November 22, 2019
Artificial Intelligence market is taken as a Trillion Dollar industry but the question that comes in our mind first is, how to avail this diamond opportunity !!!
Here are five right hand finger tips for you.
1.What to do - Start from Yourself
Who are you? A person who has little bit knowledge about market, a little knowledge about IT and a little bit knowledge about Tech Entrepreneurship. Starting from yourself is best practice of any innovation helps you to maintain your long term interest and gives you good start by keeping yourself safe from several initial barriers. After knowing your interest you have to find your wings (available resources / opportunity) around your circle to fly quickly. Let us suppose, you are a son of an eye-specialist, the first thing that comes in mind is the field of medical imaging. Working in domain of eye disease detection (or similar) will be best starting point for you to take off due to two major reasons. The first one is you don’t need to worry about domain/field knowledge (you papa is there for you , the second is most probably you can get realistic dataset more easily (again you papa will help you.
Similarly, if you are a shop-keeper or have a small e commerce website, you are already good at demand-supply management, price optimization, promotional offers, customer caring etc. So, you are ready to jump into retail industry's automation due to two above described two reasons.
Still confused? You think that you have no domain knowledge or no smart social or family circle to take off? You think you are a super lazy baby? Then, I think you are an unbeatable super-man. Why? Just start to think that how you can make this world lazier like you by automating everything after applying artificial intelligence. Even if you have wasted your whole life playing PUBG or other games, start to think how to make computer players more intelligent by applying reinforcement learning.
So, Start to think, start to do what you like to think what you like to do!
2. How to do – See what others have done so far
This world has population of more than 7 billion human-being. There are big chances the idea you have, has already been thought or implemented by someone in the world. So without wasting your time just go to Doctor (Google) for check up and tell all your whole story for recommendations.
Mr Google will facilitate you with certain relevant links, just open them and see what people already in market are doing related to it, short-list them – sort them.
Here the real game starts. Forget you are an AI expert or IT expert or whatever just become a Stupid Customer. Get trial versions of these products, use it, critically evaluate it, make calls to support department to ask more depth questions about their product. Because there are very minor points that you do not consider usually but they become bottleneck for you if you do not completely scan other people’s work. So Think Like a Thief (TLT) at this moment.
At this stage, you have domain knowledge, you have relevant data and you have other competitors’ work related to your idea, you have critically analyzed them. Now, keep the holes in existing solution in front of yourself and start to think about covering these gaps – the more refined ideas will start to reveal in your mind and you are done with Business Diligence.
More you find market holes, more you capture weaknesses of market, stronger your product will become!
3. Start to do – Integrate a team & differentiate their roles
Under the umbrella of Artificial Intelligence, there are a lot of branches to learn like Natural Language Processing (NLP), Computer Vision (CV), Reasoning, Planning etc. Now, on the basis of your idea and market research pick one branch and start to learn relevant algorithms, technically. You should learn something technically about your field even if your role is only leading the team and managing the product. A very basic program in Python takes around 100 lines of code and 1 week of time to execute a pilot task. So go for it first!
After that your job is to build a team. When the word team comes, we start to think about a cricket team of 11 players plus 4 extra with proper uniform and monthly salary. No, you can build a team with yourself, you can build a team with your one or many friends, or you can build a team with hundreds of employees. Its up to your capacity, financial capability or timeline to reach goals. But for instance, we take the above described example about someone’s eye-specialist father, in this case his father is also a team member having role of Domain Knowledge Expert (DKE) and is a Data Provider (DP). Similarly, one of your friends might be good in marketing, he is also your team member providing you access and giving you consultancy related to client dealing or product selling.
Formally, you need the role of people like Market Researcher, AI Consultant, Data Engineer, Machine Learning / Deep Learning Engineer, Software Engineer, Testing Engineer, Product Manager. You can merge the role of 2 or 3 persons into one or you can hire 2 or 3 persons for one role; its up to you.
Now the next question comes, which one software design methodology should you adapt for implementation. Unfortunately traditional waterfall methodology is difficult to implement here due to numerous reasons. So, in AI we usually go for Agile methodology due to certain reasons like iterative sprints, flexible planning, unit testing etc. In AI algorithm implementation, you do not know which one technique is going to give you best results so you should be very flexible in implementation through hit and try process to fine tune your implemented algorithms. That’s why the experts prefer Agile methodology in this field.
It is also important to note that what you are going to do will achieve enough accuracy that market will adopt your product easily. At this stage you have to completely find out technical diligence of your system that you are going to launch in market.
4. Start to shout loudly - Launch in Market
Take a step back once your product is ready for the market, and do not launch it widely. Offer your product to selected customers as a pilot so that you can better understand market usage and work out bugs — AI is a new world for your customers, too and work on case studies that will be needed for wider market adoption. Once you are confident enough with your product, start to raise your voice loudly. Use Search Engines, Social Medias and other platforms for marketing and try to sale at least 1 unit of your product; that’s the success!
Even if you need investors to further launch your product on bigger scale you should sale out at least a few units that builds the confidence of investor on you and your product.
5. Patience, Patience, Patience
This shouldn’t be the end, though. Your AI solution should continue to grow and be iterated upon, especially since you already have the team and infrastructure in place. Don’t focus on who clap on you but focus on those who do not clap and ready to accept negative feedback with patience. This will, ultimately, be helpful for not only your product’s growth but also for your personality’s growth. Since you are following agile methodologies, you have to complete one cycle till launching the product in market and then again you have to start this development cycle on the basis of market’s feedback. This process keeps on running that is the Research & Development.