The checklist for AI entrepreneurs.

Théo Szymkowiak
2 min readFeb 24, 2017

--

How to check the sanity of your market opportunity.

Artificial Intelligence is becoming a buzz word. Money is pouring and the market is getting a little bit crazy. Here is a checklist for anyone looking to start an AI startup.

How much data do you have? How much data can you acquire?

Credits: http://wersm.com/wp-content/uploads/2015/08/wersm-how-much-data-is-generated-every-minute-infographic.jpg

Do not compete with Google (or Apple, Microsoft and Baidu for that matter). In the Artificial Intelligence race, data is a scarce resource. Going against those huge data monsters is just pure suicide.

Your business must leverage data that has not been collected or exploited in a meaningful way. The business of AI is really the business of data.

It is impossible to have a working AI on day one. Your Artificial Intelligence(s) will grow with your business. You can think of their intelligence (or accuracy) as a logarithm function. It becomes harder to make them significantly better over time.

Can you exploit the data in a meaningful way?

Data is a necessary but not sufficient condition. You must be able to exploit it in a meaningful way at a somewhat early stage of your business.

How do you do in the first months if you don’t have data? Humans. You need to hire humans to perform the job of your future AI. Then you need to automate the process piece by piece until the entire pipeline is fully automated.

Take for example the famous startup that schedules meeting automatically. They have been running for months, however, all of their AI’s responses are checked (and corrected) by humans.

Data is useless unless you can exploit it.

How many good data scientist do you have/know?

Credits: https://www.slideshare.net/AdamRabinovitch/opus-data-scientist

You need data scientists, this is invariant. This is also the main problem for many non-technical entrepreneurs. The best DS (data scientists) usually work at Google, Facebook and Microsoft so you’ll have to fight to attract them.

Each DS adds a lot of value to your business. They drive innovation.

Your success is directly related to how good your data scientists are.

This is all for now. If you think I missed something please comment and I’ll update this post!

PS: AI for X is not a business plan!

--

--