In this episode of The LowCode Podcast, we explore how AI is changing the economics of software development, from how much it costs to build an app to who actually owns it. We share why we at LowCode Agency made the decision to move away from no-code platforms like Bubble and Glide toward almost exclusively building custom software with AI, and what changed to make that possible.

We break down how advances in AI models like Claude Opus have made custom development faster and more affordable, turning projects that once took a year into builds that can now be completed in three to four months. You'll learn why this shift matters for businesses looking to scale without platform limitations, avoid vendor lock-in, and maintain full ownership of your software. We also discuss the challenges behind our transition, including restructuring our development team, and why no-code still makes sense in certain situations.

If you're planning to build an app, evaluating your current tech stack, or wondering how AI should influence your software strategy, this episode offers a practical perspective on making smarter development decisions. Because at the end of the day, the goal isn't to build with no-code or custom code. It's to choose whatever delivers the best outcome for your business.

In this episode of The LowCode Podcast, we tackle one of the biggest challenges with AI automation: Building the technology is easy compared to getting people to actually use it. Why do so many AI workflows fail to gain traction, even when they solve real business problems?

We break down how our team built an AI-powered sales follow-up system that reviews 120 active deals every morning, drafts personalized messages, and lets employees approve or edit them directly from Slack or the CRM. You'll learn why integrating AI into the tools people already use is more effective than forcing them to adopt new platforms, how human oversight fits into automated workflows, and how to keep your CRM updated without adding extra work.

If you're building AI automations for your business or struggling to get your team to adopt new tools, this episode offers practical insights into designing workflows around how people actually work, not the other way around.

Running more AI agents doesn’t automatically mean you’re running a better business. In this episode of The LowCode Podcast, we break down why the number of agents in your stack matters far less than what those agents actually know, what they’re responsible for, and how closely their work is supervised. A network of dozens of agents with limited context can quickly create more monitoring, confusion, and risk than real leverage.

We also look at the much smaller system we use at LowCode Agency: two customized agents that handle a significant amount of sales, marketing, and administrative work for a 40-person agency. One supports sales and marketing inside the CRM, while the other acts as a chief of staff with more than 30 recurring responsibilities, from morning briefings and pipeline reviews to extracting to-dos and identifying content ideas.

Finally, we explore why human oversight and clearly defined responsibilities are essential if you want AI automation to create useful output instead of automated busywork. We cover the risks of giving poorly supervised agents broad access to tools like email and CRM systems, why new automations should earn trust before getting more control, and how keeping your agent setup manageable makes mistakes easier to catch. If you’re thinking about building an AI agent system for your business, this episode offers a simpler way to think about it: don’t ask how many agents you can deploy. Ask how much context, direction, and control each one has.