The Citizen Developer

Replaced by AI—Congrats on the Promotion!

• By Howard M. Cohen
• May 11, 2025

The conversation shifts again.

In recent months, we've shifted from focusing on low-code application platforms (LCAP) to artificial intelligence (AI) platforms that will build whole applications based on requirement descriptions given to them by citizen developers. We recently highlighted Mentor from OutSystems which integrates their popular LCAP with their new Mentor Generative AI platform. We explained, 'With OutSystems Mentor, developers can now build full-stack applications in minutes, without sacrificing quality, security, or governance."

In an even earlier column we suggested, "Don't Make Software, Make Agents" in which we reported how Microsoft CEO had announced "The Agentic World" at the recent Ignite conference and developers were now turning their attention toward making it easier for citizen developers to create agents, each of which could tackle a specific assignment completely.

Congratulations, Citizen Developer, You're Getting a Promotion
Those who fear that AI will replace people and eliminate their jobs may soon find themselves enjoying that fact.

It is said that the first responsibility of every great manager is to replace themselves so they can eventually move upward. Good succession planning demands it.

Now, citizen developers will be able to create their own "head count" of AI agents performing all manner of tasks for them. Those agents will require supervision as they perform over time. This defines the new role citizen developers can look forward to rising to. They will become supervisors, managers of their teams of AI agents.

Writing in a recent blogpost, AI at Work: How Human-Agent Teams Will Reshape Your Workforce Microsoft CMO of AI at Work Jared Spataro explains, "Everyone from interns to the C-suite will become an "agent boss" who oversees their own constellation of agents that power business processes." 

Spataro also points out the need to find, "the optimal ratio of humans to agents for whatever task or project your teams are working on." According to Spataro, strategically assigning specific tasks to AI agents will be of great value by "unlocking scale like never before."

Mick Costigan, VP of Salesforce Futures, takes it a step further in an interview with HR Executive, suggesting that citizen developers will find help in their new management responsibilities by turning to "'agent chiefs of staff'—AI assistants that coordinate multiple agents to optimize efficiency and reduce confusion for the user."

Emerging Platforms
In addition to OutSystems Mentor, many other platforms are emerging giving citizen developers the ability to create new agents, find accessible agents in growing marketplaces, activating those agents, integrating them, and supervising their ongoing operation. These include Microsoft Copilot Studio, IBM AgentLab, Writer, Agent.ai, Voiceflow, Relevance AI, Gumloop, Tars, Knolli, Dify, Langflow, Flowise AI, Vertex AI Agent Builder, Zenity, and more.

If this feels familiar, it is an elevated repeat of the emergence of low-code/no-code (LCNC) platforms a dozen years ago. As with that revolution, we will see some of these rise and become leaders while others eventually fall away. One of our biggest challenges will be to identify those that will reward our early investment in them.

Loving It When a Protocol Suite Comes Together
Two major announcements, which came in concert with each other, have dramatically accelerated this promotion of citizen developers to agent "bosses."

Agent-to-Agent (A2A) Protocol from Google enables agents to find other agents with needed capabilities and collaborate with them. Model Context Protocol (MCP) from Anthropic, adds the ability for multiple agents working together to access multiple large language models (LLM) while maintaining stability of context between them. Enabling agents to work together is a major advancement toward fulfilling Microsoft CEO Satya Nadella's vision of collaborating agents replacing traditional applications.

We can look forward to more new protocols being added to this nascent suite, bringing even more coordination and control to building teams of agents working together to achieve superior business outcomes.

Agentic Team Building
Microsoft's Spataro envisions the future, explaining, "Agents can research for you, provide expertise you don't have, or code for you. In the firm of the future, every employee (and every team) will manage a pool of them, with the exact number varying depending on their goals and preferences."

His advice continues: "Agents can research for you, provide expertise you don't have, or code for you. In the firm of the future, every employee (and every team) will manage a pool of them, with the exact number varying depending on their goals and preferences." Referring to management of these new citizen agent managers, Spataro suggests, "You'll also need to redefine roles and responsibilities. You might need new roles for overseeing agentic resources: tracking performance, leading deployment, and monitoring the human-agent balance. In a very dynamic labor market, employees and leaders who emerge as effective 'agent bosses' will likely get a leg up."  

Imagine a time in the near future when many professional and executive education providers offer courses on effective motivation and management of digital employees, guidance on how many digital workers a single citizen developer can effectively manage, policies regarding appropriate interaction with digital employees, and agentic performance bonus programs.

For today's citizen developers, this all begins with a survey of the agent supervision platforms including those mentioned earlier. Similar to the early days of LCNC, it's time to shop for the ideal agent platform. Enjoy!