by Tan Aik Keong (AK)
Anthropic's recent releases of Claude Code and Claude Design have sent another jolt through the software industry. Claude Code is no longer just a simple coding assistant — it's an AI programming tool that can understand a codebase, modify multiple files, run commands, and help developers ship software faster. Claude Design, launched in April 2026, focuses on collaborating with Claude to generate designs, prototypes, decks and one-pagers. In other words, AI doesn't just write code anymore — it's starting to design too.
As the head of a Main Market-listed software company, I keep getting asked the same question: when AI can write code and design faster than a person can, where should a software company actually go from here?
That question isn't really about crisis — it's about facing reality. In the past, a software company's core value came largely from "can we build it." A client wants an app, we build it; a business wants to digitise a system, we build it; a government agency or enterprise wants a platform, backend, database, payments, integration — we build that too. That was the starting point for a lot of software companies, and it was Agmo 1.0.
From our founding in 2012 through to our IPO in 2022, Agmo 1.0's core business was mobile apps, web systems and enterprise digitalisation projects. Across that decade we built up technical capability, project experience, client trust and delivery discipline. It was a period of rapid growth in demand for software development — as long as businesses wanted to digitise, there was plenty of room for a software company to grow.
After our 2022 IPO, Agmo moved into 2.0. From 2022 to 2025, we started shifting from pure project-based development toward products and SaaS, because we realised a software company can't grow forever one project at a time. SaaS isn't just a different revenue model — it's what let us move from "delivering a service" to "running a product." Things like e-Invoicing, enterprise platforms and EV-related solutions were all attempts during Agmo 2.0 to move up the value chain.
Today, AI has rewritten the rules again.
If AI can code on its own, the barrier to entry for traditional development work drops. If AI can generate designs, ordinary interface design gets commoditised faster too. In future, clients won't pay a premium just because "you can code" — because coding itself keeps getting easier. The real question becomes: do you understand the industry? Do you have the use case? Do you have the licences, the compliance, the hardware integration, the operational capability, the strategic partnerships, the position in the ecosystem?
That's the thinking behind Agmo 3.0.
Agmo 3.0 isn't rejecting software — it's redefining where a software company sits. The software company of the future can't just be an outsourcing shop that takes requirements, writes a system and ships a project. We have to become a strategic partner, a co-builder, and in some areas, a long-term participant in the value we help create. Agmo 3.0's direction is applying AI in industries with a real moat — ones with licensing, regulatory requirements, hardware integration, physical operations, real assets and strategic partnerships.
Why do those areas matter? Because AI can write code, but AI can't automatically obtain a licence. AI can draw an interface, but AI can't automatically build industry trust. AI can generate a system, but AI doesn't automatically understand the operational risk, regulatory requirements, supply-chain relationships and real-world complexity inside an industry.
The genuinely valuable software of the future won't just be software on a screen — it'll be software that connects to the real world: machines, devices, vehicles, factories, assets, financial flows, the travel ecosystem, the circular economy, logistics networks, or healthcare settings. Software will stay important, but a standalone software feature will get harder and harder to defend as a moat. The real moat comes from combining AI, software, hardware, data, industry access, compliance capability and execution.
That also means developers and designers won't simply be replaced by AI — but their value will be redefined. An engineer's value used to be building the feature; going forward, it's understanding system architecture, data security, business logic and long-term maintainability. A designer's value used to be a beautiful interface; going forward, it's understanding user behaviour, business conversion, brand experience and real-world context.
So a software company in the AI era shouldn't just ask "can we still build faster than the competition?" It should ask: "are we standing in a position that's genuinely hard to copy?"
The power of Claude Code and Claude Design is a reminder: low-level execution will keep getting automated faster, and high-level judgment will only become more valuable. AI will make software development faster — but it will also make software companies without a moat much easier to replace.
From Agmo 1.0's custom development, to Agmo 2.0's SaaS productisation, to Agmo 3.0's combination of AI and industry moats — this path actually reflects the evolution of the whole software industry. The software company of the future isn't just a company that writes code — it's a company that uses technology to enter an industry, transform its processes, connect its assets, and create long-term value.
AI can generate software, but businesses still need direction. AI can complete tasks, but the market still needs judgment. AI can raise efficiency, but real commercial value still comes from trust, execution, resources, compliance, ecosystem position, and a deep understanding of the real world.
That's where a software company should stand in the AI era.
Part of the AK AI Corner column. Originally published in Oriental Daily (东方日报) on Apr 26, 2026.