Is Your Business Ready for AI Automation?

When is the right time and what is the right way to start incorporating AI and automation into your workforce? My name is Sarah Hirschfield and I am the founder of Carly AI, the AI executive assistant. My team and I have worked with companies for the past few years to implement AI and automation for them. We have found patterns in them that we will review today. At first, it didn't always seem that way, but the adoption of technology ended up being crucial for the business. I made this video to share the lessons learned from these companies so you don't have to make the same mistakes they did. By the end of this video, you should understand exactly where your business stands and what you'll likely need to do to take it to the next level. In terms of technology, companies are divided into different stages of growth. Richard Nolan devised this framework. In the initial phase, there is some kind of cheap and easy-to-use technology that employees begin to adopt. It is not difficult to implement nor is it expensive, but the gains that employees obtain are also limited. Then comes contagion where more employees discover and use the technology, and the cost skyrockets. Management loses control of what is happening, mistakes begin to be made, and the cracks are exposed. People are using technology on their own, without supervision or control, and this is not a good position to be in. Finally, management takes control of the situation. They return to the employees, try to monitor things, centralize planning and control. This period is characterized by a lot of frustration among employees, who feel that things were going well and now they are restricted. It also reveals many of the incompatibilities between the software applications they were using. And finally, Nolan discovered that the budget isn't actually being reduced despite all the outrageous spending during the contagion period. During the control period, IT professionals now need to be hired. They are bringing in new software to centralize planning and data management. Costs remain high. Finally, a company can enter the integration phase. There is centralized planning and control, but end users also feel cared for. They also personally assume greater responsibility for their use of the applications. That is a major differentiator between the contagion and integration phases, and management has established compatibilities between software applications. They also monitor the processing and use of the data, and things are going well again. As a side note, Nolan published this article in 1973. He revised it in 1979. He added these two more stages. I'm not going to talk about them today. If you're interested in learning more about enterprise technology in theory, I'd be happy to make more videos about it, but there are a couple I'll focus on. Look at Nolan, look at MIT, look at the OECD. You may recognize that your own company is going through these stages. If you haven't already, if there's anything that's going to make you think about these issues, it's probably AI. Token spending is basically how much you'll be charged for using AI. A token is basically a text character, and AI companies will charge you for the amount of text, images, or video you input into their system and that they generate. To give you an idea of what all this might cost, let me tell you that it's a minimum of $2000 per month per user. And at the high end, it could be something like $10,000 per month per user. So, frankly, a careless employee can spend the entire monthly AI budget in a single day. That's why the question of whether to implement AI and automation is so important. They can make or break a business. And deciding correctly when and how to implement them will have a huge impact on your business's revenue, costs, and frankly, headaches. If your organization is not prepared for AI or automation, if you implement them, you will waste money, waste time, and get stuck in a bad period of contagion. And you won't see the corresponding increase in revenue. This is a very bad situation to be in and difficult to get out of because if you manage to enter the control phase, remember that budgets don't actually decrease when you change phases. Therefore, you will remain stuck with those higher costs, and your employees will also feel more frustrated because you will suddenly be blocking them and trying to micromanage things. Therefore, you must be very careful about implementing these things too soon because the costs are irreversible. If your organization is ready to adopt AI and automation and you're not, you're probably leaving money on the table, wasting time, and most likely stuck in the initiation period. This means that people here and there, maybe on their own personal accounts, have ChatGPT or Claude accounts; maybe you've purchased licenses for your organization, but that's about it as far as AI is concerned. Frankly, this isn't as dire as the case of getting into AI too early because things won't get out of control, but the downside is that you'll never progress to these other phases. Let's talk about the reasons why you shouldn't adopt AI and automation. The first thing I will point out is that this technology is like gasoline. If you have something that works, you have a fire lit and you add gasoline, the fire will get bigger. And if the fire is something good, like sales or revenue, you obviously want to add fuel to it. If the fire is not a good thing, like your costs or a business unit that is broken or not working to begin with, you want to try to put out the fire and not pour gasoline on it. The first sign I see that a company is not ready to adopt AI is that not everyone is on the same system. Half the team is on Outlook, the other half is on Gmail. Everyone calls from their personal phones, it's not recorded anywhere, they don't use RingCentral, and there's no visibility into call activity. People store things not on shared drives, but on their computers or in random locations, and nobody can find anything. There is no centralized calendar or scheduling system, nor is there any project or task management software. Each person takes care of their own tasks. If your organization is like this, the priority should be that everyone uses the same system. You don't want to optimize a system that doesn't work because there are inefficiencies and you're paying a tax for it, and you don't want to put gasoline in a system that doesn't work. A more common problem I see is that everyone uses the same system, but there is no compliance. This isn't always an obstacle for AI and automation, but if this is all the information you have, it gives you a superficial idea that people aren't updating their CRM. The first thing you should do is talk to the people who aren't complying, because in your head you might have assumed there are some reasons why they aren't updating, such as being lazy, too busy, or simply not wanting to, and sometimes that story isn't true. So, within the CRM or within the project management software, perhaps there was an email integration that you hadn't actually used, and doing that thing will be the cheapest and most efficient way to solve that problem. The last situation where it's still not obvious that you should automate is when you have compliance, you have the use of the systems, but everything is done manually. In the minds of many people, they see manual labor and believe that this is the kind of thing that should be automated. In my experience, that's not always true. And this is the checklist we more or less use when we think about whether or not we should automate. And as a reminder, the cost of automating too soon is wasted time and money. That's why it's so crucial to find out whether that manual work can be automated or not. Should we automate it? Because you could end up spending a lot of money on AI and automation that shouldn't have been implemented in the first place. Why automate? Sometimes you'll earn more money. You need to have enough volume to save money if you switch from humans to automation. An example is if your salespeople don't respond to potential customers quickly enough or miss calls, you have incoming interest, they're missing calls, and you're not capturing that revenue. Another case is when you have administrative staff who spend a lot of time scheduling and updating CRMs, but they could be doing work with a higher return on investment, so automation would be a good thing. And finally, a virtual assistant won't work for whatever reason. We can talk about that in another video. There are some cases where a virtual assistant is really the solution. Frankly, setting up a virtual assistant usually takes less time and costs less than implementing an AI solution. So these are some things to keep in mind. Another reason you might want to automate is team morale. You may have done the math and implemented AI and automation won't make you more money or save you money, but your employees hate doing manual tasks. You're worried they'll quit and you'll lose them. It could also affect their work performance because they are not happy. This can be a very good reason to automate. And what business owners often overlook is that when you have this kind of passion for AI among your employees, it means you have dedicated people on your team who will focus on implementing it. That could save you money on the implementation fee, and they can also be a trusted partner for the external providers you hire to automate your business. That said, if they don't seem to really want to help the company implement these things and just want to complain, then it seems that automation isn't the solution because if they complain about the manual process, they'll probably complain about the automated process too. If they are not willing to help implement it. Frankly, it's not unlikely that we'll go back to the owner and tell them that we don't recommend implementing AI and automation yet. This is what needs to happen in the meantime. Also, beware of vendors who will try to sell you these things too soon and waste a lot of your money. So, in summary, the group that is not ready consists of companies with underlying inefficiencies, either in technology or business, that would make AI implementation a very dangerous activity, as it could exacerbate these pre-existing problems. The good news about being in this group is that, generally, the company only needs to do one or two things, and they aren't that difficult. It could be fixing that problem within the CRM, agreeing with the team on how the software should be used, or having a better voicemail system when calls are missed. And with those unlocks, they'll be 100% ready to get things moving. Now, let's talk about organizations that are ready, but are blocked for some reason. The company already has a really good system. There is only one step in the process, which, because a certain software provider does not have automation enabled or does not have a certain capability, people cannot use. However, if you've tried talking to the software provider, they might tell you, "We're working on it." But a couple of months have passed, and you may get the feeling that they aren't really making the changes because they're worried about losing customers if they don't have complete control over customer data. However, I'm not going to comment on what I think of this practice. It's a reality in many industries. My job is to advise my clients on the best way to achieve their business goals. What I usually advise here is to keep waiting and sometimes to push, and sometimes I participate in putting pressure on the software development team because if you can convince that software company, even if you only spend $700 a month, not $70,000 a month on their software. There are ways to exert pressure, and if they can change that one thing, problem solved. Furthermore, even if people urgently request it, can you automate this? I don't think they'll ever do it. They are really annoying. The reality is that external providers, people like me, often can't do what the client wants because the software company restricts access. It's a polite way of saying that sometimes the only way to get what you want, other than changing software completely, is for the company to make a change. All this is to say that if you are a business owner and are facing this situation, my heart goes out to you. It's not a good situation and I wish you luck. The second reason why companies do not automate is that they have had bad experiences with other providers in the past. Maybe they were scammed, maybe they spent a lot of money and nothing really happened, or maybe they had something but it didn't work very well. It wasn't worth much. I understand perfectly. The only thing to say here is that you shouldn't let one bad decision hurt you twice, and if you let that bad decision prevent you from choosing the right solution, you'll be letting it hurt you twice because you'll lose the time and money you could save with AI automation. And the last problem is quite common. The company cannot make a decision. They spoke with 100 suppliers. They compared prices, they compared features, they wondered if they could trust this person, and it took them too long to make a decision. There are two main factors that explain this. The first is that it actually hides a deeper problem, which are the problems of lack of preparation that we saw. We talk about this as one of the reasons why a company is not usually ready to implement an AI solution. Therefore, you may find yourself in this situation where the reason you are so reluctant to choose a software or provider is because you know you are adding fuel to the fire of something that could explode. The second reason is that you have the systems implemented and things are working well, but you have a custom-made system. You listen to proposals from all the providers, visit their websites, see how it could work, but you struggle to imagine it with your system. If you find yourself in this situation, what I would say is to come and talk to us, as this could be the perfect candidate for a custom implementation, which means you'll have a dedicated developer working specifically on your system. This way, you won't have to change everything just to get the automation and AI you want. I'll leave you with what I think should be your first step depending on where you are on your journey. If you don't already have any AI elements for your business, I would start by obtaining a ChatGPT or Claude license. Which one should you get? At this stage, it doesn't matter so much. If you have decision paralysis, simply get licenses for both. If you've already done that, the next step I would take is to audit each department and understand where they're blocked, what's working, and what isn't. And this is the discovery sheet that you will end up giving to your internal IT staff or an external consultant or vendor that will allow them to focus on the most suitable places for automation. We'll talk about internal staff versus external staff and buying versus developing probably in another video. In summary, the decision on whether or not to incorporate AI and automation can make or break a business. If you do it too soon, you'll waste money, you'll waste time, and everyone will be unhappy and panicky. If you do it too late, you will miss out on the opportunity cost of increasing revenue and reducing costs. You'll be stuck using ChatGPT or Claude forever without opening your eyes to all the other things that AI and automation can do.