Unlocking True AI Potential – Beyond the Hype to Real Business Optimisation
Recently I was involved in a meeting where I was trying to explain to stakeholders the benefits of AI when automating and optimising business processes.
I find myself explaining that AI automation of processes in the company is not (or should not be) someone copying email text to ChatGPT or using Copilot to summarise or rewrite emails.
But I have to say that I understand where they were coming from.
For example, I often hear or read online from “AI implementation experts” that AI tools like ChatGPT or Copilot can optimise business processes or enhance team efficiency by reformatting or composing emails based on a brief description from an employee or creating a better Excel spreadsheet.
However, I believe that this approach has significant flaws. Here’s why:
1. **Efficiency Misconception:** If an employee has to manually copy text into ChatGPT or ask Copilot to summarise or write an email, the time spent might be almost identical to writing the email from scratch. Moreover, how can you effectively measure the impact of such optimisation? It’s unlikely that you’ll see a noticeable increase in the number of emails answered by the team due to these tools.
2. **Questionable Necessity:** Do you need emails to be better formatted or phrased? How often have clients complained that your employees’ emails lack formality? If complaints are rare or nonexistent, what’s the real benefit?
True optimisation should focus on clear, measurable outcomes. For instance, if you aim to automate or optimize the process of “Processing client orders received via email,” the success of your project should be defined by a measurable reduction in the time employees spend on this task—or even better, by eliminating the need for human intervention altogether. This makes it easy to conclude whether your project has succeeded.
To illustrate my point, I recorded a short demo showing how AI can optimize “Processing client orders received via email.” (video below) In this demo, an AI agent takes over the task of communicating with clients and processing orders, drastically reducing—or even eliminating—the time required from employees.
The essence of any optimization effort is that it should produce measurable effects. If you can’t measure it, you shouldn’t pursue it. This principle should apply to every activity within your company. In my opinion, there are only two legitimate reasons for any business activity:
1. To increase company profit.
2. To improve the product or service offered to the client. (Under the condition that you can verify whether the client is willing to pay for this improvement or if it will increase their loyalty to your brand.)
This framework can be applied to any internal project, even broad initiatives like “Increasing Brand Awareness.” The ultimate goal of such a project should be to drive more revenue, whether through increased orders, higher prices for existing products/services, or other tangible outcomes. If the only result is more traffic to your website or more social media engagement without a corresponding rise in revenue, then discontinuing the activity could save significant expenses, especially since these types of projects tend to be costly.
I know that many people may not agree with me, but I firmly believe that every business activity should lead to one of two outcomes: more revenue or a better product/service. Anything that fails to achieve these objectives should be considered “Muda” (a Japanese term for waste used in Lean methodology).
So, the next time an “AI specialist” proposes optimizing your business, ask yourself:
– How much time or money will this actually save?
– Will this make my product or service better? If so, how will I measure client satisfaction?
My advice is to start with one process at a time so you can clearly track progress.
In my opinion, there is no way back and businesses of all sizes will have to use AI in the feature, if not they will be left behind.
Imagine running your business today without email or a mobile phone, I think AI in the feature will be that essential.
However, implementing AI is often complex and requires thorough preparation. It’s not as simple as some might suggest.
If you’re interested in learning more about how to get your business on the right path to implementing AI, either now or in the future, please feel free to contact me. I’d be happy to help.
Autor: Pawel Jankowski