The common AI traps Vs Investor Expectations in a StartUp

This article details the common AI traps startups fall into. In general, it is very exciting to talk about Artificial Intelligence, Machine Learning and its possibilities. However, there is still a question at large as to how ready the market is.

Of the technologies that make common rounds in the startup world – Machine Learning, Quantum Computing, Block Chain are probably the most common trends. Not only startups, but the investors too seem to be swayed by the buzz words.

I recall an investor conversation about why our AI software is not going to use Block Chain technologies. It almost brought me to a question – what is the end goal of all these. If a business/ startup is not considering problem solving for the market, it is indeed a technology trap. In this article, we discuss the common AI traps under three broad questions

1. What is the value proposition?
2. Why investors like AI startups? What are they expecting?
3. Conversations with Investors/Mentors

Common AI Traps Vs Value Proposition?

This is the first thing any business/startup idea needs to be asking for. We will need to work through a business plan, thoroughly understand the market before going deep into technology.

Being an engineer, I must admit that although not the easiest choice, technology needs to be secondary. A primary mandate for the startup is to create a business solution. Just because the technology is great, doesn’t necessarily mean customers will buy it. This is where a business plan must clearly elucidate the value proposition.

Once the purpose of a startup is clearly established, the rest is about HOW. This is where technology comes into play.

Why investors value AI in Startups?

The value proposition although highlights the need for the startup, the HOW is also very important.

This is where technology comes into play. Although AI hits the technology trend, the value for an investor really is in:

• Multiplier: Investors are naturally curious for the rate of return. The common multiplier values for AI/Machine Learning based startups is quite high (Multiplier is X times the turnover of a business when ready for sale). Although I coined AI as common technology trap, it did turn a lot of heads towards us.
• Competition/Threat of new entrants: Well every investor would like the startup to be a unicorn, earn lots on their investment. This is true when a product is harder to copy. The AI technology helps contain most of itself into a black-box environment. The technology makes a product unique and AI fits in that bill very well.

The common AI traps which startups fall into

Knowing the above two, there are a few traps startups must avoid in their approach

1. Is it really feasible?: The AI technology promises a lot, but not all can be delivered at ease. From a startup perspective, we need to keep the goals smaller and consumable. Once we are able to achieve these, we can boost investor confidence and AI can make it a bit more alluring
2. Time to develop: My experience of roadmaps has taught me that no matter how conservative we think the plan is – it still is very optimistic. Technology development takes a long time. Along with that, we need to plan for testing, customer trials which makes the timeline longer.
3. Technology for Tech sake: AI must be about problem solving. I have seen startups use AI without really acknowledging the real values it can provide. We cannot afford to make a startup just about the technology. It still is about problem solving.

One additional experience, I would leave this article with:

AI Investment Conversations

Aside of the common traps and investor point of view highlighted above, a few questions I have faced in the investor discussions are:

• High level architecture: It helps to have a schematic for the high level architecture. Questions around what software language is being used, servers etc will be helpful
• Data-set for training: Most investors would be curious about this. Are you using public data, how is the AI/ML being trained etc
• Potential customers to try AI training and validation: The value of AI software is only after it has started working. Have you lined up any potential partnerships? If so, what is the commitment you’ve received from these business partners.

Well, arguably it is not easy. But such is the life of a startup – nothing comes easy. We have to find ways to wade through. For a AI startup, we have to steer clear from the common AI traps which startups fall into. After all, it is technology for problem solving and not the other way round.