How agents are upending the way we get work done
Summary
IBM Research's VP of AI and Automation, Nick Fuller, discusses the evolution and future of enterprise automation, emphasizing the transformative impact of agentic AI and large language models (LLMs). While automation has existed for decades, AI, particularly LLMs and agentic systems, has significantly boosted productivity and ROI, moving beyond peripheral applications to integrate deeply into workflows. IBM has released agent-based solutions for compliance, observability, and application insight, with anecdotal client feedback indicating substantial performance improvements. The discussion highlights the critical need for domain-specific benchmarks like IT bench and Asset Ops Bench, built on 150+ enterprise-grade scenarios, to provide clarity for customers amidst rapid technological advancement. These benchmarks demonstrate the superior performance of IBM's horizontally innovative agents compared to competitive offerings, particularly in areas like graph-aided root cause analysis and code analysis.
Key takeaway
For CTOs and VPs of Engineering evaluating AI investments, recognize that agentic AI, powered by LLMs, offers a path to "hyperautomation"—always resilient, secure, and compliant systems. Prioritize solutions that integrate deeply into mission-critical workflows and demonstrate performance on domain-specific, enterprise-grade benchmarks. This approach will accelerate application deployment from spec to day-two operations, optimizing costs and resource use, and ultimately enabling your teams to focus on strategic initiatives rather than operational toil.
Key insights
Agentic AI and LLMs are transforming enterprise automation by integrating deeply into workflows, driving significant productivity and ROI.
Principles
- Automation value comes from workflow integration, not peripheral AI.
- Domain-specific benchmarks are crucial for evaluating enterprise AI agents.
- Horizontal innovations differentiate agent performance in competitive fields.
Method
IBM's approach involves building agentic AI systems underpinned by horizontal innovations, such as graph-aided root cause analysis, and validating them against enterprise-grade, domain-specific benchmarks like IT bench and Asset Ops Bench.
In practice
- Implement agentic AI for core workflow changes, not just conversational interfaces.
- Utilize domain-specific benchmarks to assess AI agent performance.
- Explore agent-based solutions for IT automation, asset resiliency, and compliance.
Topics
- AI Agents
- Enterprise Automation
- AI Benchmarking
- Large Language Models
- Hyperautomation
Best for: CTO, VP of Engineering/Data, Executive, AI Engineer, MLOps Engineer, Director of AI/ML
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by IBM Research.