Able to communicate and collaborate, AI agents are being developed for various consumer, enterprise, scientific, and industrial purposes. Highly complex environments, such as industrial plants or hospitals, might be served by multiple AI agents, each with a different purpose. According to a June 2024 Capgemini survey of 1,100 large enterprises, one in 10 organizations is deploying AI agents, with more than 50% planning to explore their use in the next year.
Key Highlights
- Leading tech companies are building upon existing AI assistants by creating platforms and tools to support developers. Specialist development companies are emerging with the expertise to work with large language models (LLMs) and the various machine learning models needed to implement AI agents. Agentic AI is expected to shape the future of DevOps (the set of practices around software integration and delivery).
- Combined with the ability to interact with both LLMs and the external environment, AI agents are empowered to execute more general-purpose work. Retrieval-augmented generation (RAG) is a design approach that aims to make LLMs more reliable by automatically retrieving information from external knowledge bases. It can ensure more accurate and contextually relevant outputs.
Scope
- Agentic AI refers to advanced AI systems that act autonomously, making decisions and taking actions with limited or no human supervision. An AI agent is a software program that interacts with its environment, collecting data to perform specific tasks, answer questions, and automate processes for users.
- This report provides analysis of this emerging AI technology, including an overview of how AI agents work, guidance on how enterprises can implement agentic AI, and agentic AI use cases by industry.
Reasons to Buy
- Although it is still early days for agentic AI, initial best practices are starting to emerge that can help enterprises that choose to develop their own agentic AI strategies or collaborate with agentic AI vendors. This report includes simple but essential steps that will help to streamline the creation and adoption of agentic AI into existing processes.
Table of Contents
- Executive Summary
- Agentic AI: The Journey Begins
- How Do AI Agents Work?
- An Evolving Ecosystem
- How Enterprises Can Implement Agentic AI
- Agentic AI Use Cases by Industry
- Case Studies
- Challenges for the Future
- Glossary
- Further Reading
- Thematic Research Methodology
Companies Mentioned (Partial List)
A selection of companies mentioned in this report includes, but is not limited to:
- Ada
- Adept.ai
- Affineon
- Agenttech
- AIQ
- Amazon
- AMD
- Anterior
- Anthropic
- Avantia
- Automation Anywhere
- AWS
- Axanol AI
- Bizdata
- Boosted.ai
- ChatDev
- Cisco
- Cohere
- Composable
- Concourse
- Connex AI
- Context.ai
- Counterpart
- Crew.ai
- Databricks
- DataRobot
- DeepSeek
- Deviniti
- Didero
- Elastic
- Ema
- Ericsson
- Finnomena
- Fourkites
- Ghostwriter
- Gretel
- H Company
- HappyRobot
- Harvey
- HCL Tech
- HSO.ai
- Harness
- Hippocratic.ai
- IBM
- Indem
- Indico Data
- Innovaccer
- Intuit
- K Health
- LangChain
- Legora
- Luminance
- Lumber
- Kanerika
- Maisa
- Meta
- MetaGPT
- McGill
- Microsoft
- Mistral
- NASA
- Netdocument
- Neutrinos
- Nokia
- Norm AI
- Nvidia
- Nuclia
- Open AI
- Optica
- Oracle
- Pando
- Perplexity
- Procore
- Project 44
- Rapidminer
- Relevance AI
- Salesforce
- SAS
- Servicenow
- Shopsense
- Spellbook
- Synthpop
- Sysdig
- Talkdesk
- Thoughtful.ai
- Toyota
- Twin Knowledge
- Trunk Tools
- University of Zurich
- Vercel
- Vida
- Vstrom
- Zapier