AI is getting better but it doesn't replace the need to understand the problem
There is barely a week now without another announcement about artificial intelligence becoming faster, smarter or more capable.
But perhaps the more interesting change is happening somewhere else.
AI is beginning to change the economics of getting things done.
Reuters recently reported that AI is already reshaping the global IT services industry. Clients are expecting greater productivity for lower cost, traditional labour-heavy development models are coming under pressure, and smaller, more agile technology companies are increasingly able to compete with much larger suppliers.
That reflects something we are seeing ourselves at Isoblue.
For many years we have helped businesses and organisations understand processes, solve operational problems and design better ways of working. Quite often we could identify exactly what software should do - but developing a bespoke system from scratch could require a team and budget that simply didn't make commercial sense for a smaller organisation.
AI is changing that equation but it doesn't replace the need to understand the problem. In many ways, that part becomes more important.
Someone still needs to work out what the organisation actually needs, understand the users, challenge assumptions, structure the information, design the experience and decide what shouldn't be built.
What AI does is give experienced people significantly more leverage once those decisions have been made.
Research, specification, coding, testing, documentation, interface development and problem-solving can all now be accelerated with AI assistance. Work that might once have required several specialist roles can increasingly be undertaken by a smaller, multidisciplinary team.
That doesn't necessarily mean doing the same work for less. More interestingly, it means being able to undertake projects that might previously have been uneconomic.
We've experienced this directly while developing IsoStack, our own technology framework, and applications such as SeasonPro, our football league management system.
Neither exists simply because AI can write code. They exist because years of business-analysis, database, design and user-experience experience can now be combined with AI-assisted development to create and improve software much more efficiently.
And there is an interesting contradiction emerging…
AI itself is progressing extraordinarily quickly, but businesses aren't necessarily changing at the same speed. OpenAI CEO Sam Altman recently acknowledged that adoption across the wider economy has been slower than he originally expected.
That shouldn't really surprise us.
Businesses don't change simply because a new technology exists. They change when someone can demonstrate a practical, affordable and sufficiently low-risk way of using it to solve a real problem.
OpenAI's latest research into business adoption makes a similar point. Organisations making the greatest progress aren't necessarily using different AI models from everybody else. They are finding better ways to incorporate the technology into real work and processes.
That, for us, is where things become genuinely interesting - AI will undoubtedly continue to become more capable but the lasting opportunity may not be the technology itself, it may be what experienced people and small organisations can now afford to do with it.