How to Build a Startup

This is the approach I would try if I was starting an AI software startup from scratch. It’s more of a strategy than a business plan. It doesn’t have to be complicated (easier said than done). Here’s the template. This is the approach I would try if I was starting an AI software startup from scratch. It’s more of a strategy than a business plan. It doesn’t have to be complicated.

It’s easier said than done but it’s easier to play McKinsey and making real decisions. Here’s the template.


Phase 1 – Have a workable product

Have something to show. A prototype. Everybody has good ideas and good intentions (e.g. peace on earth). Executing is the important part. You need to deliver something to demonstrate that the product could  be applied successfully to solve a complex problem (or save money, or to deliver substantial economic value in a short period of time).

“AI-X: Our application address a wide range of predictive analytics use cases. The app is designed to integrate and process highly dynamic data sets from networks (e.g. sensors, satellites, enterprises), and enable advanced machine learning capabilities.”

How do you get started? Do a trial. It’s like those sample stands at Costco or when you take a car out for a test drive. Trial projects typically consist of several phases including project kickoff, design, data integration, configuration, validation and final demonstration.

Not sure where to start? Study your potential client/customer. You have a deep understanding of your users’ needs. Successful startups understand some group of users and can make what they want. Make something people want. It genuinely need to delight its customers. Otherwise it will never get off the ground. In a market economy, it’s hard to make something people want that they don’t already have. That’s the great thing about market economies. If other people both knew about this need and were able to satisfy it, they already would be, and there would be no room for your startup. So you need to address either a new need (possible uncertain need), or a new way to satisfy one. Can’t find a need? Look for problems and solve them.

Don’t try to do too much. You lose focus and important resources. This sells well and attracts attention. Do one thing and do it very well. People are attracted to success and VCs will start replying to emails.

Stripes, Shopify, Square all started with a single product addressing a problem, then expanded to solve other business needs.

Phase 2 – Customer concentration

Get a few high profile clients. You want to focus on “lighthouse” customers. They are a trophy you can display. If bank X likes us, then that’s a seal of approval. Try to get clients from a few different industries (banking, energy, industrial, military, government etc..). Get them as partners if possible. If you are trying to save the planet, it will look good for them in their glossy ESG report. It also gives your product credibility. Having a solid partner sends a signal to the market “hey, this is important, pay attention.”

Phase 3 – Scaling the business

After having a few juggernauts clients and establishing your credibility, start going after middle-sized companies. The business is not a startup anymore. It’s on the path to be structurally profitable and cash flow positive. If you are doing financing round, by this time you should be a hot ticket.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s