Choosing whether to build or buy AI solutions isn’t a new type of problem, even if the technology is.
A few years ago, it wasn’t unreasonable to build bespoke systems to computerise most business needs. Then, as the IT industry matured, pre-built software became more effective, especially for commodity functions like accounting.
AI is in the early days of such an evolution, and most AI work today seems to be bespoke development. “Off-the-shelf” solutions are appearing, but are mostly “platforms” to adapt and configure, not products to unwrap and use.
But in due course, that will change, and the arguments for building your own AI will become more tenuous. At that stage, most businesses will buy AI solutions, with AI custom-build only applying to specific circumstances.
The difficult question for businesses today is whether they get on with their own AI work in the meantime. Alternatively, is it better to wait for AI solutions to become more reliable and common. And cheaper. Or does the a middle ground make sense, building in-house AI solutions on AI “platforms”?
Below is a run-through of three types of “off-the-shelf” AI solution commonly used today, and when each might be a suitable option for your business.
This article is aimed at AI business readers. The topic may seem technical, but the decisions affect timescales and costs, and will have strategic implications.