Adam Danyal’s post on LinkedIn got my attention. He says, most leaders think they are building AI Agents when they are actually building fancy chatbots. So you need to understand the tiers of AI capability.
Building AI maturity is something I talk about a lot. In 2026, decide your AI ambition.
- Level 1 – Model: generates content (good for drafting and ideation, low real-world permissions).
- Level 2 – Assistant with tools: can execute bounded actions you approve (good for booking, pulling reports, posting drafts).
- Level 3 – Autonomous workflow system: can decide steps, run loops, and coordinate multiple actions (high leverage, highest governance needs).
In my experience, most organisations aren’t stuck at Level 2 because the tech is not ready. They’re stuck there because trust, controls, and accountability are not ready.
So let me share how to move through the tiers.
Tier 1 to Tier 2 is the permissions jump
Moving from Tier 1 to Tier 2 is the moment AI stops being only a content tool and starts interacting with your real systems, which means it can take actions rather than just produce text.
That shift is powerful because it moves work forward, but it also changes the risk profile, since an error can now land in someone’s inbox, calendar, CRM, or social queue.
Most organisations stall here because they worry about mistakes, brand damage, and data access, and because nobody has a clear mandate to own the workflow end-to-end.
They also get stuck when the organisation has not agreed on what the AI is allowed to touch, which tools it can use, and what needs a person to approve before anything happens.
To push through, start with a shortlist of safe, repeatable tasks where the inputs are predictable, and the outputs are easy to review, like booking meetings within fixed constraints, pulling standard reports, turning transcripts into clips, tagging content, or generating UTMs.
Then put human approval points in the workflow so the AI can draft and prepare, but a person presses send, publishes, or commits changes, because that is how you build trust without pretending the tool is flawless.
Finally, track a small set of success metrics such as time saved, error rate, and rework required, because hard evidence is what earns you permission to expand access and move to the next tier.
Tier 2 to Tier 3 is the autonomy jump
Moving from Tier 2 to Tier 3 is not mainly about adding more tools; it is about giving the system the ability to decide what to do next in pursuit of a broader goal.
At Tier 2, you define the task, and the agent executes it, but at Tier 3, you define the outcome, and it plans, sequences steps, checks progress, and adapts when something fails.
Organisations stall here because errors can compound across multiple steps, and because accountability becomes murky when the system is making decisions rather than following a checklist.
They also run into compliance and privacy concerns, cost and complexity, and the reality that most tool stacks are messy, poorly integrated, and hard to govern as one coherent workflow.
To push through, you need clear goals and explicit stop rules so the system knows when to halt and escalate, especially when data is missing, uncertainty is high, or an action could create reputational or legal risk.
You also need audit logs and traceability, plus evaluation methods that score both outputs and actions, so you can see what it did, why it did it, and whether it met your standards.
And you must lock down data governance and role clarity, because persistent state means storing context over time, and autonomy only works when someone owns the agent, someone approves key decisions, and someone monitors performance and incidents.
So what’s your 2026 AI ambition?
Do you want more support with AI in your organisation, or to boost your career? Get in touch.
Book a call in my calendar.
Get a copy of AI Playbook.
About the Podcast Host
Joanne Sweeney is a specialist in public-sector marketing and digital communications. She’s the author of Public Sector Marketing Pro and the creator of the Professional Diploma in AI for Public Sector Marketing. With years of experience in training, consulting, and advocating for digital transformation, Joanne provides insight and guidance for government agencies navigating the digital landscape. She is also the host of the award-winning Public Sector Marketing Show podcast.
