The AI Summit London 2026: Real Conversations About AI at Scale
July 1, 2026
The AI Summit London celebrated its 10th annual edition in 2026. The sessions were full, the floor was busy. Many people stopped by our booth to see what was new, and we showed them Scolta: Tag1's open source AI site search with no SaaS fees, no vendor lock-in, and true data sovereignty. Some conversations went longer, and interest in Scolta at the event outpaced the time we had, so we are hosting a live demonstration webinar.
Another crowd favorite was Jesse, our open source AI assistant. It's private by design: run it solo or deploy it across an organization, and either way only you can see your own vault, locally. Setup takes 15 minutes, then Jesse watches everywhere your work piles up (email, Slack, calendar, and more) and hands you one prioritized list each morning. No more juggling five apps and hoping nothing slipped. And you can evolve how it works simply by talking to it.
The people who lingered weren't exploring AI in the abstract. They were managing it, and the problems they were working through were specific.
Themes That Stuck
Across two days and dozens of conversations, the same themes kept surfacing.
Governance. For leaders running operations across global enterprises that already had teams using multiple AI tools across regions and functions, adoption was settled. The open question was how to govern what already exists and define a path forward. The recurring feedback was that most vendor narratives don't address it well.
Visibility. Governance assumes you can see what you're governing, and in regulated environments like financial services, that assumption was failing. The people responsible for system access were finding it increasingly difficult to track what each AI vendor could reach, how data moved between systems, or which permissions were active. Part of the problem arrived uninvited, as AI features switched on inside software the organization had already cleared. The models weren't the worry. The blind spot around them was.
Cost at scale. An increasing number of teams had rebuilt their development workflows around AI. The trouble was the math. Token usage had become hard to control, and the costs weren't adding up. Many scaling adoption hit this wall.
What Stood Out on Stage
Our CEO and Founding Partner, Jeremy Andrews joined a panel titled “When Everyone Can Build Software: Democratizing Development.” The discussion focused on a practical reality many teams now face: AI has made it much easier to start building, but it has not made production any less demanding.
The clearest ideas from the panel was that AI has widened the doorway into software creation. More people can now prototype, generate, and iterate quickly with low-code, no-code, and AI-assisted tools. But the harder side of the doorway, getting reliable systems into production, still depends on architecture, review, governance, and accountability.
That is where Tag1’s perspective stood out. In open source, trust has always had to be earned through process, verification, and experience. AI introduces a similar challenge: the output can be useful, but it still needs guardrails. Existing tools such as linters, testing practices, and structured review remain essential, especially when teams are tempted to let AI validate its own work.
The panel also reinforced a point that is often lost in the hype cycle: human oversight is still critical. AI can accelerate delivery, but it cannot replace architectural judgment, production sign-off, or the responsibility to ship systems that are secure, maintainable, and fit for purpose. As software creation becomes more accessible, those disciplines matter more, not less.
Democratizing development is a meaningful shift. Making software easier to build does not automatically make it reliable, secure, or maintainable. That gap is where governance, standards, and operational discipline come into play.
Scolta is Tag1's open source AI site search. No SaaS fees, no vendor lock-in, your data stays yours.
Image by the AI Summit London.
Related Insights
-
/