AI doesn't fail—broken business processes do, accounting expert warns

So there is a lie that is being sold about artificial intelligence and the interesting thing that we want to discuss is that the lie is not that AI doesn't work. It works. AI can write, AI can quote. AI can analyze documents. AI can make summaries of meetings. AI can answer questions, customer queries. AI can generate marketing material. where I can analyze enormous amount of content and increasingly AI system can do more than just answers. They can execute tax, interact with softwares and at the same time operate with a lot of degree of autonomy but that creates basically an enormous opportunity but then also new risks and management challenges. So the problem isn't that AI is useless. The lie is much more subtle and we are here to discuss that because AI is powerful. Putting AI into your business will automatically make your business better. If you buy the AI, buy the software, give everybody access, add a chat box, automate the tax, put AI power to your website, and suddenly your company is supposed to be more productive, more efficient, and more competitive. Well, that is not actually how transformation works as business is concerned. What if we are confusing having AI with knowing how to design a business around AI, right? Because there is a massive difference between those two just go and subscribe for an AI tool. And so that's why we discuss about the lie. Now the first lie is if we adopt AI, productivity will automatically explode. It sounds logical given all what we already enumerated AI can do. It sounds reasonable. But then if an employee used to spend 3 hours doing something and AI reduces that to 30 minutes then surely the company has become more six times more productive. But there is a problem with that assumption. The employees become faster. that does not necessarily mean that the organization become faster because you have just taken a work that used to be 10 hours and you have reduced it to one hour. So let's take for a simple example imagine an accountant who previously basically spent time 3 hours preparing financial statements and information and now it is reduced to 30 minutes. That is a fantastic situation. But the report sits on somebody's desk for two days waiting for approval. So the time that has been taken to reduce to 30 minute doesn't still solve the problem because there is still a challenge somewhere. So AI has been used but has it benefited the organization. Now another department receives the information and manually enters it to an Excel sheet. a manager doesn't trust the numbers and ask someone to verify again because it's AI that drafted this information. So eventually the C received the information 5 days later you discover that despite the fact that AI has did some did some work the organization has not automatically become more efficient. Now where did AI fail? Now AI did not fail. It didn't. AI improved the tax but the business processes remains broken and that is one of the things that we are trying to understand that when we talk about AI and so people basically want to think that AI will magically improve the situation. If one aspect of the business has become more efficient and the rest of the business is not then we have not helped the business in any way because at the end of the day the law of averages will set when one person is under overperforming and the others are underperforming performance will still be normally pulled down. So the business process when it remains broken and the distinction is not just to bring in AI is to be able to understand that bringing in AI you must look at AI integration as a holistic process within the organization. So we need to distinguish three things that matter when it comes to AI integration. Tax productivity organizational productivity and business performance. So anytime you want to bring in an AI model or you want to bring AI into your business, you must ask yourself whether what you are trying to do is to improve a tax or improve organizational productivity or business performance in general. They are related but they are not the same. A worker finishes a tax but a company still takes longer time to process will not help. And when you are bringing AI and that AI doesn't automatically translate to profit then there's still an issue. So until your AI implementation become profitable then you are not working. Now sometimes when we bring in AI and we discover that because something else in the system may not have the right results, we have administrative bottleneck, we have poor data, we have procurement challenges, we have customers demands, maybe cash flow challenges, we may blame the AI. No, the AI is not a problem. It's just that the organization took AI and implemented it without taking into account the other dimensions of AI so that management can benefit from the money that has been spent to improve the system. So AI can save us time but that time can be spent and wasted elsewhere. So it is important that is the first lie which is that when we implement AI we automatically get better. No. The second lie is that AI will fix a bad business. No AI cannot fix a bad business. I come from the data science background and understand. Imagine a company with poor inventory controls, weak accounting records, poor business processes, maybe customer service, slow customer service. employees doesn't know exactly what they should do and they are maybe confused. Now if managers don't know how to integrate AI to improve these different systems no matter what AI you bring in it might not solve the problem. Customers pay for you to reduce time and to for only for speed. Speed is important but no customer is trying to look for where to only do something as fast as possible. Customer is looking for quality. Customer is looking for care. the customer is looking for a wow experience and so when your AI implementation doesn't give customers a w experience then that bad business design cannot be solved by AI. Now if you a business is not also collecting data then you know that no matter how you bring in AI you will not be able to solve the problem because even though you are brought in AI AI also needs data. Now you may be relying on generic data but the modules of the future are those that are trained with company's data. You still using notebook, you still using Excel sheet. You are still using maybe WhatsApp messages. You are still using maybe your brain and your head and you are bringing AI. So if we need AI, it is good. But the first thing you need to ask yourself when you think of bringing AI into your business is but what exactly is AI supposed to fix. Once you ask that question, you will not just go out to subscribe to an AI tool. You know exactly what you want to do with AI and that gives you an idea of what you should do. Please thank you and if we appreciate if you also subscribe to the channel. Suppose your inventory records are wrong. AI can analyze those wrong records but it cannot magically what make correct information from wrong records. So you need to have right records to be able to translate that and so your customer database is incomplete. AI cannot process can process faster but will not give you valid results or complete the database. So if your purchasing processes are inefficient, automating the processes may simply allow inefficiency to do what to happen more faster. This is one of the most dangerous misunderstanding about management trying to adopt technologies into the organization. They think that technology accelerate processes without improving the processes or without putting systems that will help improve. So you have to think about how that is designed. If you give a customer a faster car, have you solved transportation problem? The answer will basically be no. That is how AI comes in. You may simply think that giving a customer a faster car will be able to solve the problem. No. But what about traffic? If there's traffic, no matter how fast the car is, you still have a challenge. So that is what happens with AI. So you can automate confusion. You can automate duplicates. You can automate what? Bad decision making. You can estimate poor data and so you can even automate processes that nobody even needs. So what we are saying is before you ask where can we put AI management should start by saying what business problem are we actually trying to solve. That is important and if you handle that then you are fine. Now the third line about AI is that AI means automation. Yes, we have been tricked into this belief that AI simply means automation. The conversation around AI is now that you can go online and be able to create content, automate everything and everything will function wonderfully. Uh that companies need fewer workers. We need fewer employees, lower payrolls, more automation. And yes, automation is happening. It will happen more and more. But only one possible strategy. Consider two companies so that you understand how this translate. Let's assume that now we have two companies and we're looking at these companies are like how many employees do you have? I have 10 employees. Another one have five employees. Now even though these companies were looking about the number of employees if you want to automate you must first start by understanding the processes and see how these processes can be automated without necessary being coming what just automation for automation sake. So if a sales person wants to automate a process, you need to understand how that process will give customer experience because an automated sale process with bad customer experience will not still help. Right? So we are looking at that there is the place of human in any AI system. So we don't talk about we are like AI is competing to take over. No, AI still need human input. No matter how good it gets, it will still need human input. So we have AI versus human. So we need to say it is not AI versus human. It is what AI plus human or hu