- The Generative Beings
- Posts
- AI Safety and AI Agent Startups in SEA
AI Safety and AI Agent Startups in SEA
My keynote at AIQURIS event on 6th March

AI Agent startups in Singapore and South-East Asia
AI Agents - the 2025 phenomena
While 2024 was the year of Generative AI, 2025 has been voted the year for Agentic AI. 2026 is still up for grabs for Embodied AI, AGI, or Quantum...

2022: The mass adoption of ChatGPT drove a fundamental shift in our perception of what AI can do, and ignited the fire for every individual-to-company to transform completely, and be AI-ready.
2023: The rise of Generative AI (of course it’s been there since before), but, the early indicators of success were ‘POCs’, most of which never scaled to production. This was the phase of ‘solution looking for a problem to solve’.
2024: GenAI became mainstream, with all the major technology giants forging new alliances, utilising their existing userbase as rapid distribution channels. Meta rolled out the Blue circle across a billion WhatsApp users and my parents started to create images and messages using GenAI. While Microsoft, Google, AWS, made major enhancements to their foundational LLMs, and business models, a new age of technology giants became household names - OpenAI, Anthropic, Perplexity, Groq, Grok, and many more.
2025: is the year AI Agents become mainstream - by delivering smarter automation, productivity gains, and creation of new use cases. Rule-based ERP workflows are now monolithic structures of the past. What comes next are AI Agents - an intersection of smarter LLM models and decision-making to orchestrate set of actions to achieve the intended outcome.
Connect with AI founders building Agentic AI startups, leaders looking to adopt AI Safety frameworks in their organizations, and stay updated with the latest in South East Asia with The Generative Beings community.
AI startups have moved beyond GenAI workflows to AI Agent workflows
Here are some of the startups I know are knocking it out of the park when it comes to building AI Agents within their product experience, targeted at enterprises.

These founders are in The Generative Beings community. You can connect with them, try their products, even get a special discount as a TGB member.
Over 1k AI Agent startups globally, and rising
It was amusing to discover that the global AI Agent ecosystem is so vast and diverse. There are over 1k AI Agent startups across 50+ categories (Source: AI Agents Directory), and the list is growing daily.

There are some wild estimates on the potential future market size estimation, one being ~$2 trillion, by 2030 (Source: All about AI).
Now, you can debate on the accuracy of these estimates, but, there is simply no way for us to know. What we should get out of this is the opportunity, upside, and a hope that this technology will help make current processes and products a lot better.
With great power comes great responsibility
We are still not in the ‘Agentic AI’ stage, where AI can reason and take all the right decisions, without human-in-the-loop. But, there are clear and easy wins on smarter automations for manual/ops-heavy/redundant tasks. Autonomy is still not given away, so, humans have a clear role to play - choose!
To build smarter and safer AI agents, there has to be more effort put in to define scope of decisions an AI can make, ethical considerations, guardrails, and rigorous testing. AI startup founders and developers need to integrate these testing and safety frameworks from the get-go.
It may have been an afterthought till now, as the focus was on delivering UX quickly, and it sufficed for now. But, there are grave implications of not building safety nets into the AI workflows.
As the keynote was followed by a panel of esteemed AI Safety and enterprise leaders from AI Verify Foundation, Nomura, SC Venture, and Resaro AI, it only made sense for me to close with my understanding of building safer AI Agents.

AI Safety Events
There is much more to talk about regarding AI safety and AI agents this year! You can start by attending our next community event on AI Safety and Testing with AI Verify Foundation and AWS. It’s going to be a ‘Safe space’ for AI builders and leaders ;)
Event: Testing Your GenAI application: A Beginner’s Guide for Non-technical Users
Date/Time: 19th March, 4:30 PM (SGT)

Bloopers: Don’t focus on the funny expressions - but, on the slide ;)
Stay tuned for more updates on AI Safety from the region.
Cheers,
Shivang
Reply