What does an AI agent actually do once you hand it a task? In this episode, I break down how AI agents work, how they differ from chatbots, and why briefing and supervising them is becoming one of the most important workplace skills you can build right now.
The Numbers That Should Get Your Attention
By the end of 2026, 40% of enterprise applications will have task-specific AI agents built in. That is Gartner’s prediction, up from less than 5% in 2025. And it does not stop there. By 2027, around a third of agentic implementations will combine multiple agents working together on complex tasks. By 2029, at least half of all knowledge workers will need skills to work with, govern or create agents on demand.
Gartner also warns that more than 40% of early agentic AI projects could be cancelled by 2027 due to rising costs, unclear value or weak risk controls. That is exactly why the human skills around briefing and supervision matter just as much as the technology itself.
The Difference Between Answering and Doing
Here is the distinction that changed how I think about all of this. A chatbot answers your question and then waits for the next one. You do all the thinking in between. An AI agent takes a goal, breaks it into steps, uses tools to carry out those steps, and checks its own work along the way. You supervise the outcome rather than every single keystroke.
An agent runs a loop with four steps:
- Perceive — it reads whatever you have given it access to
- Plan — it decides the steps needed to reach the goal
- Act — it uses tools like calendars, spreadsheets, search engines or document editors
- Check — it reviews its own output before flagging anything to you
A Real-World Morning for a Comms Team
Let me make this concrete. Picture a communications team with an agent set up on their shared inbox. At 8am, the agent reads the 40 messages that arrived overnight — that is the perceive step. It plans: three media queries need holding responses, two are meeting requests, one is a complaint that must be escalated, and the rest are newsletters to file.
Then it acts. It drafts the three holding responses using the team’s approved lines and saves them as drafts for a human to review. It checks the diary and proposes meeting slots. It files the newsletters. The complaint? It does not touch it. It flags it to the duty press officer, because that boundary was set on day one.
Finally, it checks — confirms every draft names the right journalist, notes one query it was unsure about, and posts a summary to the team channel. That is a full morning’s triage done before anyone sits down, with a human still approving every word that leaves the building.
How I Think About Agents After Two Years of Training Teams
My honest take: an agent is a capable junior colleague. It needs a clear brief, defined boundaries, and someone reviewing the work. Give it those three things and it earns its place on the team. Skip them and you have handed the keys to an eager intern with no manager.
Anthropic’s Economic Index found that 45% of conversations on Claude are automated — the person hands over a task and lets the AI complete it — while 52% are collaborative. Both modes have their place. The skill is knowing which task belongs in which mode.
My test since my Harvard programme has always been this: AI plus me must be greater than me alone. If it is not, we do not deploy it. The agent handles execution. You hold the judgement, the context and the accountability.
If you want to build your AI literacy and skills, I can help
📚 Grab AI Playbook on Amazon, the perfect podcast companion.
📰 Subscribe to Human in the Loop, my LinkedIn Newsletter
🎓 Take the CPD Accredited AI Playbook Diploma
📞 Book a call at digitaltraining.ie
Sources
Gartner: Gartner Predicts 40% of Enterprise Apps Will Feature Task-Specific AI Agents by 2026. https://www.gartner.com/en/newsroom/press-releases/2025-08-26-gartner-predicts-40-percent-of-enterprise-apps-will-feature-task-specific-ai-agents-by-2026-up-from-less-than-5-percent-in-2025
Gartner: Over 40% of Agentic AI Projects Will Be Canceled by the End of 2027. https://www.gartner.com/en/newsroom/press-releases/2025-06-25-gartner-predicts-over-40-percent-of-agentic-ai-projects-will-be-canceled-by-the-end-of-2027
Anthropic: Anthropic Economic Index report — Economic primitives, January 2026. https://www.anthropic.com/research/anthropic-economic-index-january-2026-report
Key Statistics
– By end of 2026, 40% of enterprise apps will have task-specific AI agents built in, up from less than 5% in 2025 (Gartner).
– By 2027, around a third of agentic implementations will combine multiple agents on one complex task (Gartner).
– More than 40% of early agentic AI projects could be cancelled by 2027 due to costs, unclear value or weak risk controls (Gartner).
– By 2029, at least half of all knowledge workers will need skills to work with, govern or create agents on demand (Gartner).
– 45% of conversations on Claude are automated — users hand over a task and let the AI complete it (Anthropic Economic Index, January 2026).
– 52% of Claude conversations are collaborative, with humans and AI working together (Anthropic Economic Index, January 2026).
About the Podcast Host
AI and Communications Specialist | Author | Podcaster | International Speaker
Joanne Sweeney is a recognised leader in digital marketing and digital transformation, specialising in the practical application of AI to enhance communication strategies. The founder of Digital Training Institute, Joanne, empowers leaders and communications professionals to leverage AI responsibly and effectively, ensuring transparency, trust, and tangible results.
She completed the Harvard Kennedy School Executive Programme in AI and Public Policy this summer before launching her fourth book AI Playbook, in September. Joanne has also published two editions of Public Sector Marketing Pro, which were preceded by her first title, Social Media Under Investigation Law Enforcement and the Social Web.
Her books have become go-to resources for governments, public sector and policing agencies, as well as marketing and communications professionals, all navigating the challenges of the changing Internet.
Joanne is the host of the AI SIX Podcast, a daily six-minute podcast that provides actionable insights, expert interviews, and practical advice to help professionals navigate the new world of work.
