Using a citizen developer program to boost AI deployments
In this podcast, Fabien Cros explains how consulting firm Ducker Carlisle gave business users low-code tools to jumpstart AI development, cut costs and free up IT staff time.
The idea of citizen developers sounds promising -- nontechnical workers employing user-friendly programming tools to write and customize software -- but how realistic is it?
At Ducker Carlisle, a management consulting firm, the results in both user adoption and bottom-line results are very real. Late last year, it started a citizen developer program when IT's data science and AI specialists became overwhelmed with requests for AI applications. Around 80 of the 200 employees signed up from a broad range of departments, including research, sales, HR and finance.
The resulting AI apps automated dozens of tasks, cutting operating costs by 3%, breaking the AI logjam and freeing IT staff for other projects, according to the company.
In this episode of Enterprise Apps Unpacked, Fabien Cros, Ducker Carlisle's chief data and AI officer, explains how the citizen developer program works and offers tips on starting one.
Fabien Cros
Building a cyber workforce
The commercial low-code/no-code development tools typically provided to citizen developers have been around since the early 2010s, but their origins go back to the visual programming languages introduced several decades earlier.
Cros acknowledged that low-code tools often failed to deliver on their initial hype because they couldn't handle complex programming jobs. The arrival of generative AI in late 2022 caused an "earthquake," he said, with large language models capable of generating sophisticated code. "Now you can ask anyone that is non-technical to describe something, and you have an output that is pretty good."
But Cros was careful not to overstate the level of programming carried out by Ducker Carlisle's citizen developers. "We are not trying to rush directly into business users creating full enterprise software with an SLA [service-level agreement] and Kubernetes services. That's way too advanced."
The real aim is to help business users build enough AI agents to create a "cyber workforce" that supports the human workforce. "If we can extract all of the value just from doing that, it's a massive, massive opportunity for our organization," he said.
However, agents and workflows that catch on with other users and perform critical functions can be adopted by IT and turned into an asset for the company, with proper coding for front and back ends and infrastructure. "We will take it out of the third-party platform and build it directly, using our own tech," Cros said.
Other topics discussed in the podcast include the following:
- Essential elements of a citizen developer program.
- Examples of agentic AI apps developed through the program.
- Whether the approach changes the traditional relationship between in-house developers and business users.
- How Ducker Carlisle's program could evolve in the future.
David Essex is an industry editor who creates in-depth content on enterprise applications, emerging technology and market trends for several Informa TechTarget websites.
