Don’t automate too soon!

Can you think of a cooler trend in business recently than AI, ML and generally: automation? Artificial Intelligence, Machine Learning and all types of ‘big scaling’ is something that seems to be a vital part of the ‘pitching bingo’. Most startups are proud to be including these while trying to convince investors to join their projects.

Recently I’ve heard a great summary of these actions. How to differentiate machine learning from artificial intelligence? Well, apparently if it’s written in Python it’s ML, when it’s AI, then it’s done in Power Point. I love this comparison, as it shows how artificial it is and how little ‘junior entrepreneurs’ know about when to automate their businesses. In other words, in most of the cases AI still exists only in Power Point presentations and it’s the machine learning that barely starts it way up in the daily reality. More on that below.


There is a good (audio)book that is worth your consideration when it comes to choosing the right moment for automation and scaling. It’s “Nail It Then Scale It” by Nathan Furr & Paul Ahlstrom. Dive deeper when you have a moment.


In simple terms, real AI is something most businesses will still not need within the next decade nor would it be financially efficient to be investing in that field. When it comes to ML, a ‘machine’ usually takes a lot, a lot of data in order to ‘learn’ its ways to help in process automation. By ‘a lot of data’ I don’t mean tens and hundreds of records of data, but hundreds of thousands and more. For the majority of businesses, this will take years to gain enough data to be even thinking about the proper use of ML.

Having said that, there is much you can do in between. Believe it or not, there used to be business before artificial intelligence and machine learning. I am not talking only about pure human intelligence, but also about the algorithms and process optimizations that can be done much cheaper, quicker and in a more financially effective way.

While meeting my clients via (yes, in a sauna) or while listening to startup pitches as an investor at I often observe the same scenario. The Team that has merely started its business and is in a very early stage wants to scale up their business through AI, ML and automations. Much too soon. Why?

There are few major reasons why you should not consider automation (or its higher forms) too early.

  1. Cost of automation, incl. planning & IT development is extremely high when compared to manual work.
  2. In most cases, the lack of full knowledge of the process means also that the resources invested in automation are simply wasted. You’re not building the thing that is needed at that time.
  3. Automation often means increasing the distance between you and your customers. You don’t want to be doing that too soon.

There is of course a point at which the scale of your business will justify automation, but in most cases, that moment is further than you initially think. As an example – at AirHelp most of our processes were very manual even when we were serving 3k customer enquiries a month. It started to be useful to automate at the level of 10-50k customers and became essential only after reaching the goal of 100k customer enquiries a month. Now, having helped over 14M air passengers, the company’s processes are very much automated. Today even with c.a. 100 product & tech specialists on board, it’s still not always so obvious which part of the process should be automated next.

In the initial years it was way easier and more effective (time & money) to increase the size of a Customer Service Team. That has given us more data and opportunities to realize what parts of the process are the real bottle necks and require automation in the first place. The priorities here were changing pretty much every quarter as we got wiser as a Team. For that to work well, you need the best people in place. At AirHelp, that would not have been possible if it were not for Natalia Laskowska – best VP of Operations ever. Her knowledge of the processes and ability to build a great Team is best in class and has helped the company grow in the right direction at an amazing speed. I’ll touch on the people element of automation and growing business in a separate article, as it definitely requires dedicated attention.

To sum it up. Learn from this process. Whenever you think of automation, consider the cost of that process and the alternative cost. Can’t you use your IT resources better and in the meantime deal with increased manual workload in a more efficient way? Make a simple business case whenever you’re about to add a new ‘automation task’ on your roadmap. Delay tech driven automation without stopping process optimization.


Boleslaw Drapella

Founder at, exCEO of AirHelp Poland.

Let’s meet in a sauna to talk business.


Don’t want to miss another article? Leave your email below.