Web Summit 2025: How AI Is Turning Everyone Into a Builder
This November, I had the chance to attend Web Summit 2025 in Lisbon — and wow, what an experience. The 10th edition was the biggest yet: 71,000 attendees, 2,000+ investors, and participants from 130+ countries. The scale was staggering, with 13 stages and 20 content tracks running in parallel. For four days, it felt like the entire tech world gathered in one city.
With so much happening, it was easy to get lost in the buzzwords, especially around AI. But the sessions that really stayed with me weren’t about hype; they were about how AI is reshaping the way we build, collaborate and innovate. Here are the ideas that made me stop and think.
- Daria Barsukova de MeloSoftware Engineer
“Demo, Not Memo”: From Ideas to Working Products
In large organizations like ours, developers make up only a small fraction of the workforce. Yet some of the most innovative ideas come from colleagues in HR, marketing or business teams. The problem? Most of those ideas never make it past the pitch stage, let alone reach engineering backlogs or production. The journey from concept to implementation is often slow, expensive and tangled in complexity.
That’s why the phrase “Demo, don’t memo” resonated strongly with me. Why not build something tangible right away? Anton Osika, founder of Lovable, showed how their vibe coding platform lets anyone turn an idea into a working demo in minutes. No waiting for engineering backlogs, no endless PowerPoints: just create, test with real users and see whether the idea actually works. With AI-driven tools and agent-based coding, both tech and non-tech professionals can validate ideas quickly before committing resources, instead of burying them before they even get a chance.
Everyone Is a Builder. And That’s Exciting (and Messy)
Anton’s point connects perfectly with what Anu Bharadwaj, President of Atlassian, shared in her keynote: AI has democratized building. It’s no longer just developers who can create solutions — now marketers, HR, finance and business analysts can build, too. She illustrated this with the example of Rovo, an app created by an HR specialist at Atlassian to improve performance review conversations. No coding background, just an idea and the right tools. This product is now extremely successful, with over 30,000 Rovo agents in use.
This is empowering, but it also introduces complexity. When AI stops being a special project and becomes the default way of working, how do we manage the chaos of innovation at scale? The answer isn’t just teaching people to use AI tools. It’s about orchestrating change responsibly. Large companies are complex systems. Aligning early adopters with slower-moving teams while staying true to strategy is hard.
In Anu’s own words:
“AI has brought the power of building to everyone in the company, not just technical teams. What happens when AI stops being a special project and becomes the way every team works — marketers, HR, finance, developers, all building their own solutions. It’s exciting… and a little chaotic.”
Beyond Coding: Orchestrating Workflows
From my experience, developers spend around 20% of their time actually coding and the rest coordinating tasks with other colleagues, analyzing and problem-solving. How can we use AI moving forward? It’s no longer just a developer’s tool for writing code. Its potential lies in how we design and manage work itself. We should explore how to leverage AI at a higher level of abstraction to streamline the remaining 80% of tasks related to workflow management. Instead of layering AI on top of existing processes, we need to rethink workflows from the ground up. This approach could be a game-changer for large organizations like ours, where driving change is often slow and challenging.
Michele Catasta, President of Replit, an AI-powered platform for building web apps, brought up that autonomy of AI agents is a key for non-tech builders. Autonomy, according to him, consists of capacity (or intelligence), verification and context management. For non-tech builders, verification of newly built features or products is critical. They shouldn’t have to debug broken features or deal with AI hallucinations. The experience needs to be simple and reliable. That’s where sub-agents come in: They can iterate on results, detect mistakes and create self-correcting feedback loops. They can even manage context across tasks, so that the main agent can focus on execution.
The Elephant in the Room: Security
Of course, none of this works without trust. AI-built products are not immune to mistakes and security can’t be an afterthought. The same goes in terms of permissions to knowledge assets. We would need granular permissions control: per agent and per task. And yes, security audits remain essential, just like with human-built software. Anton Osika summed it up perfectly: “AI-built products are prone to errors and require the same security checks as traditional software.”
Final Thoughts
Web Summit 2025 made one thing clear: AI is no longer just a developer’s tool. It’s a company-wide capability. The challenge now is to harness it responsibly orchestrate workflows intelligently, and create an environment where everyone can build, without chaos.
The future isn’t about replacing humans with AI. It’s about amplifying creativity, accelerating innovation, and building together. Maria Sharapova (yes, the tennis legend) said something at the opening of Web Summit that stuck with me: “AI should help humans avoid burnout and contribute to one’s longevity on their own terms." It reminds me that progress should enhance life, not exhaust it.
