As said earlier that the process is not very different from the traditional user-centred design process. But following design thinking framework definitely adds a structure to your problem-solving approach. The only difference I felt is the way designer needs to collaborate with the product management and engineering and the kind of question we should ask to understand why Machine learning? How it is going to solve users and business problem uniquely. And collaboration, I feel it’s an extremely critical factor here because you are trying to augment the machine for user needs. So as a designer, you need to think from the users’ and machines’ perspective, while keeping the business goal in mind.
So there is a lot of alignment that you have to juggle with. But yeah, it’s fun. So as long as we keep things and all these questions in mind, the process again follows, it’s a standard design thinking process, there is nothing different, apart from the questions. The focus and parameters I talked about — first you have to know the business and user goal, understand user needs their pain points, what define and this is something I would say the additional thing that we have to understand in machine learning is define what user intent and intent I talked about earlier, what kind of intent I’m saying, define a design guideline guiding principles for ML experience, for example, like hyper personalization or Intelligent Automation, or trust, meaningful data is. Of course, trust and meaningful data are the heart and soul of it, and it should be there. But to keep you aligned when you are going through this design process. These are the design guiding principles that keep you aligned with that. Next is identifying what type of data we need, and what we have then I also work with a product team actually to prepare the experience for training a model.
So how do you train the model, you actually have to prepare a checklist which when you train the model, it goes through that checklist and then it confirms that whatever data you have is a good data and it can train your model on that. So something like that you need that guidance, looking at the end to end customer journey. And identifying the potential to embedded machine learning? So where can you really add machine learning in that normal customer journey where it can amplify the user’s performance? Identifying what the big idea is? So as I talked about, like, what’s the theme? What is that big idea around that you are building this experience? So as I say, like maybe, for example, hyper-personalization, so then everything, run or everything, basically thought through around that one point, and then the dead data, everything you need around that.
Another thing is identifying patterns in the data. So this is something beneficial for us to understand what type of patterns as you also scale the design, and then the standard design phase, where you actually Lo-Fi & Hi-Fi mockups and paste those before you actually get into development, because it is very critical to test it before you actually model it. Because again, the elementary fact is you don’t want to develop something which is not working for users, right? And the cost factor. So testing, refining, and then standardizing once you refine based on user feedback standardize it again, when I say standardized it is basically thinking about a design system and all of that, and then develop and the same process, every cycle, you keep rotating or keep doing the same process. So that is the process which I follow.