or a firmwide user count with a named owner such as Global Head of AI Amy Challen, would move it to Piloting.
How the stages are defined
Ownership: Global Head of AI Amy Challen; CIO Robbert van Rutten reports to the CEO
Platform: Hybrid: Shell.ai Platform on bought C3 AI and Databricks; Shell e built for generative AI
Models: In-house ML models; GPT-4 via Azure OpenAI; Microsoft Copilot; a NeMo-tuned chemistry LLM
Hosting: Microsoft Azure for C3 AI and Azure OpenAI; Databricks on AWS and Azure; SLB subsurface tools on AWS; Shell's own HPC cluster in Bangalore; NVIDIA edge devices at sites
Reach: AI community of 12,000+ members (Jan 2025); Copilot users not disclosed

How they deploy

Buy the platform, build the models

Shell considered building its own AI platform but chose C3 AI on Microsoft Azure in Sep 2018 because of the work needed to maintain one, Dan Jeavons said in Mar 2022; Shell's own data scientists build and run the models on it. Shell CIO Robbert van Rutten said in Dec 2025 that Shell now follows a rule of market standards unless there is a reason to build, and Shell's TechXplorer Digest said in Jan 2026 that Shell is a digital company but not a software company, which is why its inner source team open-sourced its AI code discovery tool.

Operations first

Shell's AI is concentrated in its plants and fields: predictive maintenance on more than 13,000 pieces of equipment, real-time production optimization at LNG plants, and machine vision that reads gauges, spots hazards and checks protective equipment. The World Economic Forum named Shell's Pearl GTL plant in Qatar to its Global Lighthouse Network in Sep 2025 for its use of smart technologies and AI, including machine vision for hazards and AI equipment monitoring.

Partner where Shell lacks the data

Amy Challen said in Nov 2023 that Shell focuses its own AI work on use cases where it holds strong internal data and intellectual property, and relies on partners, foundation models and open-source models for the rest. Shell works with C3 AI and Microsoft on its core platform and generative AI, NVIDIA on simulation, vision and language models, Databricks on data, SLB on subsurface software and Yokogawa on plant robots.

AI as a cost lever

Shell CFO Sinead Gorman told investors on Mar 25, 2025 that the company was excited by technology and AI opportunities to help it reach a raised target of $5 billion to $7 billion in cumulative structural cost cuts by the end of 2028 against 2022. Wael Sawan said on Feb 5, 2026 that further performance improvements would be supported by AI, and on Jul 30, 2026 that Shell was halfway through the range, was pushing for the top end and was asking how AI could add more value. Shell cut $5.1 billion of structural costs by the end of 2025 and $700 million more in the first half of 2026 without attributing any of it to AI.

Moving to agents

Shell's Jun 2026 extension with C3 AI adds AI agents that diagnose the root cause of equipment problems and suggest remedies, beyond spotting anomalies. Shell's Dec 2025 collaboration with SLB aims to build agentic AI tools for subsurface, well construction and production experts on SLB's Lumi platform, and Shell CIO Robbert van Rutten named agentic AI and low-code tools that let employees build their own solutions as the shifts he watches most (Dec 2025).

Deployment model: Shell runs AI through a central platform and AI team with delivery in the businesses. Shell.ai is Shell's internal change programme for a common data science platform, shared ways of working and a community of practice. Shell's Information and Digital Technology organisation, whose CIO Robbert van Rutten reports directly to CEO Wael Sawan (Dec 2025), has four functions working from Houston, The Hague, London and Bangalore: Enterprise IT, Business IT, Information and Digital Engineering, which runs capability centres including Information, Data & Analytics and citizen development, and Information and Digital Services and Operations. Shell lets employees build their own solutions through citizen development within risk zones, and in early 2026 it moved the technical divisions of Projects & Technology into its Integrated Gas, Upstream and Downstream, Renewables and Energy Solutions businesses.
Model providers: Shell's production AI is mostly its own machine learning: more than 10,000 production models detected equipment anomalies on the C3 AI platform in Mar 2022, with three to four candidate models trained for each. For generative AI, Shell's undated AI web page says Shell uses GPT-4 through Microsoft's Azure OpenAI Service and Microsoft Copilot, and that Shell e, a generative AI platform Shell developed, lets teams build applications on open-source and other large language models. Shell fine-tuned a custom chemistry language model with NVIDIA NeMo (the base model is not named), runs inference for its Visual Remote Sensing program on NVIDIA Triton, and tested open-weight Gemma and Qwen models in an archive search pilot described in Jan 2026. Shell has not said how widely Copilot or Shell e is used.
Data and infrastructure: Shell's predictive maintenance runs on C3 AI on Microsoft Azure, fed by the Shell Sensor Intelligence Platform, which is built on Delta Lake and in Mar 2022 took in 20 billion rows of data a week from more than 3 million data streams; Shell open-sourced the platform's ingestion code as RTDIP through LF Energy in 2022. Shell's GPT-4 use runs through Microsoft's Azure OpenAI Service. Databricks is part of the Shell.ai Platform on AWS and Azure, and in Oct 2023 Shell signed a three-way agreement with SLB and AWS to run SLB subsurface workflows on AWS using the OSDU data platform standard. Shell uses an in-house high-performance computing cluster in Bangalore, cooled with its own immersion fluids, for large-scale AI computing for R&D hubs in Amsterdam, Bangalore, Houston and London (Sep 2025), and runs vision models on NVIDIA Jetson Orin devices at sites with poor connectivity.

Platforms and internal tools

Shell.ai Platform2018
The Shell.ai Platform is Shell's enterprise data science platform, built on C3 AI on Microsoft Azure from Sep 2018 with Databricks as a key component for data engineering, analytics and model building. Shell describes Shell.ai as an internal change programme for a common platform, consistent ways of working and a community of practice across its businesses.
Users
About 800 citizen data scientists, more than 250 data analysts and about 600 professional staff (Databricks case study, undated)
Models
C3 AI and Databricks tooling for in-house models
Source
Shell e
Shell e is a generative AI platform Shell developed that lets teams customise, build and deploy applications on open-source and other large language models, according to Shell's AI web page, which also says Shell uses GPT-4 through Microsoft's Azure OpenAI Service and Microsoft Copilot. Shell has not said when Shell e launched or how many people use it.
Models
Open-source and other large language models
Source
Predictive maintenance on C3 AI Reliability2018
Shell's predictive maintenance program monitors control valves, compressors, pumps and other critical equipment for anomalies across upstream, integrated gas and downstream assets. The program covered more than 10,000 pieces of equipment, with more than 10,000 production models and 15 million predictions a day, in Mar 2022; C3 AI agreed in Oct 2023 to add Shell's models to C3 AI Reliability for other customers; and the program covered more than 13,000 pieces of equipment when Shell extended it under a multi-year agreement in Jun 2026 that adds agent-based root cause analysis and remediation.
Users
13,000+ pieces of equipment (Jun 2026)
Models
C3 AI Reliability and the C3 Agentic AI Platform
Source
Shell Sensor Intelligence Platform
The Shell Sensor Intelligence Platform, built on Delta Lake, collects sensor data from Shell's assets and fed the C3 AI models 20 billion rows a week from more than 3 million data streams in Mar 2022, according to C3 AI. Shell open-sourced the platform's data ingestion code as RTDIP through LF Energy in 2022, Databricks said in Oct 2023.
Source
Chemistry language modelBy Aug 2025
Shell built a custom chatbot with chemistry expertise for technology staff in its R&D organisation, curating 300,000 technical documents down to 154,000 with NVIDIA NeMo Curator and then using domain-adaptive pretraining and fine-tuning. NVIDIA's Aug 2025 case study reports a 30% accuracy gain over the base model on a custom benchmark and 20% less training time; Shell said in Dec 2024 that it was experimenting with custom language models on NeMo.
Users
R&D technology staff
Models
Custom LLM trained with NVIDIA NeMo; base model not named
Source
AI-Powered Discovery Tool and Project FlemingNov 2024 (open-source variant)
Shell's inner source enablement team built the AI-Powered Discovery Tool so developers can find reusable code across GitHub, Shell's internal Stack Overflow and standards documents by natural language search, with a language model that summarises repositories. Shell runs it on CPUs at a monthly cost in the hundreds of dollars and released an open-source variant, Project Fleming, in Nov 2024.
Models
Sentence transformer embeddings and Hugging Face-compatible language models
Source
Operator Round by ExceptionJun 2025 (Yokogawa agreement)
Operator Round by Exception is a machine vision tool Shell built over two years so robots and drones can read gauges and check for leaks and machinery faults on operator rounds. Yokogawa is adding it to its robot management software under a long-term agreement signed in Jun 2025, with pilots at two Shell facilities.
Source
Visual Remote Sensing (VERS)By Dec 2024
Visual Remote Sensing is Shell's deep learning program for satellite and remote sensing data, used for areas such as nature-based solutions, e-mobility and finding unregistered renewable natural gas sources (BioVERS), with batched multi-model inference on NVIDIA Triton Inference Server.
Models
Shell computer vision models on NVIDIA AI Enterprise
Source
T-Pulse2020
T-Pulse is an AI safety monitoring tool from Detect Technologies that reads closed-circuit camera footage to flag unsafe acts, unsafe conditions and potential leaks in real time; Shell had installed it at 26 sites by 2024 and kept expanding it in 2025.
Users
26 sites (2024)
Source

Vendors and partners

C3 AI2018
Core AI platform for predictive maintenance and other industrial AI · Shell selected C3 IoT, now C3 AI, as its AI platform in Sep 2018, renewed the partnership for five years in Jun 2021, let C3 AI add Shell-developed predictive maintenance models to C3 AI Reliability in Oct 2023, and signed a multi-year extension adding agentic root cause analysis on Jun 4, 2026.
Microsoft2018
Azure cloud under the C3 AI platform; GPT-4 through Azure OpenAI Service; Microsoft Copilot and Office 365 for employees · Shell picked C3 IoT with Microsoft Azure as its AI platform in a Sep 2018 deal that also covered Office 365 tools for its then more than 84,000 employees; Shell, Microsoft, C3 AI and Baker Hughes founded the Open AI Energy Initiative in 2021. Shell's undated AI web page says Shell uses GPT-4 through Azure OpenAI Service and Microsoft Copilot.
Databricks
Data and AI platform inside the Shell.ai Platform · Databricks lists Shell as a customer running petabyte-scale data on AWS and Azure for inventory simulation and loyalty program recommendations; the case study is undated.
NVIDIA
Accelerated computing, NeMo, Modulus, Triton and Jetson for simulation, vision and language models · Shell says it has worked with NVIDIA for more than a decade; a Dec 2024 summary covers digital rock simulation, CO2 storage surrogate models, machine vision at sites and custom language models.
SLB2023
Subsurface software and agentic AI for upstream · SLB, AWS and Shell signed a subsurface workflow agreement in Oct 2023, SLB announced a partnership in 2025 to deploy its Petrel software across Shell's assets, and the two signed an agentic AI collaboration on SLB's Lumi platform on Dec 11, 2025. Shell India's chair said in Apr 2026 that the partnership was delivering AI-enabled seismic interpretation workflows.
Amazon Web Services2023
Cloud for SLB subsurface workflows and part of Shell's Databricks estate
Yokogawa2025
Robot and drone management for plant rounds, with Shell's machine vision
Detect Technologies2020
T-Pulse AI safety monitoring
Baker Hughes2021
Co-founder of the Open AI Energy Initiative

Leadership

AI, data and platform leaders

Robbert van Rutten

Executive Vice President Information and Digital Technology and Chief Information Officer

Unconfirmed

Robbert van Rutten became Shell's group CIO on Jul 1, 2023, after serving as CIO for Downstream and as Shell's chief information security officer, and has worked at Shell since 1998. Van Rutten is the first CIO in Shell's history to report directly to the CEO, and told CGI in Dec 2025 that Shell takes a pragmatic, business-led approach to AI, with agentic AI and low-code tools among the shifts he watches most. Shell's 2025 annual report refers to the role without naming him.

Amy Challen

Global Head of AI

Unconfirmed

Amy Challen joined Shell in 2019 after a decade in academia, including a PhD and postdoctoral work in data science, and a role helping build McKinsey's data science practice; Mind the Product described her as Shell's Global Head of AI in Jan 2024. Challen set out Shell's generative AI approach at Web Summit in Nov 2023, and AI Magazine listed her in that role in 2026.

Gerben de Jong

CIO, Integrated Gas, Upstream and Projects & Technology (Jun 2025)

Unconfirmed

Gerben de Jong spoke for Shell on the Jun 2025 Yokogawa robotics agreement as CIO for Integrated Gas, Upstream and Projects & Technology, and appeared in a 2022 Shell careers video as SVP and CIO for Integrated Gas and Renewables and Energy Solutions in Information and Digital Technology.

Detlef Hohl

Chief Scientist, Computation and Data Science

Unconfirmed

Detlef Hohl, Shell's Chief Scientist for Computation and Data Science, featured in a Mar 2024 NVIDIA session on Shell's work with NVIDIA, which spans simulation, machine vision and custom language models, according to Shell's Dec 2024 summary of that work.

Dan Jeavons

Former Vice President, Computational Science and Digital Innovation

Dan Jeavons began working on data science at Shell in 2013 and, as Vice President of Computational Science and Digital Innovation, led the scale-up of predictive maintenance on C3 AI to more than 10,000 pieces of equipment by Mar 2022. Jeavons led Shell's core AI program for 13 years and about 350 researchers before leaving Shell in 2025 to become President of Applied Computing, an energy AI company, according to tech.eu (Dec 2025).

Executives

Wael Sawan

Chief Executive Officer

Wael Sawan has been CEO of Shell since Jan 1, 2023 and joined Shell in 1997 as an engineer with Petroleum Development Oman; he later ran Shell Qatar, Deep Water, Upstream and Integrated Gas and Renewables and Energy Solutions. Sawan told investors on Feb 5, 2026 that further performance improvements would be supported by AI, and on Jul 30, 2026 that Shell was asking how AI could add more value as it pushed for the top of its cost reduction range.

Sinead Gorman

Chief Financial Officer

Sinead Gorman joined Shell in 1999 and has been CFO since Apr 1, 2022, after finance leadership roles in Upstream, Projects & Technology and Integrated Gas. Gorman named AI among the levers for Shell's $5 billion to $7 billion structural cost reduction target at the Mar 2025 Capital Markets Day.

Robin Mooldijk

Former President, Projects & Technology

Robin Mooldijk led Projects & Technology, which held Shell's technology development, and stepped down on Feb 28, 2026 after 35 years once its technical divisions had been integrated into the Integrated Gas, Upstream and Downstream, Renewables and Energy Solutions businesses, shrinking the Executive Committee from nine to eight members.

AI org map

Central AI team and Shell.ai platform; delivery through business-line IT and, since 2026, technical teams inside the businesses

Shell has no AI executive on its eight-member Executive Committee. Shell's CIO, Robbert van Rutten, reports directly to CEO Wael Sawan, according to a Dec 2025 interview, and heads Information and Digital Technology, which runs enterprise IT and business-facing IT teams with their own CIOs. Shell's Global Head of AI, Amy Challen, leads the central AI team. In Jan 2026 Shell said the technical divisions of Projects & Technology had been integrated into its Integrated Gas, Upstream and Downstream, Renewables and Energy Solutions businesses, and Robin Mooldijk left the Executive Committee on Feb 28, 2026.

Shell's AI workforce sits in three layers: a professional group of about 600 mathematics, computational and data science staff supporting more than 100 AI applications, according to an undated Databricks case study, with about 350 researchers under Dan Jeavons before he left in 2025 (tech.eu); business-facing IT teams in Information and Digital Technology, each with its own CIO; and an AI community of more than 12,000 members in Jan 2025, several thousand of whom build AI themselves through citizen development. Shell has not published a current firmwide count of AI or machine learning specialists.

Reporting lines

Chief Executive Officer
Wael Sawan
- Chief Financial Officer
  Sinead Gorman
- Information and Digital Technology
  Robbert van Rutten, Executive Vice President and CIO
  Enterprise IT, Business IT, Information and Digital Engineering, and Information and Digital Services and Operations; reporting line per a Dec 2025 interview
  - Business IT for Integrated Gas, Upstream and Projects & Technology
    Gerben de Jong, CIO
    Title as of Jun 2025
- Projects & Technology
  Robin Mooldijk, until Feb 28, 2026
  Technical divisions integrated into the business lines in early 2026
- Integrated Gas
  Cederic Cremers, President
- Upstream
  Peter Costello, President
- Downstream, Renewables and Energy Solutions
  Machteld de Haan, President
Global AI team
Amy Challen, Global Head of AI
Reporting line not published
Computational Science and Digital Innovation
Dan Jeavons until 2025; successor not published
Unit named in Mar 2022; current form and reporting line not published

Shell's 2025 annual report names an Executive Vice President Information and Digital Technology and Chief Information Officer among those who report to the Audit and Risk Committee but does not name the person; a Dec 2025 CGI interview names Robbert van Rutten as CIO and says he reports directly to the CEO. Shell has not published where the Global Head of AI reports. Shell's executive committee page lists eight members, none with a data, digital or AI title.

Where the AI workforce sits

Professional AI, computational and data science staff
An undated Databricks case study counts about 600 mathematics, computational and data science staff supporting more than 100 AI applications, and tech.eu said in Dec 2025 that Dan Jeavons had led about 350 researchers; Shell has published no current figure.
Not disclosed
Citizen data scientists and analysts in the businesses
The same Databricks case study counts about 800 citizen data scientists and more than 250 data analysts on the platform, and Bernard Marr reported several thousand employees building AI in Jan 2025.
Not disclosed

Named people

Executive sponsors

Wael Sawan
Chief Executive Officer
Sinead Gorman
Chief Financial Officer · Set out AI as a lever for structural cost cuts (Mar 2025)

AI, computational and data science

About 600 professional staff (Databricks case study, undated)

Amy Challen
Global Head of AI
Detlef Hohl
Chief Scientist, Computation and Data Science · Title as of Dec 2024

Information and Digital Technology

A CISO who is also responsible for end-user and enterprise IT took over on Apr 1, 2025.

Robbert van Rutten
Executive Vice President Information and Digital Technology and CIO · Reports to the CEO; role as of Dec 2025
Gerben de Jong
CIO, Integrated Gas, Upstream and Projects & Technology · Title as of Jun 2025

Former leaders

Projects & Technology's technical divisions moved into the business lines in early 2026.

Robin Mooldijk
President, Projects & Technology, until Feb 28, 2026
Dan Jeavons
Vice President, Computational Science and Digital Innovation, until 2025

Use cases

Shell says it has more than 100 AI applications in development and deployment, led by predictive maintenance on more than 13,000 pieces of equipment, production optimization at LNG plants, machine vision for safety and equipment health, generative AI through GPT-4, Microsoft Copilot and its own Shell e platform, and AI in R&D, supply chain and internal audit. Shell has published no business return for any named AI use case and no firm-level AI return; the only value figure for its predictive maintenance program, hundreds of millions of dollars, comes from vendor C3 AI (Jun 2026).

Use case	Area	Status
Predictive maintenance Shell's machine learning models watch control valves, compressors, pumps and other critical equipment for signs of failure on C3 AI Reliability, and under the Jun 2026 extension AI agents will diagnose root causes and suggest fixes.	Operations (Upstream, Integrated Gas, Downstream)	In production
Real-time production optimization at LNG plants Shell's LNG process optimiser takes sensor data and uses AI to calculate efficient equipment settings; Dan Jeavons said in an undated Shell film that an algorithm deployed on an LNG train in Nigeria cut emissions by the equivalent of taking about 28,000 vehicles off the road.	Integrated Gas	In production
Machine vision and equipment health at Pearl GTL Pearl GTL in Qatar uses machine vision to detect hazards and alert supervisors in real time and AI to monitor the health of critical equipment, work the World Economic Forum recognised in its Global Lighthouse Network in Sep 2025.	Integrated Gas	In production
T-Pulse AI safety monitoring T-Pulse reads camera footage at Shell sites to flag unsafe acts, conditions and potential leaks; by 2024 it had raised more than 20,000 alerts since 2020, with interventions in more than 1,300 cases, and sites saw an average 70% fall in T-Pulse observations after deployment (2025).	Safety	In production
Generative AI assistants and Shell e applications Shell says it uses GPT-4 through Microsoft's Azure OpenAI Service and Microsoft Copilot, and that teams build their own generative AI applications on Shell e; Shell has not said how many employees use them.	Enterprise	In production
Robot and drone operator rounds Shell's Operator Round by Exception machine vision lets robots and drones read gauges and check for leaks on Yokogawa's robot management software, piloted at two Shell facilities.	Operations	Pilot
AI with SLB for subsurface, wells and production Shell and SLB plan agentic AI tools for technical experts and an open data and AI infrastructure across subsurface, well construction and production on SLB's Lumi platform; Shell India's chair said in Apr 2026 that the partnership was already delivering AI-enabled seismic interpretation workflows.	Upstream	Pilot
AI surrogate models for CO2 storage Shell built machine learning models with NVIDIA Modulus that simulate CO2 storage in reservoirs about 100,000 times faster than physics-based simulation, to screen potential storage sites.	Upstream and carbon storage	Pilot
Visual Remote Sensing Shell's VERS program applies deep learning to satellite and remote sensing data; its BioVERS offshoot has found unregistered sources of renewable natural gas.	Renewables and Energy Solutions	Pilot
Inventory demand planning Shell uses AI to plan inventory demand and cut waste of products and raw materials, and Databricks says Shell runs more than 10,000 spare-parts simulations across its parts and facilities in hours instead of days.	Supply chain	In production
Loyalty program recommendations Shell built a recommendation engine for its loyalty program from customers' full transaction histories on Databricks.	Mobility	In production
Chemistry language model for researchers Shell's R&D staff can query a custom language model trained on 154,000 curated technical documents to find chemical research and data.	R&D	Pilot
TechXplorer archive search An internship pilot with Huracán Technical Advisory put more than 50 volumes of Shell's TechXplorer Digest behind a retrieval-augmented search tool running open-weight models on locally hosted servers, with no external API calls.	R&D	Pilot
AI code discovery for developers Shell's AI-Powered Discovery Tool lets developers search code and documentation in plain language and summarises repositories with a language model.	IT	In production
Methane emission localisation Shell is testing an AI solution that continuously monitors methane emissions from sensor networks and open-sourced its pyELQ localisation and quantification code in Mar 2024.	Environment	Pilot
Internal audit reporting Shell's Chief Internal Auditor prepared reports to the Audit and Risk Committee with AI in 2024 and 2025.	Corporate	In production
Board meeting preparation Shell increasingly uses AI to prepare for Board meetings and is exploring more use of AI in the Board environment, according to the 2025 annual report.	Corporate	In production

Adoption

Shell spread AI in layers: data science on cloud computing from 2013, an enterprise AI platform on C3 AI and Microsoft Azure from 2018, citizen development so employees can build their own solutions, and generative AI through GPT-4 on Azure OpenAI, Microsoft Copilot and the Shell e platform. Shell's AI community had more than 12,000 members in Jan 2025, out of about 96,000 employees at the end of 2024, and a Digital Skills programme was open to virtually every employee. Shell has not published how many employees use a generative AI assistant; Amy Challen said in Jan 2024 that Shell's change management for generative AI was not there yet, after webcasts in 2023 and training for general managers and IT managers, and Shell India's chair said in Apr 2026 that Shell was moving from experimentation to enterprise deployment.

By the numbers

Equipment under AI predictive maintenance: 13,000+; C3 AI announcement of Shell's multi-year extension, Jun 4, 2026

AI community members: 12,000+; Several thousand of them actively building AI (Bernard Marr, Jan 2025)

AI applications: 100+; In development and deployment, per Shell's website (undated)

Employees: 85,000; Dec 31, 2025, down from 96,000 a year earlier (2025 Annual Report)

Rollout, first pilot to today

2013
Shell began building data science on cloud data and computing, building on a statistics group it formed in the 1970s, Dan Jeavons said in a Shell film.
Shell
Sep 2018
Shell selected C3 IoT, now C3 AI, on Microsoft Azure as its AI platform, starting with predictive maintenance, and moved employee communication tools to Office 365.
Microsoft
Jun 2021
Shell renewed its C3 AI partnership for five years, the same year Shell, C3 AI, Microsoft and Baker Hughes founded the Open AI Energy Initiative.
C3 AI
Mar 2022
Shell's predictive maintenance reached more than 10,000 pieces of equipment, with more than 10,000 production models generating over 15 million predictions a day.
C3 AI
Oct 2023
C3 AI agreed to add Shell-developed predictive maintenance models for control valves and critical equipment to C3 AI Reliability, freeing Shell's engineers and data scientists for new applications.
C3 AI
Nov 2023
Amy Challen, Shell's head of AI, said at Web Summit that Shell was applying generative AI to knowledge management, R&D, support functions and personal productivity, with human checks on output.
Rigzone
Nov 2024
Shell released Project Fleming, the open-source variant of its AI-Powered Discovery Tool for developers.
Shell TechXplorer Digest
Dec 2024
Shell said it had deployed machine vision models on NVIDIA Jetson Orin devices at sites and was experimenting with custom large language models built on NVIDIA NeMo for R&D.
Shell
Jan 2025
Shell's AI community passed 12,000 members, with AI applications sorted into green, amber and red risk zones.
Bernard Marr
Mar 2025
Shell raised its structural cost reduction target to $5 billion to $7 billion by 2028, with CFO Sinead Gorman naming AI among the levers.
Capital Markets Day 2025
Jun 2025
Shell signed a long-term agreement with Yokogawa to put its Operator Round by Exception machine vision on robots and drones, with pilots at two Shell facilities.
Yokogawa
Aug 2025
NVIDIA published a case study on Shell's custom chemistry language model, trained with NeMo on 154,000 curated technical documents for R&D staff.
NVIDIA
Sep 2025
The World Economic Forum named Shell's Pearl GTL plant in Qatar to its Global Lighthouse Network for machine vision and AI equipment monitoring.
Shell
Dec 2025
Shell and SLB agreed to develop agentic AI for upstream operations on SLB's Lumi platform.
SLB
Jun 2026
Shell extended its C3 AI deployment under a multi-year agreement covering more than 13,000 pieces of equipment and adding AI agents for root cause analysis.
C3 AI

Economics & controls

Spend and ROI

Technology budget

Shell does not publish an IT, digital or AI budget. Shell spent $1.17 billion on R&D in 2025 and says it pairs R&D with digital solutions, often powered by AI.

What AI itself costs

Shell has not disclosed what it spends on AI or what its C3 AI, Microsoft, Databricks or NVIDIA contracts cost. The only cost figure in the public record is small: Shell's TechXplorer Digest said in Jan 2026 that its Shell-wide AI code discovery tool runs on CPUs for hundreds of dollars a month.

Reported ROI

Shell has published no financial return from AI, for any use case or for the company. Dan Jeavons said in Mar 2022 that predictive maintenance pays for itself and that the return is compelling, without a figure, and C3 AI's president said on Jun 4, 2026 that the expanded partnership shows enterprise AI at scale delivering hundreds of millions of dollars in economic value, a vendor claim Shell has not repeated. Shell reported $5.1 billion of structural cost reductions from 2022 to 2025 and $700 million more in the first half of 2026 without crediting any of it to AI.

AI in the cost plan

Shell CFO Sinead Gorman said on Mar 25, 2025 that most of the planned $5 billion to $7 billion of structural cost cuts by 2028 would come from non-portfolio activities, including procurement, technology and AI opportunities and a leaner corporate centre. Wael Sawan said on Feb 5, 2026 that lower costs and further performance improvements, supported by AI, give Shell confidence for the road ahead, and on Jul 30, 2026 that Shell was halfway through the range, aimed for the top end and was asking how AI could add more value.

Headcount

Shell employed 85,000 people at the end of 2025, down from 96,000 at the end of 2024 and 103,000 at the end of 2023. Shell's 2025 annual report attributes the drop to divestments in Singapore, Nigeria and the UK, ended activities in Downstream, Renewables and Energy Solutions, and efficiencies in Projects & Technology, IT and Finance, and does not attribute it to AI.

Governance and risk

Risk zones for AI applications

Shell sorts AI applications into green, amber and red zones, with light-touch controls for simple, low-risk green-zone tools and more oversight and expertise for amber and red, so thousands of employees can build their own AI solutions inside the framework (Bernard Marr, Jan 2025).

Code of Conduct

Shell introduced a refreshed Code of Conduct in Jan 2026 that clarifies its expectations for the responsible use of AI, according to the 2025 annual report. Amy Challen said in Nov 2023 that staff should not enter confidential data into public AI tools and that people check AI output until Shell is confident it works.

Board and audit oversight

Shell's Board review, discussed in Feb 2026, set a sharper focus on technology and AI, including in the Board environment, as a 2026 priority, and the Audit and Risk Committee lists AI compliance and finance systems and AI among its 2026 priorities. Shell's Audit and Risk Committee discussed AI compliance with the Chief Ethics and Compliance Officer in 2024, and management and the Board reviewed AI and quantum computing as emerging risks in 2025.

Cyber security and regulation

Shell's 2025 annual report says agentic and generative AI add data leakage, unauthorised access, data manipulation and system exploitation to its cyber risks, and that wider AI misuse raises the challenge of complying with rules such as the EU Artificial Intelligence Act. Shell centralised its cyber defence teams in a Cyber and Information Security Office from Mar 1, 2025 under its Information and Digital Technology standard, and a new CISO took over on Apr 1, 2025.

Keeping data local

Shell's TechXplorer archive search pilot, described in Jan 2026, ran open-weight models on its partner's locally hosted servers with no external API calls, to meet privacy requirements. Shell's machine vision safety models run on NVIDIA Jetson Orin edge devices at sites where internet connectivity is unreliable.

Recent signals

Jul 30, 2026
Wael Sawan told analysts that Shell was halfway through its $5 billion to $7 billion structural cost reduction range, after $700 million of cuts in the first half of 2026, and that Shell was asking how AI could add more value, without putting a figure on AI's part.
Q2 2026 earnings call transcript
Jun 4, 2026
C3 AI said Shell is extending its predictive maintenance program under a multi-year agreement, covering more than 13,000 pieces of equipment on Microsoft Azure and adding AI agents for root cause analysis and remediation.
C3 AI
Apr 10, 2026
Mansi Madan Tripathy, chair of Shell's companies in India, told Platts that Shell was moving from AI experimentation to enterprise deployment and that its SLB partnership was delivering AI-enabled seismic interpretation workflows.
S&P Global Platts, via Shell India
Mar 12, 2026
Shell's 2025 annual report said the organisation increasingly uses AI to prepare for Board meetings and listed AI compliance and finance systems and AI among the Audit and Risk Committee's 2026 priorities.
2025 Annual Report
Mar 12, 2026
Shell's 2025 annual report added agentic and generative AI to its cyber security risk description and said mainstream generative AI could change competition in its commodity markets.
2025 Annual Report
Feb 5, 2026
Wael Sawan told investors that lower costs and further performance improvements supported by AI give Shell confidence, without putting a figure on AI's contribution.
Q4 2025 results speech
Jan 23, 2026
Shell's TechXplorer Digest described its AI-Powered Discovery Tool for developers and a retrieval-augmented archive search pilot run on open-weight models.
Shell TechXplorer Digest
Jan 20, 2026
Shell said the technical divisions of Projects & Technology had been integrated into its business lines and that Robin Mooldijk would step down on Feb 28, 2026, cutting the Executive Committee to eight members.
Shell
Jan 2026
Shell introduced a refreshed Code of Conduct that clarifies its expectations for the responsible use of AI.
2025 Annual Report
Dec 12, 2025
Tech.eu reported that Dan Jeavons, who led Shell's core AI program for 13 years as Vice President of Computational Science and Digital Innovation, had left Shell and joined Applied Computing as President.
Tech.eu
Dec 11, 2025
SLB and Shell signed a strategic collaboration to develop agentic AI and an open data and AI infrastructure for upstream operations on SLB's Lumi platform.
SLB
Dec 8, 2025
Shell CIO Robbert van Rutten told CGI that he is the first Shell CIO to report directly to the CEO and that Shell takes a pragmatic, business-led approach to AI, with agentic AI and low-code tools among the main shifts.
CGI

Quick answers

How does Shell deploy AI?: Shell deploys AI mainly in its operations: since 2018 it has run predictive maintenance on the C3 AI platform on Microsoft Azure, which covered more than 13,000 pieces of equipment when Shell extended the deal in Jun 2026 and is adding AI agents. For office and research work, Shell says it uses GPT-4 through Azure OpenAI, Microsoft Copilot and Shell e, a generative AI platform it built. Shell has published no usage count and no financial return from AI.
What does Shell use AI for?: Shell says it has more than 100 AI applications in development and deployment, led by predictive maintenance on more than 13,000 pieces of equipment, production optimization at LNG plants, machine vision for safety and equipment health, generative AI through GPT-4, Microsoft Copilot and its own Shell e platform, and AI in R&D, supply chain and internal audit. Shell has published no business return for any named AI use case and no firm-level AI return; the only value figure for its predictive maintenance program, hundreds of millions of dollars, comes from vendor C3 AI (Jun 2026).
Who leads AI at Shell?: AI at Shell is led by Robbert van Rutten (Executive Vice President Information and Digital Technology and Chief Information Officer); Amy Challen (Global Head of AI); Gerben de Jong (CIO, Integrated Gas, Upstream and Projects & Technology (Jun 2025)). Ownership of AI on the org chart: Global Head of AI Amy Challen; CIO Robbert van Rutten reports to the CEO.
What AI platform does Shell use?: Shell runs its industrial AI on the Shell.ai Platform, built on C3 AI on Microsoft Azure since Sep 2018 with Databricks for data and extended under a multi-year C3 AI deal in Jun 2026; Shell also built Shell e, a platform on which teams build generative AI applications.
Which AI models does Shell use?: Shell builds its own machine learning models for equipment monitoring, says it uses GPT-4 through Microsoft's Azure OpenAI Service and Microsoft Copilot, and fine-tuned a chemistry language model with NVIDIA NeMo; Shell e runs open-source and other large language models.
How many Shell employees use AI?: Shell's AI community had more than 12,000 members in Jan 2025, against about 96,000 staff at the end of 2024; Shell has published no count of employees who use Microsoft Copilot or Shell e.
How far has AI spread at Shell?: Shell has reported no business return it credits to AI, for any use case or for the company: none of its $5.1 billion of structural cost cuts from 2022 to 2025, or the $700 million more in the first half of 2026, is credited to AI; CFO Sinead Gorman named AI as one of several levers for the $5 billion to $7 billion cost target and CEO Wael Sawan has put no figure on its part; and the hundreds of millions of dollars of value put on predictive maintenance in Jun 2026 is a claim by vendor C3 AI, not by Shell. Shell has not reported a company-wide AI rollout with a user count either: it says it uses GPT-4 through Azure OpenAI, Microsoft Copilot and its own Shell e platform without saying how many of its 85,000 employees use them, and the AI community of more than 12,000 members is a membership figure from Bernard Marr (Jan 2025), not a count of tool users. That leaves Shell at Experimenting although predictive maintenance has run in production since 2018 and covers more than 13,000 pieces of equipment; one cost, revenue or output figure Shell credits to AI, or a firmwide user count with a named owner such as Global Head of AI Amy Challen, would move it to Piloting.

Open questions

Who speaks for AI on Shell's Executive Committee, which has no CIO or AI member, and where does the Global Head of AI report now that Projects & Technology has been folded into the businesses?
Does Robbert van Rutten still lead Information and Digital Technology, which Shell's 2025 annual report does not name, and who succeeded Dan Jeavons in computational science and digital innovation?
How many Shell employees can use Microsoft Copilot, GPT-4 through Azure OpenAI or Shell e, and how many use them each week?
Which models run on Shell e beyond GPT-4 and the chemistry model fine-tuned with NVIDIA NeMo, and where does that inference run?
What business return does predictive maintenance on more than 13,000 pieces of equipment deliver to Shell, given that only C3 AI has put a value on it?
How much of Shell's structural cost cuts since 2022 ($5.1 billion by the end of 2025 and $700 million more in the first half of 2026) comes from AI?
What does Shell spend on AI, and what did the Jun 2026 multi-year C3 AI extension cost?
Is the count of more than 100 AI applications current, and how many of them run in production?
What business return do Shell's other deployed tools deliver, such as LNG production optimization, T-Pulse and inventory planning?
What will the SLB agentic AI collaboration and the Yokogawa robot pilots deliver, and when?

Sources

Shell

Filings, reports and investor materials

Partners

News

Data confidence

Confirmed: Stated in a primary or reputable public source.
Estimated: Derived or approximated from public reporting.
Unconfirmed: Reported, but not yet corroborated.

Shell profile · Enterprise AI diffusion report by Imagine · facts checked against their sources on Nov 1, 2026

More company profiles

All profiles