The answer is slowly and carefully.
Most big businesses ignore this advice and are spending millions on the most expensive consultants and then designing expensive software that will never work and be overtaken by better technology before it is implemented.
Who remembers the early days of e-commerce? Around the year 2000, Standard Bank spent millions creating an e-commerce platform that never worked. Why did they do it? They knew that e-banking was coming so they had to do something.
Another analogy: You know that Voyager 1 was launched in 1977. It is now 25 billion km from earth. People are asking why we don’t send another Voyager to our closest star, Proxima Centuri? So far, in 48 years Voyager has travelled nearly 1 light day (the distance light travels in one day). The closest star is 4.3 light years away. If we were to send another Voyager now it would probably be twice as fast. However if in 1000 years time earth (if it still exists) were to send a spacecraft then it would at a wild guess be 900 times faster and catch the previous Voyager within 19 days! Why send another Voyager today? The Voyager of 3025 is way ahead of what we can currently conceive.
The point is technology is improving at a rapid rate and it is sometimes better to wait than to waste money today on something that can be done better in 5 years time.
In the AI field, AI is still very new, but improving at a faster rate than Voyager 3025. Whatever you are doing with AI today, you will be able to do 100 times better in 2030. Does that mean you should do nothing today? No, not at all. However do not start a billion Rand project today that will only be finished in 2030. Start a project that will be usable this year. In 2030 you can improve it with new technology.
So, what do you do now?
– The quick and easy things.
– Get to know and use the LLMs. Make sure every staff member knows how to use them and make sure they do.
– Use and implement the personalised chatbots and agents. Aim to use it where you have a problem that needs solving today, for example, my favourite, customer support. Don’t waste time getting it perfect! If it handles 80% of customer support, it’s a good start. Make sure you monitor the 20% that does not work. Put your customer support staff onto those areas and incrementally improve your AI agent once you know what issues your customers have.
– Use AI tools that are currently available: e.g. Image and video creation, voice to text, creation of minutes, automated agents.
– Keep up to date with available technology, but don’t let it intimidate you.
As much as I sound critical of businesses wasting money on vanity AI projects, I’m also critical of businesses that are too conservative to even consider AI. Far too many CEOs say AI is an IT function so they do not need to know it. Yes, IT will implement AI, but your IT guru will never know as much about marketing your business as your marketing director.
Subscribe to our newsletter and keep up to date with useful AI news.