AI tools are flashy but token-based. SaaS is reliable but has vendor lock-in. Here’s why Excel still belongs in every entrepreneur’s toolkit.

These are the 10 ultimate AI tools to automate your business.

Here are the 5 SaaS tools every entrepreneur needs.

How to build a 1 million ARR business without a team.

The internet is full of catchy headlines of this type. And yes, there are plenty of AI and SaaS tools out there that make life easier for entrepreneurs. But no, AI will not entirely replace SaaS (I will tell you why later in this article). And yes, there is a surprising third option for entrepreneurs out there (I will not spoil the last paragraph of this article yet).

Are you curious about the third option? Don’t jump to the last paragraph of this article yet; let’s go through the options one by one.

The Contemporary Choice: AI

Needless to say, the whole world talks about Artificial Intelligence nowadays. AI labs claim that their models will solve every IT problem in the world, and that soon, all Software-as-a-Service (SaaS) tools will be dead. Plus, all software developers will be made redundant. And on top of this, AI will write better software than any human ever could.

Well, that’s marketing blah blah. The future will hold a place for both SaaS and AI tools. We’ll come back to SaaS a little later in this article.

But first, let’s look at the advantages of AI tools. First of all, entrepreneurs can build their own simple tools relatively easily by using a 20$ monthly ChatGPT or Claude subscription. Building is easy even for non-technical entrepreneurs, and results are visible almost immediately. If you’re not into building your own AI tools, there are many ready-made AI tools you can subscribe to the same way as any SaaS tool.

So much for the advantages. What are the disadvantages of all those AI tools?

First of all, never forget that all AI tools are based on probabilistic language models. This means that no two outcomes for the same operation will ever be 100% identical. That isn’t a problem per se, but it can be if you’re working in a safety-critical domain such as aviation or med tech.

Second, all AI tools consume tokens of text. The more tokens you consume, the higher your bill will be. That’s the case for input tokens (the prompt you use to talk to the AI tool) as well as for the output tokens (the response you get from the AI tool). And we all know that all the major AI labs are still highly unprofitable. Therefore, they will raise the token prices significantly at some point in the future. The question is not if, but when. When token prices rise significantly, you can either pass on the cost of AI tokens to your customers, which will probably lead to some customer churn. Or you are paying for the higher costs yourself, which will probably push you into unprofitability or even bankruptcy. It’s that simple.

Third, don’t confuse self-built AI tools with production-ready software. What might work in an early-stage company on your local machine will most probably fail gloriously at a later stage when your company has grown, and more users put a higher workload on your home-built AI tool.

The Reliable Choice: SaaS

And that’s where good-old-SaaS enters the stage. With Software-as-a-Service, as the name suggests, you’re not just buying software, but also a service. The service consists of a team at the SaaS company who fixes bugs, upgrades components, applies security patches, and restores your data at 2 am when disaster strikes. Furthermore, most SaaS tools are deterministic, i.e., outputs for a specific operation will always be 100% identical. In certain industries, predictability is more important than AI fanciness.

Where there is light, there is shadow. Just like AI, SaaS has a few disadvantages.

First of all, just like with AI tools, you are dependent on your SaaS providers. If they raise the prices, shut shop, or discontinue a product you depend on, there is nothing you can do. You cannot transfer the software to your servers; in many cases, it’s even difficult to extract all your data from a SaaS tool in a useful format.

Second, the world has shrunk in recent years: Geopolitics is back. Gone are the times when you could do business globally without any restrictions, and gone are the times when you could rely on a contract with a SaaS company from a foreign country. Digital sovereignty is back, which means many companies favor SaaS solutions from their own countries or regions, or open-source solutions they can install in their own server rooms.

Third and last, SaaS tools cannot adapt to user requests as quickly and as individually as AI tools. Whereas AI models can learn from a user’s behavior and adapt, SaaS tools are usually based on a single code base for all users. This means that feature requests and feature improvements take longer to implement than an AI tool able to adapt its behavior to individual users.

The Pragmatic Choice: Excel

Now, what’s the third option? It’s surprising, but it’s Excel (or LibreOffice, if you’re surfing on the digital sovereignty wave).

For many tasks in a small or mid-sized company, Excel is good enough: There is no need to buy a risk management software tool if you’re tracking just a dozen risks. There is no need to buy a license tracking tool if you need to track the license expirations of just a few tools. You can set up a simple Excel tool for this, and put the update of those Excel tracking sheets on your end-of-month task list. Subscribing to an AI or SaaS tool for such simple tasks would be a complete overkill.

With Excel sheets, you get the benefits of both AI and SaaS in one go: They are easy to build and maintain, and easy to adapt. And beyond the benefits of AI and SaaS, Excel sheets don’t carry a running cost — not even a Microsoft 365 license if you use LibreOffice.

The downside of Excel sheets? They are easy to screw up, especially when multiple people work with them.

Conclusion

Just because you have a hammer, not every problem is a nail.

Frontier AI models are hammer tools. But tracking a dozen risks in a startup is not a nail problem. Therefore, you can do very well with an Excel sheet for this problem.

On the contrary, try managing compliance with regulations and policies in a multinational airline with Excel sheets. I’d recommend a specialized SaaS tool for this (and yes, at Yonder, the B2B SaaS company I co-founded, we have a suitable solution for you).

In today’s world, it’s not either-or. You will benefit most from using a healthy mix of AI tools, SaaS subscriptions, and self-built Excel sheets. Don’t be guided by AI marketing talk; think for yourself which tool fits your business and your problem best. 

And never forget the pragmatic Excel option before you make a final decision. It’s suitable for more problems than you think.