The New GeneXus: Modeling Intent, Not Just Writing It
In the new GeneXus, the conversation does not directly produce code. It produces a formal, versioned model; code is one of its projections.
Almost forty years ago, GeneXus promised future-proof technology. Keeping that promise today means helping companies become agentic organizations based on what they have already built.
And that is the starting point of GeneXus for Agents: the set of skills, tools, and interfaces that allows AI agents to understand and operate on organizations’ Knowledge Bases (KBs), and leverage that digital asset that has years of evolution behind it.
Does it work? We asked some developers who are already working with GeneXus for Agents on real KBs what their impression of the product is. These are their responses:
“It is really very good and I have already adopted it into my way of working.”
[Read the use case: Labor evaluation system with GeneXus for Agents and Claude Code]
“I have found it very addictive and it brought me feelings similar to those from 30 years ago, at the first Distributors Technical Meeting, when we tried GeneXus for Windows for the first time.”
“I have been following GeneXus for Agents since the first versions available. The proposal struck me as right from the beginning. I always argued, and communicated it in the community, that having something like this provided by the manufacturer itself was essential. Community solutions are valuable, but they cannot on their own sustain the spread of a culture of using GeneXus with AI. That needs to come backed by the product.”
“I felt excited, with many positive expectations.”
“Regarding my first impression of using GeneXus for Agents, it seemed like a very powerful tool, although with a configuration learning curve. After the MCP server became available, it became much easier to perform analyses and pre-execute tasks.”
“I took the initiative to bring this in, and when it worked I was very happy hahaha. With use, I could see that it had a future and that it could help the team quite a bit, despite the limitations it had at the time and some bugs that could occur.”
“My first impression was that it seemed to be a product that would allow us to speed up analysis and development timelines significantly. That first impression became a reality for most of the small requirements we worked on; however, the product has areas of opportunity that we will continue exploring to maximize its value.”
“It seemed like something good, to make it easier to translate ideas for changes or things I want to implement into something concrete more quickly than with manual coding.”
“I am just starting to try it out, and the first impressions are good. I applied it to a web application in production (first tested in a testing environment) to solve a specific problem with a payment form. I asked it to analyze the KB – I personally use Claude Cowork – and I was able to add the GeneXus for Agents MCP to Cowork. I am now applying it to a more complex project: totem-based sales, real-time stock management and stock reservations, and interaction with other sales systems that share that same stock.”
GeneXus for Agents is available at no additional cost for all customers who are up to date with maintenance and use GeneXus 18.
If your KB is on an earlier version, GeneXus for Agents can also help you with the migration. And once migrated, it is available to keep expanding your solution.
And if you are already using GeneXus for Agents and want to share your experience, you can leave your comment below or write to me at warambule@genexus.com. The next opinions we publish could be yours.
The New GeneXus: Modeling Intent, Not Just Writing It
GeneXus and Neuro-Symbolic Architecture
The Problem with Prompt-Based Development
GeneXus for Agents: 15 key questions answered
Leave a Reply