In a recent LinkedIn post, Andrew Bolis explores the evolving landscape of artificial intelligence in the workplace, specifically distinguishing between AI agents and what he terms “smart workflows.” Bolis emphasizes that many organizations believe they are leveraging advanced AI when, in reality, they are often utilizing more rudimentary automated processes.
Understanding the Nuances of AI Implementation
Andrew Bolis highlights a common misconception regarding AI adoption, noting that the term AI is frequently applied to systems that are not truly intelligent agents. He points out the prevalence of AI in business, citing a statistic that suggests “By 2027, 82% of organizations will use AI agents (as reported by Auth0).” However, Bolis clarifies that this figure might encompass more than just genuine AI agents.
Differentiating Workflows: From Manual to Agentic
To illustrate the differences, Bolis uses the example of scheduling a quarterly business review with a marketing client. He breaks down the process into three distinct categories:
A. Non-Agentic Workflows
In this scenario, Bolis describes a process where humans are central to every action, using tools based on their direct commands. This involves manual steps like checking CRMs, drafting emails with AI assistance (e.g., asking ChatGPT to “write an email to schedule a client call”), and sending them manually. As Andrew Bolis notes, “Humans think. AI simply delivers output based on the user’s request.” This highlights that while AI tools can assist, the human remains the primary driver of the workflow.
B. Agentic Workflows
Andrew Bolis then defines agentic workflows as systems that operate with a defined set of rules and triggers. An example provided is a calendar detecting a 90-day period has passed since the last review. The system then automatically checks the manager’s availability, drafts an email, and sends it. While time-saving, Bolis cautions that these workflows “can’t adapt when conditions change.” This implies a limitation in flexibility and proactive decision-making.
C. AI Agents
The most advanced category, according to Bolis, is AI agents. These entities are defined by having a clear, overarching goal, such as “Keep clients happy.” An AI agent, as described by Bolis, would autonomously monitor client interactions, recognize the need for a review, and proactively manage the entire process. This includes sending tailored emails, following up, confirming the meeting, and even creating an agenda. Bolis argues that AI agents are “intelligent, focused on outcomes, not just individual tasks.” This distinction emphasizes their ability to understand context, make decisions, and achieve objectives with minimal human intervention.
The Future of AI in Business Operations
Andrew Bolis’s analysis suggests a significant shift in how businesses will interact with AI. Moving beyond simple task automation, AI agents promise a more integrated and intelligent approach to operations. By clearly delineating between automated workflows and true AI agents, Bolis provides a valuable framework for understanding the current and future capabilities of AI in the professional sphere.
The post concludes with a call to action, encouraging readers to follow Andrew Bolis for more insights and to repost the information.
📝 About This Content
This article is based on insights shared by Andrew Bolis on LinkedIn.
📅 Originally posted on November 7, 2025 | View original post on LinkedIn →