Today, that picture is starting to change. Modern platforms have lowered the barrier to entry by introducing:
- Simplified drag-and-drop interfaces with Lego-style modular building blocks
- Ready-to-use templates for common processes
- Robust governance features to define exactly what non-technical users can and cannot deploy to production
These improvements are making automation and app-building more accessible to business-side teams. But a new wave of capability is about to supercharge this movement: AI agents.
From citizen developer to AI-enabled citizen developer
The next evolution will see process owners and business analysts wield AI agents to design and execute processes simply by describing their intent. Why does this matter? Because process owners already know:
- What their processes need to achieve
- What steps should happen in what order
- What outcomes are required
What they don’t know is how to configure complex flows, write conditional logic, or manage API calls. AI agents change that by becoming an intelligent intermediary and turning natural language into executable workflows.
What is an AI agent?
An AI agent is a goal-driven software entity that can:
- Understand objectives in natural language
- Reason about the steps needed to achieve those objectives
- Use tools or “skills” to execute relevant actions in the correct sequence
Tools in this context can be as simple as functions, which take inputs, execute logic, and return outputs. They might:
- Call APIs
- Query databases
- Perform calculations
- Trigger actions in external systems
Agents select the right tool at the right time based on a description of what the tool does, when it should be used, and how to use it.
The analogy is simple: think of a handyman and their toolbox. The handyman knows the job, selects the right tool, and uses it in the correct order.
AI agents inside automation flows
Many automation platforms are now starting to allow AI agent steps in workflows. At a given stage in your process, you can hand over control to an AI agent that has a defined set of tools at its disposal. The agent then determines the right sequence of actions based on your high-level instructions.
Putting it all together – A practical example
Let’s imagine an organisation that sets up a catalogue of AI agents, each with:
- A clearly documented purpose
- A list of tools available for it to use
- Descriptions of how and when to use each tool
A citizen developer or process owner can now design processes by assembling these building blocks, without knowing the technical details behind each tool.
Example: A simple order processing flow
Business requirement when an order is received:
- Check if the customer is valid and passes credit checks
- If not, cancel the order and notify the customer
- If valid, check stock for all items
- If any item is out of stock, cancel and notify
- Otherwise, process the order
Traditional approach
A developer explicitly wires together API calls, data mappings, conditional logic, and message templates.
AI agent approach
We have an order agent with the following skills:
- checkCustomerRecord – validates the customer across multiple systems
- checkStockAvailability – checks if each ordered item is in stock
- processOrder – triggers order fulfilment and generates a confirmation
- cancelOrder – cancels the order and sends a message to the customer
Instead of writing logic, the business user passes the order and a prompt like this to the agent:
“This is an order. Check if the customer is valid and passes credit checks. If valid, check stock for all items. All items must be in stock to process the order. If any check fails, cancel the order and notify the customer. Otherwise, process the order.”
The agent takes it from there deciding which tools to call, in what order, and with what parameters.
Why this matters
With AI agents, the gap between process knowledge and technical execution shrinks dramatically. Process owners no longer need to understand system APIs or complex branching logic. They just need to describe what they want to happen—and the agent figures out the rest.
Certainly, there will be considerations around cost, performance, and governance. But the potential is clear:
- Greater self-sufficiency for business users
- Faster time-to-automation
- Reduced dependency on specialised technical resources
In short, AI agents are not just another tool, they represent a major shift in how automation is designed, deployed, and controlled.
The citizen developer era has been promised for over a decade. AI agents might finally be the catalyst that makes it real.