Analyst(s): Keith Kirkpatrick
Publication Date: July 25, 2025
Many of today’s agents are marketed as agentic, but most fall short of those promises. But given the pace of innovation and the cautious nature of customers, that may not matter in the long run.
Key Points:
- While 2025 has elevated “agentic AI” to the forefront of enterprise software—with vendors universally marketing autonomous AI agents—the majority of offerings fall short of full autonomy.
- Despite heavy vendor promotion, agentic capabilities rank only around 8–10% as a top vendor selection criterion (rising to 15% among the largest enterprises), lagging behind features/functionality and generative AI, signaling buyers still prioritize proven value drivers over emerging autonomy.
- Major SaaS players (Microsoft, Google, Salesforce, Zendesk, Zoho, ServiceNow, SAP, Oracle) offer agentic tools—ranging from customizable frameworks to domain-specific bots—but most require human-configured goals, rule-based oversight, and manual updates.
- Enterprises and vendors alike are adopting a crawl‑walk‑run strategy around agentic AI.
Overview:
Agentic AI has emerged as the defining theme of 2025 enterprise software, building on 2024’s generative‑AI momentum. While nearly every major vendor now touts “agentic” capabilities—systems that operate independently, adapt to changing conditions, initiate actions, and learn over time—the reality often falls short of this promise. For buyers, understanding where vendors truly deliver autonomous, self‑optimizing behavior versus where they merely wrap LLM‑driven workflows in marketing gloss is critical to making informed purchasing decisions.
Market Context and Buyer Priorities
Futurum’s mid‑2025 survey of 895 decision-makers shows agentic AI capabilities cited by just 8.8% as a top‑3 selection criterion overall (up slightly from 8% in 2024).
- Larger enterprises (annual revenue >$25 billion) place greater emphasis on agentic features (15%), matching their focus on core functionality. Smaller organizations cite it in the 7–10% range.
- Traditional buyer priorities—features/functionality, generative AI, cost/TCO, customization, and implementation speed—remain dominant.
Vendor Offerings Today
Most SaaS platforms provide LLM‑backed “agents” that can invoke tools and orchestrate multi‑step workflows, but they require substantial human setup, supervision, and manual retraining:
- Microsoft Azure AI & Copilot: Offers “autonomous agents” in documentation, yet actual tools (GitHub and 365 Copilot) act as assistants; Azure AI Foundry demands goal‑setting, tool integration, and policy enforcement by developers.
- Google Vertex AI: Supplies building blocks for chaining LLM calls and API integrations, but agents lack true self‑improvement.
- Salesforce Agentforce, Zendesk AI Agents, Zoho Zia Agents, ServiceNow AI Agents, SAP Joule, Oracle AI Agents: All promise varying degrees of autonomy—planning, decision‑making, skill libraries—but still depend on human‑defined rules, dashboards, and periodic tuning rather than real‑time learning or weight updates.
Key Shortcomings
Today, AI agents don’t self‑optimize; model weights remain frozen, and “learning” comes from manual rule or prompt updates. Any adaptation in agentic behavior is manually configured, as improvements require human‑driven configuration. In many cases, actions pause for human approval rather than executing fully autonomously. Most notably, human oversight and governance layers remain essential.
Outlook
Enterprises will adopt a “crawl‑walk‑run” approach—piloting semi‑autonomous agents under tight controls before entrusting them with unsupervised workflows. True agentic AI with safe, online self‑learning is likely still quarters or years away, as vendors balance innovation against risk. What to watch next: broader real‑world action‑oriented use cases, unified agent‑management platforms, and the gradual introduction of controlled self‑optimization in production environments.
The full report is available via subscription to Futurum Intelligence’s Enterprise Software & Digital Workflows IQ service—click here for inquiry and access.
You can read a news release from Salesforce touting its success with Agentforce on its own Help Platform.
Enterprise Software & Digital Workflows
Futurum clients can read about it in the Futurum Intelligence Platform, and non-clients can learn more here: Enterprise Software & Digital Workflows Practice.
About the Futurum Enterprise Software & Digital Workflows Practice
The Futurum Enterprise Software & Digital Workflows Practice provides actionable, objective insights for market leaders and their teams so they can respond to emerging opportunities and innovate. Public access to our coverage can be seen here. Follow news and updates from the Futurum Practice on LinkedIn and X. Visit the Futurum Newsroom for more information and insights.
Author Information
Keith Kirkpatrick is Research Director, Enterprise Software & Digital Workflows for The Futurum Group. Keith has over 25 years of experience in research, marketing, and consulting-based fields.
He has authored in-depth reports and market forecast studies covering artificial intelligence, biometrics, data analytics, robotics, high performance computing, and quantum computing, with a specific focus on the use of these technologies within large enterprise organizations and SMBs. He has also established strong working relationships with the international technology vendor community and is a frequent speaker at industry conferences and events.
In his career as a financial and technology journalist he has written for national and trade publications, including BusinessWeek, CNBC.com, Investment Dealers’ Digest, The Red Herring, The Communications of the ACM, and Mobile Computing & Communications, among others.
He is a member of the Association of Independent Information Professionals (AIIP).
Keith holds dual Bachelor of Arts degrees in Magazine Journalism and Sociology from Syracuse University.
