Barysa S.A. helps you apply the same principles PwC used to scale generative AI on Microsoft Azure AI. PwC designed a secure architecture that enables broad adoption while protecting data, so employees can safely use GenAI tools and proprietary solutions to boost productivity. Drawing on our soporte tecnológico especializado, we work with your team to build a scalable, secure AI framework on Azure that fits your governance, compliance, and integration needs.
What is ChatPwC and how is it built on Microsoft Azure AI?
ChatPwC is PwC’s private Generative AI assistant, designed to help its people handle day‑to‑day work more efficiently while keeping client and firm data protected.
PwC built ChatPwC on Microsoft Azure AI with a few key components:
- **Azure OpenAI Service** – Provides the large language models that power natural language conversations and content generation.
- **Azure AI Document Intelligence** – Delivers text extraction and optical character recognition (OCR), so employees can pull structured insights from complex documents.
- **Azure AI Search with retrieval‑augmented generation (RAG)** – Adds semantic search and ranking so ChatPwC can ground answers in PwC’s own content and domain knowledge.
- **Azure Kubernetes Service (AKS)** – Orchestrates these services as APIs and integrates them with PwC’s proprietary plugin architecture.
From a security standpoint:
- All data ingested into ChatPwC is stored behind **PwC firewalls**.
- Data used in ChatPwC is **not used to train** the underlying large language models.
- PwC’s internal governance organizations provide quality assurance and oversight of how GenAI is deployed.
This combination lets PwC scale GenAI across the firm while keeping control over data, performance, and compliance.
How widely has ChatPwC and GenAI been adopted inside PwC?
PwC has seen broad, structured adoption of ChatPwC and other GenAI tools across its global workforce.
**ChatPwC adoption**
- Started with a **four‑week pilot** led by the low‑code and GenAI team.
- Now used by **75,000 professionals in the United States** and another **125,000 employees globally**.
- In total, around **200,000 PwC employees** are using ChatPwC.
**Adoption journey**
PwC designed a four‑phase journey to help people build skills and confidence:
1. **Learn it** – Educate on GenAI fundamentals, responsible use, and prompting basics.
2. **See it** – Demonstrate real use cases and outcomes.
3. **Try it** – Encourage experimentation through pilots and guided sessions.
4. **Do it** – Embed GenAI into everyday work.
To support this, PwC:
- Ran large‑scale learning sessions, including one with OpenAI CTO Mira Murati that drew **35,000+ attendees**.
- Built the **ChatPwC brand** as part of firm culture.
- Introduced a **prompt catalog** so users can save and share effective prompts.
**Broader GenAI tools**
PwC extended similar adoption programs to:
- **Copilot for Microsoft 365** – Popular for email coaching and first‑draft creation in Word and Outlook.
- **GitHub Copilot** – Adopted by about **70% of relevant engineering staff**.
- **Microsoft Power Platform GenAI features** – Rolled out with targeted enablement.
Through the **My AI** upskilling initiative, PwC ran hands‑on events like “prompting parties,” helping drive **95% engagement** with My AI. During a Microsoft GenAI pilot:
- **88% of users** reported time savings.
- **86%** were highly satisfied.
- **69%** said GenAI improved accuracy and quality of their work.
PwC leaders view adoption as ongoing, with continued investment in features, training, and new use cases.
What business impact is PwC seeing from its GenAI and ChatPwC programs?
PwC is using GenAI to rethink how work gets done across internal and client engagements, and it is tracking tangible outcomes.
**Productivity and speed**
- PwC reports that top ChatPwC users are completing certain tasks **up to eight times faster** than before.
- In the Microsoft GenAI pilot, **88% of users** said the tools **saved them time**.
**Quality and accuracy**
- **69% of pilot users** said GenAI **improved accuracy and quality** of their work.
- **86%** reported being **highly satisfied** with the GenAI experience.
**Breadth of use cases**
- PwC has seen **thousands of internal and client use cases** across industry domains.
- The firm has not needed to **fine‑tune large language models** so far; instead, it relies on:
- Strong base models from Azure OpenAI Service.
- RAG patterns using Azure AI Search.
- Document processing via Azure AI Document Intelligence.
**Workforce transformation and skills**
- Through the **My AI** upskilling program, PwC has achieved **95% engagement**, indicating that most employees are actively building GenAI skills.
- Events like prompting parties and large‑scale learning sessions help employees reimagine how they approach tasks, from drafting documents to analyzing complex materials.
**Technology evolution**
- PwC is expanding its architecture with **Microsoft AutoGen** to build agentic systems that can coordinate multiple GenAI capabilities.
- Similar adoption programs for **Copilot for Microsoft 365**, **GitHub Copilot**, and **Power Platform GenAI** are extending the impact beyond ChatPwC.
Overall, PwC sees GenAI as a way to reshape how its people work: handling information‑heavy tasks more efficiently, improving quality, and giving employees opportunities to develop in‑demand AI skills while maintaining strong security and governance controls.
Sign in with LinkedIn.jpg)