After seeing demos of agentic AI, I wonder how many software engineers are ready for this paradigm shift. AI agents developing applications, APIs, and integrations is a major step forward in simplicity and capability. And it will leave many old coding and software development methods behind. Much like how assembly coding, Fortran, and Cobol are in the rearview mirror.

I know there’s plenty of Cobol out there, and assembly is used for embedded systems. But these are small and specialized use cases.

AI Agents are transforming low-code experiences

Last summer, I wrote how my team used genAI to develop this blog and migrate 700+ blog posts. GenAI was far from perfect, and most coding required several iterative prompts to achieve the desired results. None of us code JavaScript or can configure cloud-based redirect proxies. And we wouldn’t have been able to without genAI.

The State of Web Dev AI 2025 reports that 91% of developers use AI for code generation. According to another report, DevOps teams accept between 20% and 35% of the code recommendations.

So there’s no doubt that developers are catching on to using genAI for code development. But that’s just the beginning of how agentic AI is transforming software development.

Code generators are great productivity boosters. But agentic AI software development innovations from no-code and low-code development tools showcase the future of app development.

For example:

• Adobe Experience Platform Agent Orchestrator supports the integration and collaboration of several customer experience agents, such as audience, content production, and data engineering agents.
• Appian AI agents help organizations build apps, optimize business processes, and track actions in real-time.
• Quickbase app builder provides an industry-oriented agent for developing no-code applications in manufacturing, construction, government, and other professional services.

In my last post, I shared details from Appian World about their app composer and AI agents. The composer allows building and iteratively improving the app’s UI and data without coding. The developer can publish the app and then use AI to work on new releases with improvements. Not only is this low-code, it’s a genAI-first experience that illustrates how organizations will plan, build, and enhance applications.

Agentic AI software development: A significant paradigm shift

I wrote an article for InfoWorld last year on how generative AI will change low-code development. I asked whether genAI embedded in low-code would help organizations modernize applications faster, improve software quality, and transform the developer skillset.

One year later, I wonder if my questions weren’t bold enough. Perhaps we’re seeing the beginning of the end of coding as we know it? Maybe, the evolution of low-code, no-code, and process automation is a converged experience, driven by genAI, and requiring even less of a developer’s skillset.

And my question is, are engineers ready for agentic AI software development?

AI Agents beyond code generators and SDLC accelerators

Consider three predictions on how agentic AI software development may evolve.

AI Agents build real-time data visualizations

“Low-code platforms are entering a new era with GenAI at their core—enabling faster application development, intelligent automation, and deeply personalized user experiences,” says Rajan Nagina, head of AI practice at Newgen. “These platforms empower enterprises to make agile, data-driven decisions by learning from data patterns and offering real-time insights to meet the demands of today’s dynamic business environment.”

My translation: Will citizen data scientists still create data visualizations? Maybe not. Data visualizations will still be needed to help people make decisions, but genAI will be developing them based on the question or decision being made at that time.

AI Agents reducing technical debt

“I think there are a ton of emerging generative AI features that are making huge impacts on organizations outside of simple productivity improvements, but if I had to select one, it would be the ability for generative agents to follow heuristics and leverage tools, just as human workers do today,” says Simon Margolis, associate CTO of AI and ML at SADA. “An agent can have a ‘feeling’ that a certain approach is best and then act on it, validate its assumptions, and make real changes.”

My translation: Will genAI be able to analyze code –  their code developed by an AI agent or legacy code created by developers – and recognize when the code smells bad and there is technical debt rotting below the service?

AI Agents improving business process automations

“Over time, agentic AI will transform low-code platforms into true business partners — learning from day-to-day operations and helping businesses move from reactive to predictive, adaptive models,” says Sunil Senan, SVP and global head of data, analytics, and AI at Infosys. “As these AI agents mature, they will not only scale enterprise and human expertise but also drive autonomous decision-making across increasingly complex environments.”

My translation: Will GenAI monitor workflows and user experiences and self-improve them without developer involvement?  

The future role of software developers in an agentic AI world

GenAI’s ability to autonomously develop and improve applications without a human-in-the-middle or human-at-the-helm remains a future projection. But it’s hard to guess if and when these capabilities will emerge. Which use cases will they provide early and easy business value? What will the roles of product owners, software developers, QA engineers, or SREs look like in this agentic AI future?  

“Teams will move faster, adapt better, and innovate more consistently, and businesses that embrace this shift will be better equipped to thrive in a constantly changing environment,” suggests Senan of Infosys.

Call me somewhere between a pragmatist and a pessimist, but I don’t see The Matrix as a near-term evolution of software development.

However, I see a paradigm shift where genAI is a force multiplier for organizations that use or move to low-code development tools. There will be more citizen development, and genAI will take on more complex tasks in app migration. Agentic AI software development may not be sci-fi, but it will be a game changer.

I love software development, but the skill and craft are changing rapidly.