Reported impact
Business outcome
Repsol reported substantial employee productivity gains and better document output from the Copilot pilot, alongside strong user adoption and a broader company rollout of Copilot Chat.
- Productivity Improvement
MicrosoftEnergyDocument ProcessingEnterprise Knowledge SearchAI and Machine LearningGenerative AI / LLMProductivity Improvement2025 MicrosoftMS#0078d4EnergySpainRepsol wanted to measure whether generative AI could reduce time spent on routine knowledge work such as finding information, summarizing documents, drawing conclusions, comparing supplier proposals, and preparing…121 minutes a weekAug 12, 20256Document Processing|Enterprise Knowledge SearchCopilot Chat · Excel · Microsoft 365 Copilot0 2025-08-12 6
Repsol
Document Processing · Enterprise Knowledge Search
Repsol wanted to measure whether generative AI could reduce time spent on routine knowledge work such as finding information, summarizing documents, drawing conclusions, comparing supplier proposals, and preparing presentations, while improving document quality and employee experience.
121 minutes a week
16.2% higher in quality
600 people
four-month trial period
Solution
Repsol ran a 600-person, four-month pilot of Microsoft 365 Copilot after training participants to use the assistant in daily work. The implementation placed Microsoft 365 Copilot inside Word, Excel, and Teams, where employees could issue natural-language prompts for tasks such as drafting supplier emails, summarizing recorded meetings, comparing supplier proposals, analyzing information, and preparing presentations. According to the source, Copilot combines large language model capabilities with company data held in the cloud and information available on the Internet, while honoring each user's existing access permissions so responses are limited to content the employee is authorized to see. Repsol also describes a broader rollout of Microsoft Copilot Chat to more than 25,000 professionals, including secure web-style querying, file upload for prompt context, and agent creation, extending the same generative AI workflow beyond the original pilot.
Data flow
Employees submit prompts inside Microsoft 365 applications such as Word, Excel, and Teams. Microsoft 365 Copilot pulls context from Repsol's cloud-held company data, Internet information, and in specific workflows artifacts such as recorded meetings, supplier proposals, news, reports, and uploaded files, then returns summaries, drafts, comparisons, and conclusions directly to the user in the application they are using.
Solution architecture
- Serving
- Microsoft 365 Copilot The generative AI assistant used by employees to find information, summarize content, compare proposals, draft communications, and help create presentations.
- Word One of the Microsoft 365 applications where Copilot is embedded for document drafting and content creation.
- Excel One of the Microsoft 365 applications where Copilot is available for analysis tasks.
- Teams One of the Microsoft 365 applications where Copilot can work with meeting-related content such as recorded sessions and summaries.
- Copilot Chat The Microsoft assistant Repsol is expanding to more than 25,000 professionals for secure natural-language search, file-grounded prompting, and agent creation.
Architecture clues
- Copilot Chat is described as supporting agent creation and file upload for prompts.
- Copilot uses company data in the cloud plus Internet information as context.
- Participants received training before the pilot began.
- Repsol is deploying Copilot Chat to more than 25,000 professionals.
- Responses are limited to information each user is authorized to access within the organization.
- The pilot tested real business situations including supplier proposal comparison and presentation preparation.
Evidence from the source
"A pioneering study reveals that, on average, each Repsol employee who used Copilot, Microsoft's smart assistant, saved 121 minutes a week in carrying out their tasks."
"Copilot was particularly useful for three types of tasks: finding information, summarizing documents, and drawing conclusions."
"The research also included experiments on real business situations, such as using Copilot to compare proposals from different suppliers or to prepare a presentation by compiling certain information."
"The research showed that the documents presented by people who used Copilot were 16.2% higher in quality."
"This practical research involved 600 people, who received training before beginning a four-month trial period with this assistant."
"This smart assistant for the Internet 'is changing how users interact with search engines, using natural language in their queries, and obtaining more elaborate and precise answers... together with the possibility of creating agents or uploading files that can be used as a basis for prompts,' highlights Antonio Cruz."
Transparency Evidence & methodology
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
How scores are calculated
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Customer reference value100/100
- Problem clarity20/20Measures whether the customer's operational problem is specific enough to compare.
- Implementation depth25/25Rewards named components, data flow, integrations, and architecture detail.
- Outcome specificity20/20Measures whether the reported benefit is concrete rather than a generic claim.
- Quantified impact25/252 metrics describe measurable impact.
- Deployment scale10/10Rewards evidence of a live or scaled deployment and meaningful integration scope.
Evidence confidence97%
- Source authority25/25The deployment is described by the adopting organization or a government source.
- Claim traceability23/2511 of 12 key fields link to a named source.
- Direct evidence20/206 direct source quotes retained.
- Date provenance10/10The publication date comes from the source page or publisher.
- Source transparency9/102 source references document publisher, URL, authority, retrieval, and supported fields.
- Implementation specificity10/10Rewards source-backed components, data flow, quantified impact, and a specific outcome.
Traceability
Sources & provenance
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
- https://www.repsol.com/en/energy-move-forward/innovation/smart-copilot-help-at-work/index.cshtmlRepsolFirst-party
Repsol customer story
Supports- Customer
- Business problem
- Solution
- Products
- Solution components
- Data flow
- Integrations
- technical area
- use case category
- Business outcome
- outcome category
- Metrics
- Architecture
- evidence quotes
- Publication date
- https://www.repsol.com/en/footer/contact/faqs/index.cshtmlRepsolFirst-party
Repsol FAQs
Our main headquarters is the Repsol Campus, which is located on Calle Méndez Álvaro, 44 in Madrid, Spain.
Supports- Region
Date evidence
- Published on source page Aug 12, 2025 · Repsol · First-partyOpen date source
Not separately sourced Industry, Company size. Treat these normalized attributes as lower-confidence until a dedicated source is added.
Related
Similar use cases
Microsoft Sample Sep 25, 2024 National Grid National Grid is helping the UK achieve its net zero ambitions. Azure · Microsoft Copilot Microsoft Sample Oct 15, 2026 Bar-Ilan University Students who miss lectures struggle to reconnect new lessons with prior material when course information is spread across notes, recordings, slides, and readings, and lecturers need a reliable way to answer routine… 71 courses Azure AI Services · Microsoft Azure · Microsoft Copilot Studio 
Deutsche Telekom Sample Apr 14, 2026 Thüga AG Public utilities in Thüga's network manage sensitive data from network operations, customer management, and regulatory processes, and need a secure way to introduce AI without losing digital sovereignty or forcing each… around 100 public utilities Enterprise GPT solutions · T Cloud Public