Schneider Electric is harnessing the power of Microsoft’s Azure OpenAI platform to create advanced chatbot solutions aimed at enhancing employee productivity and enriching customer interactions. Key tools in this initiative include Resource Advisor Client, a data analysis and decision-making copilot; Jo-Chat GPT, an internal conversational assistant; and Knowledge Bot, which supports customer care representatives. They are also launching Finance Advisor, designed to assist financial analysts in their accounting tasks, along with Conversational Search, allowing customers to inquire about products in a natural conversational style. Furthermore, Schneider Electric plans to integrate GitHub Copilot to streamline their creation processes and operational workflows. The Azure OpenAI platform enables companies to craft secure versions of ChatGPT that utilize proprietary data.
Philippe Rambach, Schneider Electric's Chief Artificial Intelligence Officer, shared insights into the company’s approach to developing these AI tools and offered guidance on implementing large-scale generative AI projects. Here’s an excerpt from that conversation.
**On Embracing Generative AI and Managing Risks**
Philippe Rambach explained that the development of their proprietary version of ChatGPT, Jo-Chat GPT, was a strategic move to mitigate data leakage risks while still leveraging AI's value. They have implemented robust training programs for staff, emphasizing the importance of verifying AI outputs before usage. The Knowledge Port tool exemplifies this strategy, enabling customer care personnel to extract reliable information from user manuals and FAQs.
To prevent data leaks, the training data for their systems consists solely of internally sourced information, ensuring greater accuracy and reliability. The implementation of vectors for embedding and direct links to source material provides an extra layer of validation for customer care employees.
**Why Microsoft Azure?**
Schneider Electric's choice of Azure was influenced by their need for a secure industrial solution. With a strong existing partnership with Microsoft, Rambach indicated that opting for Azure was a logical choice for launching their generative AI journey, emphasizing safety, efficiency, and ease of use. While they plan to explore other large language models (LLMs), starting with Microsoft’s technology minimized risks and complexities.
**Exploration of Additional AI Systems**
While details on future systems remain confidential, Rambach revealed that Schneider Electric is evaluating various AI solutions, including Google Bard and LangChain for knowledge management applications. The company is committed to finding and utilizing the best technologies available.
**Identifying Use Cases for AI Implementation**
Schneider Electric prioritizes business needs over technical capabilities when exploring generative AI. For instance, they plan to purchase existing solutions for software code generation instead of developing one in-house. However, they recognize the necessity of developing tailored internal solutions for proprietary knowledge management, essential to enhancing customer offerings.
**Challenges and Strategies in Data Integration**
During implementation, Schneider Electric emphasized a unified approach, incorporating insights from both AI specialists and business leaders to ensure effective deployment. They focused on delivering value at scale rather than merely conducting pilot projects. This collaborative strategy helps bridge data silos and guarantees that the developed solutions meet business requirements.
**Safety and Cost Considerations**
Safety is paramount; AI products undergo rigorous cybersecurity and data privacy evaluations through Schneider Electric’s standard protocols. In terms of costs, the company opted for GPT-3.5 for their customer care center, emphasizing cost-effectiveness and environmental sustainability. By prioritizing efficient solutions, Schneider Electric aims to minimize both expenditure and carbon emissions.
**Future AI Developments**
Looking ahead, Schneider Electric plans to expand Copilot offerings, enhance knowledge management, and further support employees in utilizing generative AI. These developments will serve to bolster both internal processes and customer interactions.
**Lessons Learned for Future Deployments**
Rambach’s key takeaway from these initiatives stresses the importance of agility in technology adoption. Companies should be ready to pivot away from established solutions if new technologies deliver superior capabilities. This means recognizing when to halt continued investments in a project and embrace fresher, more effective alternatives—an essential mindset in the rapidly evolving landscape of artificial intelligence.
By embracing innovative technologies and focusing on structured implementation, Schneider Electric is actively shaping the future of AI in industrial contexts, ensuring that both operational efficiency and customer satisfaction are prioritized.
Ranking
