IBM Embraces Multi-Model AI Strategy to Match Large Language Models with Targeted Enterprise Use Cases

IBM has observed a significant shift in how enterprise customers are approaching AI technology, particularly in the realm of generative AI. Over the last 100 years, IBM has navigated numerous tech trends, and the current trend in AI is no exception. At VB Transform 2025, Armand Ruiz, Vice President of AI Platform at IBM, highlighted that the key to success in AI lies not in selecting a single large language model (LLM) provider but in adopting a multi-model strategy that matches specific LLMs to targeted use cases. The Multivendor Approach Enterprise customers are increasingly rejecting the idea of being locked into single-vendor AI solutions. Instead, they are using a variety of LLMs, often from multiple providers, to address different business needs. For instance, Anthropic’s models might be preferred for coding tasks, while others like o3 are favored for reasoning. For LLM customization and fine-tuning with proprietary data, enterprises might use IBM’s Granite series or smaller models like Mistral or Llama. IBM’s position in this landscape is evolving from a foundation model provider to a control tower for AI workloads. They have introduced a new model gateway that offers a single API to switch between different LLMs, ensuring observability and governance across all deployments. This technical architecture allows customers to run open-source models internally for sensitive tasks and access public APIs like AWS Bedrock and Google Cloud’s Gemini for less critical applications. Agent Orchestration Protocols To address the growing complexity of agent-to-agent communication, IBM has developed the Agent Communication Protocol (ACP) and contributed it to the Linux Foundation. This protocol, similar to Google’s recently donated Agent2Agent (A2A), aims to standardize interactions between AI systems, reducing the need for custom development. Ruiz believes that these protocols will eventually converge, simplifying integration at enterprise scale. IBM customers already have over 100 agents in pilot programs, and without standardized communication, managing these interactions would be unsustainable. Transforming Workflows with AI Ruiz emphasizes that simply adding chatbots to existing systems does not represent true AI transformation. Instead, he sees AI as a tool for fundamentally restructuring business workflows and automating tasks. IBM’s own HR department exemplifies this shift: specialized agents now handle routine queries about compensation, hiring, and promotions, interfacing with internal systems and reducing the need for human intervention. This change signifies a move from human-computing interaction to computer-mediated workflow automation. AI agents are being designed to execute multi-step processes independently, only escalating to human handlers when necessary. Strategic Implications for Enterprise AI Investment IBM’s findings suggest several critical changes for enterprise AI strategies: Move Beyond Chatbot-First Thinking: Enterprises should focus on transforming entire workflows using AI, rather than merely enhancing existing systems with chatbots. The goal is to automate and streamline processes to eliminate redundant human steps. Architect for Multi-Model Flexibility: Integration platforms that support seamless switching between different LLMs based on specific use cases are essential. This approach ensures that enterprises can maintain governance standards while leveraging the best models for each task. Invest in Communication Standards: Supporting emerging protocols like ACP and A2A is crucial to avoid vendor lock-in and ensure that AI systems can communicate effectively across different platforms and vendors. Industry Evaluation and Company Profile Industry insiders view IBM’s multi-model and protocol-driven approach as a strategic move that aligns with the evolving landscape of enterprise AI. By not forcing customers into proprietary ecosystems, IBM is positioning itself as a versatile and future-proof partner. This strategy also highlights IBM’s commitment to open standards and collaboration, which is increasingly valued in the AI community. IBM, known for its long history of innovation in technology and computing, continues to adapt to new paradigms. The company’s AI Platform division, led by Ruiz, is at the forefront of developing scalable and flexible AI solutions that address the diverse needs of enterprise customers. IBM’s recent investments and contributions to open-source protocols underscore its role as a thought leader in the AI space, helping businesses navigate the complexities of AI adoption and integration. Overall, IBM’s approach reflects a pragmatic and forward-thinking mindset, recognizing that the future of AI in enterprises is about versatility, standardization, and deep integration into business processes.