On (AI) Business Cases
Can AI really revolutionise the business case? And are the C-Suite really the best people to do that? Danni and Dettman investigate.
We are appearing this week as ourselves for the first time. We cast aside the cloak of anonymity, the confidence gained from a costume, the masquerade ball… [This is exactly why we stopped all of that nonsense. Danni.]
We keep hearing that AI is the number one worry for senior professionals like you. Do not let this be the case. It must not be a worry. It is a big opportunity. But it probably isn’t your No. 1 opportunity, it’s just another change programme, just something else to factor in. Do not heed those who suggest we will all be subservient to the AI, tending the daffodils instead of working. Do not get fooled again. [Pete Townshend. Dettman.]
Did the Victorian mill lead us all to lives of leisure? No. It just caused different problems. The mill did eventually lead to a 5-day working week. Down from a previous 6, in a time when only Sunday was a day off and that was because you were expected in church. If those mill owners had it their way, they would have done five minutes of prayers in the mill on Sunday and made you work all 7 days.
Maybe one day, one day… and not soon, AI might allow all of us to try the sunlit uplands of a 4-day week. If that is so, if you want the productivity gains that today’s rudimentary tools can provide, read on.
The theme of this post is really “less is more”. Your first point of action is to figure out what doesn’t work for you in your company today. You can easily write down 2-3 workflows in your business that you are either responsible for or you play a key role in. You could even do this right now if you’re at your desk. What you do not want is 10 or 20 different workflows to think about.
Having done that, for each workflow, write down the half a dozen or so key steps. These should be only the key steps, the most important ones. If you have more than 6, ditch the lesser important ones.
Now identify those key steps where there are known issues. Bottlenecks, time-wasting, uncertainty, confusion. Perhaps there will be 1 or 2 per workflow. This exercise is just normal process mapping. Note the total absence of AI.
You should have now 3 workflows or business processes which are mapped at enough detail to get specific on, and start to solution for. What you are very likely to find is that all of these challenges can be overcome without using AI at all. If there is another way, then take it. It will be cheaper and more reliable and less susceptible to all the storms and heavy weather coming to AI in the next few months.
What you might possibly end up with is a couple of points in just one of these workflows were AI is actually the best answer. This will be a task which requires human judgement. Perhaps it is poorly defined or cannot be resolved using simple IF-THEN binary logic. Before firing up Copilot, is there a way to simplify this task or split it up, so that it can be addressed without AI?
Do you see where we are going? In this AI newsletter, written by people who genuinely love AI and all of its imperfections, who use it all day every day, we are going out of our way to suggest that you use anything else except AI. This is the only rational path with the state of the market as it exists today.
You should only use AI when it is the right tool, and what this means is, the only tool for the job. If there is another way, take the other way.
Danni’s Point of View
So, like Dettman said in too many words, obviously AI creates efficiencies. That might mean what used to take five days now takes four. And honestly, who doesn’t want a four day working week? We could write a whole article on the pros and cons of that alone (and probably will), but let’s not get side-tracked for now. [Please stop adding more articles to the backlog. Dettman.]
The point is, you should take the efficiencies where you can get them. He's absolutely right in saying don’t try to fix everything at once. Pick 2–3 workflows in your business and focus there. But while Dettman's looking at the workflows, I’m looking at the people.
Because for me, the key to effective AI rollout isn’t the tech: it’s your people.
You need leadership buy-in, yes. But that doesn’t mean it has to be your senior team running every AI initiative. In fact, some of your best AI champions are probably sitting quietly in operations, customer service, HR, or any number of teams where people already know what’s broken and have ideas on how to fix it.
So build AI working groups made up of volunteers/champions who’ve shown interest or curiosity. These are the people who’ll talk positively about AI around the business. They’ll be the ones sharing real, relatable use cases and helping drive actual adoption, not just theoretical excitement.
And while we’re on that educate everyone. Make it fun. Make it simple. Don’t do “death by PowerPoint” with long, dry workshops. Instead, share short, snappy examples, interactive training sessions. Drop one use case into your weekly meetings. Something someone did that saved time, made a process better, or avoided a task that no one wanted to do in the first place.
Even better, ask a different team to share something useful in under 60 seconds. Keep it real. Keep it relevant. I’ll be honest, I don’t care what Finance are doing most of the time (sorry, Finance), but if someone from that team tells me they automated a dull report and saved two hours a week, I am absolutely listening, I want to know how I can use their case study to save me 2 hours a week as well. And so are others who maybe aren’t seeing the benefits yet.
The point is: use all levels of your business. Don’t make AI the exclusive playground of senior people or the technical team. When you bring your people with you, when you show them value in real life, that’s when the magic happens.
So yes start small, stay focused, but most importantly, start with your people.