Find a topic that the group has knowledge within. The other alternative is to find a topic that interests the group.
Next, discuss among the group members what area of expertise the group wants to find a problem.
There is also an intermediate step where you need to discuss the idea of trying to invent something new or if you would like to improve an already existing idea.
Next, the group should look for the profitability of the problem to solve. If the problem is too niche or not important enough then it will not make enough money to warrant a solution.
Steps that I took to finding a problem!
I wanted to find a product that worked in the cyber-security field but since it was so general I needed to narrow it down a bit.
I also needed to think about what problem appears often to the every day person.
Problem that came into my head came from my image processing class where I thought that it would be cool if there was a facial recognition and motion activated door bell that acted as a door knob and a sensor that can take pictures of people who are trying to break into the house or steal packages from the front door.
I looked through the current solutions and found that the products are either on at all times or they are activated when a button is pushed or it activates through phone detection but some of these are either insecure or easily breakable.
I wanted to make a smart lock that could be tricked with jammers or by stealing the owner's phone. I only way to do this is by making sure the person entering the house is the person associated with the device or is a verified person!!
Solution Formulation!
The door bell could read faces and determine whether the person who came to the door is authorized and if they are not, the owner will be notified. Also, if there is a person who comes to the door and steals a package it will notify the owner and record.
This might be a stretch but there will need to be machine learning to verify if a person is verified or not so that may be super difficult.