Skip to main content Skip to footer


Accenture and Cohere

Focus on Data Security in Generative AI

5-minute read

March 4, 2024

Accenture and Cohere are joining forces to accelerate how we can help companies incorporate AI into their operations at ​scale, prioritizing​ privacy and security.

More and more companies are implementing generative AI solutions in their businesses in innovative ways. I have seen many enterprises experiment with the use of generative (gen) AI beyond simple chatbots and automation tools over the past year, and this wave of transformation is likely just the beginning.

As both interest and adoption grow, companies are seeking to tailor and deploy large language models (LLMs) at scale ,  while remaining vigilant about data protection.  This new approach is setting the stage for significant developments in the field of generative AI.

Building on Accenture’s $3B investment in AI, we have teamed up with Cohere, a leading AI platform for enterprise, to accelerate how we can help companies incorporate AI into their operations at scale, prioritizing privacy and security.

Accenture and Cohere are joining forces to provide businesses with generative AI solutions powered by Cohere’s leading foundation models (Command, Embed and Rerank), and Accenture’s cloud, data, AI, industry, and functional expertise. Companies can implement these models alongside other Large Language Model (LLM) frameworks and applications, including customization options and Cohere’s leading Retrieval-Augmented Generation (RAG) capabilities that connect the models with relevant proprietary data sources.

Addressing the trust and verifiability challenge

As adoption of AI increases, several challenges remain evident, including Trust and Verifiability. It is essential to establish confidence in the responses generated by LLMs and ensure the accuracy of these responses. Companies like Cohere are leading substantial advances in addressing these obstacles with advanced RAG capabilities that anchor their LLMs to enterprise data sets and improved search results, resulting in responses that are well-founded, highly relevant, and include source citations. This contributes to a more reliable, verifiable, safe, and responsible use of AI.

Meeting enterprise-grade security and data privacy requirements is also crucial when launching AI solutions. Companies can deploy Cohere models closely aligned with their specifications. For instance, using SaaS APIs in their cloud provider of choice, as a private deployment in their VPC (Virtual Private Cloud) on any of the hyperscale cloud providers, or using their own on-premise servers and infrastructure.

Enterprises can explore customized solutions for their specific needs, optimizing accuracy and results by orchestrating different types of models. By choosing the right ones, with the added ability to customize and scale them, enterprises can reduce overall TCO (Total Cost of Ownership), while improving efficiency.

Cohere has capabilities that enable enterprises to meet these challenges:

  • Private deployment of Cohere’s LLMs along with supporting frameworks for enterprise use.

  • Retrieval augmented generation to improve relevance and reliability of LLM generated results grounded in enterprise data.

  • Ability to customize LLMs of different model sizes which allows for lower fine-tuning and inferencing costs.

We recognize that companies might face complexities when implementing generative AI solutions, hence our joint team of Cohere and Accenture AI experts can work with clients to help them identify the most effective way to deploy AI for their businesses. This guidance minimizes the time it takes to get the models up and running in a manner that is secure, scalable, and affordable.

“Cohere is proud to team up with Accenture to help enterprises leverage AI technology to deliver real business value. Generative AI for enterprise has moved past the experimenting stage and reached the point where companies can rapidly and effectively implement these solutions in a secure way, which is why Accenture is working closely with Cohere to offer their clients business solutions that are secure and available on every major cloud provider or on-site.” - Martin Kon, President and COO of Cohere

Powering the Accenture Treasury Insights platform through AI

Accenture has built a knowledge agent for our finance and treasury teams. With AI capabilities, this knowledge agent empowers our global treasury operations team to detect financial variance and make better decisions. Cohere's Command model powers summarization for the knowledge agent delivering key financial metrics in natural language. The solution then sends customized alerts and chat notifications to our finance and treasury teams, summarizing key insights that require attention.

By providing timely and actionable alerts on key financial metrics, this capability enables our treasury professionals to make well-informed decisions promptly, leading to better financial management and strategic planning.

We believe the following capabilities will be key to maximizing value and driving differentiation within enterprises, including the ability to:

  • Provide a conversational interface optimized for business interactions.

  • Customize models to domain-specific knowledge and proprietary training data sets.

  • Connect to data stores in real-time to provide more timely and accurate information and ground responses, including citing data sources for verification.

  • Deploy on private infrastructure (self or cloud-managed) to protect sensitive data, where required by strategic or regulatory considerations.

By collaborating with enterprise-focused generative AI technology firms such as Cohere, Accenture can deliver advanced capabilities and innovative solutions to our enterprise clients, enabling them to maintain a competitive edge in an ever-evolving digital landscape.


Lan Guan

Chief AI Officer