IBM and Meta are implementing the combined power of IBM’s watsonx AI and data platform and Llama to help businesses reach their AI goals. We’re seeing strong adoption of Llama on watsonx.ai as companies increasingly turn to open AI models and tools to drive real business value.

This builds on the work IBM and Meta started last year to advance open innovation around AI—bringing together leading organizations, startups, researchers, and academia across industries in the AI Alliance to advance responsible, open AI around the world.

Today, IBM offers Meta’s Llama as part of its watsonx.ai model catalog. Watsonx.ai provides a next-generation enterprise studio for AI builders worldwide to train, validate, tune, and deploy AI models. From telecom to financial services and local governments, enterprises across a wide range of sectors around the world are benefitting from watsonx.ai and Llama models.

Dun & Bradstreet, a leading global provider of business decisioning data and analytics, is transforming its lead generation solutions by integrating Generative AI technologies. As a part of its proprietary process, Dun & Bradstreet is utilizing Llama on watsonx.ai to generate detailed company profiles that are easily searchable using natural language. This approach surpasses traditional methods that rely on fixed taxonomies or keyword searches, enabling users to identify potential customers and suppliers with greater precision. By selecting Llama, Dun & Bradstreet has been able to achieve the necessary performance in a cost-efficient way. This innovative collaboration with IBM, not only broadens the scope and quality of the Dun & Bradstreet’s proprietary data and analytics, but also keeps information up-to-date and relevant across diverse customer applications.

IBM and Sevilla FC introduced Scout Advisor, an innovative generative AI tool that Sevilla FC will use to provide their scouting team with a comprehensive, data-driven identification and evaluation of potential recruits. Built on watsonx, Sevilla FC’s Scout Advisor will integrate with their existing suite of self-developed data-intensive applications. The solution’s natural language processing capabilities have enabled Sevilla FC to use multiple large language models (LLMs) including Meta’s Llama 3 to help enhance the accuracy and effectiveness of their player identification. Using language prompts from Sevilla FC scouters that describe the key characteristics of the players searched, Scout Advisor generates curated lists of candidates based on stated requirements and summarizes the full set of scouting reports for each individual player. Additionally, Scout Advisor links every player to Sevilla FC’s own data applications to obtain deep insights about their quantitative performance figures.

Meanwhile, AddAI is a Czech Republic-based customer interaction platform that enables companies to deploy and operate omnichannel AI assistants. With Llama as the foundation model and using a retrieval-augmented generation (RAG) methodology, IBM engineers supported a pilot to create an AI-driven Q&A chatbot to ensure precise responses in the Czech language, with a 50% reduction in unanswered client queries.

Businesses and governments face similar struggles to handle increasing paperwork efficiently while accelerating their work. One IBM business partner created a new AI data discovery solution built on watsonx.ai using Llama to expedite large document processing. The solution helps officials find information, generate documents, and spend less time in the office and more in the field.

Here are just a few other examples of how developers are leveraging watsonx.ai and Meta’s Llama models across a wide range of use cases:

Multinationals

• A major telecommunications company significantly improved its HR assistant solution using watsonx.ai and Llama. The update led to a 75% containment rate, reducing the need for human intervention. The tool also expanded from English-only to supporting Italian and German, expediting the rollout of assistants to new markets.
• In a six-week generative AI pilot with IBM, a financial services company extracted key information and summaries from phone calls for bodily injury claims leveraging Llama2-13b-chat. The client created a tool to summarize victim calls that saved the handler an hour per case and allowed the victim to receive a copy of the report on the same day.

Governments

• In a pilot, a major European city’s local government found it could match or exceed Watson Language Translator’s capabilities in a Q&A assistant that translates and summarizes questions.
• Through a six-week legal research automation pilot that involved four judges and dozens of active court cases, a European city’s Ministry of Justice used generative AI to help simplify and speed up the judiciary’s work, particularly in the context of mass lawsuits. Using a RAG approach, the project resulted in an estimated 50% reduction in case preparation time.

Software as a service (SaaS)

• A major enterprise SaaS vendor for contract lifecycle management is partnering with IBM to embed watsonx.ai as the unified backbone for its platform. This solution, with Llama as the foundation model, will make generative AI contract processes available to everyone in the organization, allowing them to extract actionable insights from contract data.

Talent acquisition

• In a proof of concept, a major talent acquisition firm found that it could lower costs by 90% by switching its HR automation to watsonx.ai on IBM Cloud.

Finance

• Financial analysts and asset owners struggle to understand an event’s impact on an asset’s value. With watsonx.ai’s generative AI, an enterprise digital asset service provider can give asset owners a real-time valuation of their portfolio. Summarization of relevant events has enabled analysts to value a portfolio in five minutes instead of hours each day.
• In a four-week pilot, a major cloud banking and financial services platform worked with IBM Client Engineering to evaluate the impact watsonx.ai could have on its tools for monitoring bank transactions for suspicious activity—estimating more than 80% improvements in transaction insight generation.

Information management

• In a collaboration with IBM Client Engineering on an archive digitization and data cataloging use case that leveraged Llama, a major information management solutions provider used watsonx.ai to automatically classify documents and extract details while migrating projects to save time and gain new insights.

Media

• A Middle Eastern press agency is using watsonx.ai and Llama to save editors time and enhance their work. The organization has reduced redundant article reviews, automated translations, adapted articles for social media distribution and generated more compelling titles for stories. Editors saved up to 95% of their time previously spent on manual searches and reviews.

Tech and IT

• A major global technology and investment group worked with IBM to create a speech- and text-based Q&A chatbot using watsonx.ai and a RAG approach. This level of human-like Q&A will lead to more customer engagement and enable human customer service agents to optimize their time.
• To offload service desk agents from trivial tasks, a major European IT services provider and IBM developed a digital assistant solution that could search through its enterprise data to solve customer problems. The assistant reduced the duration of a support ticket from around three hours to mere minutes. An evaluation of the answers indicated 93% accuracy before any prompt tuning.