There’s too much talk in a startups world when it comes to AI – and mostly it’s a talk from people who have never done anything in the space. That is the perspective of Konsta Saarela, Director of Accenture Liquid Studios and moderator of the Arctic15 AI & Data track. It is time to really take action and do something about all of these theories and business ideas floating around our heads.
What’s holding back AI enthusiasts from starting companies?
Do you have that Artificial Intelligence business, a friend who is willing to help you start it; are you at all of the AI meetups in your city, talking tech with the other enthusiasts? Maybe you should finally go ahead and do it? Stop being an enthusiast and become a practitioner? Problem is you don’t have the resources to start a business, you don’t know how to sell or market. Sure, you and everybody else in your city might have a business idea, but what in the world do you do with it? You don’t know. Instead, you turn on Netflix re-binge the entire Game of Thrones. Well, put Jon on pause for a while – or maybe ever. This piece might help you finally start off your AI business. The focus is primarily on B2B sales, however even B2C business ideas may benefit, so take a look.
First, You Need The Idea
Duh. You may already have this, but even if you don’t, the next step is still the same: look for what’s viable and current. No pricey androids that can raise children from infancy to adulthood; don’t bother (for now) replacing the entire Helsinki police force with a single computer system in the next three years. Instead, think of what you and your college buddy could produce which would help your client today (not ten years from now), and which you can prototype quickly.
Yes. Quickly. That doesn’t mean producing a low-quality product just so you can get it done in a week, says Lars Selsas CEO and Founder of Boost AI. It’s about figuring out what quality product you can produce quickly to make sure it gives value to your customer.
Don’t ask for a big commitment from that potential business upfront, either. Lars assures that even the largest corporations, while they may have a lot of capital, will need buy-in and strong data for any kind of sizeable commitment. You won’t have that buy-in simply by cold-emailing their CEO, or sending a LinkedIn message to their IT manager. They have no interest in putting their reputation on the line for some unknown startup working out of their apartment. What you need to do then, Lars says, is “create value early” for little to no commitment. “Go in and create good numbers” as fast as possible, because then you “get that trust” and have the data which will enable you to request a larger budget for bigger projects.
A statement from Telia Finland AI Lead Kia Tahvanainen corroborates Lars’s point. So many startups are walking in during a pitch and providing nothing more than really really rough estimations, but what a contact in a corporation needs to get buy-in is an actual estimation at least. Comparisons to other products and services, especially those your client are already using, would be beneficial as well.
Now, what if you don’t know your consumer?
Our speakers didn’t give a direct answer to this question, but what you can do is consider who would benefit from your product or vision.
It’s a cliche, but cliches exist for a reason: businesses succeed when they solve a problem for their customer or make the customer’s life easier. Look around for the problems you face each day at your 9-5. Or go to your local bank and look for issues in their customer service process. Could you build an AI program which could solve their problem?
Lars Selsas made a comment which might help:
“Play the game you understand, because that’s where you have the advantage.”
Lars and his team always wanted to deliver products to enterprises and other businesses. Why? They felt more prepared for that context.
Okay. So, if I do want to approach a business or corporation, will they even respond?
Maybe. The thing is, it usually takes longer than you would like.
We already learned from Kia at Telia Finland that a business may need real estimations and real data before committing to anything. But let’s assume for a moment that you are able to produce authentic estimations, or are willing to give value to that business without asking for anything upfront. You still might have a good amount of waiting and working before you get confirmation from the company.
Product Owner for Nordea‘s Virtual Assistant Nova, Noora Vesterinen confirmed that the department she is working with does wish to collaborate with startups and hear new ideas. One small caveat though: “there are processes, so don’t get discouraged if it takes a while…if you can help us get started fast, that’s always a bonus”.
Kia added that lofty future plans of a dream AI program which wouldn’t lead to a return in the next few years are unlikely to garner attention. Focus on communicating and delivering practical AI solutions which will impact their present-day challenges.
Well, okay, how do I communicate that value to the clients? I’m not very into sales…
You must sell to make money. This was the exact issue Lars and his team faced early on. In fact, it wasn’t even until late last year that they hired their first-ever salesperson. For that reason, they struggled to communicate the value of their work to businesses.
Then how did they get people to work with them?
They focused on delivering value upfront as we already mentioned. Secondly, even though they didn’t have the best sales force, the value of their product got people talking. “When you deliver real value to enterprises, they talk together,” Lars said.
Now, if you really want to scale or if you have a friend or family member who is skilled at sales, by all means, go get that salesperson. Just remember to always make sure that you can back up what you say with real work.
Do you have any tips on contacting enterprises?
Again, be patient. Be clear on the value you deliver to the decision-maker and make it enticing and seamless for them to try your solution. Also, besides the empathy and understanding for your client, you will also need empathy and understanding for yourself. Lars Selsas notes that the first company you work with needs to be the “nice guy”.
As in, that business should be the one you believe will be patient with you, because you won’t know exactly what to do in the beginning.
That’s not to say that business needs to lower their standards, but you want to find an enterprise which is willing to tell you what they need and explain further if you don’t get it right immediately. After that first client, once you have real numbers and real use-cases, you have an opportunity to be less particular with what businesses you work with.
The privacy of your potential client’s customers. Like AI, the topic of privacy is very much on the minds of the general public. It’s not something your client can ignore, and for that reason, neither can you.
When you get that meeting with their decision-maker or contact, make sure to go in with an understanding of how you intend to address the privacy of their customers.
There you have a short summary of how you can start working with corporations and enterprises as an AI startup. Click here to watch the entire talk.