The crux is this – as designers and product managers we need to explore the idea space offered by today’s Information and Communication technology to ensure productivity growth – through design driven innovation (among other means).
In other words, we need a 0.5 to 1 approach. Peter Thiel differentiates between zero to one and one to n innovations. The former is creating something new and the latter is just copying what already worked. But I think we have another space. At 0.5 level, the technology is present, but the applications are not fully realized.
Today, these technologies are Machine Learning and Blockchain. But even the internet in general (social, mobile) might offer more areas for good products – like a social network for maximalist careers.
We need to look at technology as an key ingredient of products, and not as a flavor – ‘added on top’. Many startups brand themselves as [x] + AI or [x] + blockchain, but that is incremental thinking. (Even while we compare AI to electricity, we almost never think of a product as [x] + electricity, because we know that electricity is just there). This ‘add on top’ thinking gives rise to the idea of ‘assistants’ – which is surely not the best we can do with these powerful algorithms. Rather than have an assistant help me, I would rather have tools which make me better at my work.
So to really see how algorithms could make our lives better, we need to think about specific and niche use cases, while being aware of the general capabilities of machine learning algorithms – like being effective at recommending content.
But while we see this power in consumer internet products, we don’t see them enough in productivity software. Why is there little or no use of these capabilities in MS Excel, WordPress, Sketch, Intuit, Evernote or even Google Analytics? Apart from Google Search and Gmail (somewhat), I find it hard to see applications of machine learning in productivity software. There are attempts, but solutions are not commonplace yet.
Let’s briefly outline a specific case to understand this better. If you were designing a UX design tool from scratch, how would you utilize the power of algorithms? (I don’t know of any UX design tool doing this, yet).
I start with my own workflow. When I design, I draw sketches on paper as well as make wireframes in design software to clarify ideas. Then I usually look for inspiration on sites like dribbble or behance to see if there is a better visual implementation of the same idea. Or, are there some new peripheral ideas related to the problem I am solving. Sometimes I consult sites like Useronboard.
What if I had a design tool which would take my sketch as an input, and gave me ‘recommended’ designs based on the sketch? This would quickly speed up my workflow and make the output better.
Similarly, there could be an algorithm based ‘design-critic’, sort of a spell-check for wireframes and visual designs. We need a lot of design critique and collaboration because one person alone can fail to see key aspects of a design (say, accessibility). What if the design tool can highlight these ‘errors’ in a design and ‘suggest’ improvements and examples?
Google can read all the text in websites, and also the images, surely advanced machine learning ability can also ‘read’ the design of the best designed apps and websites?
This idea is just a hint. It needs more working out, but this could very well be a valuable company which no one is building (yet).
I am quite sure if we investigate the detailed workflow of financial analysts, radiologists, or marketers, we would come across opportunities to create products which make use of algorithms, but with the people and their workflow at the centre – and this would give us better ideas
to apply these technologies for breakthrough products.
Adding algorithms ‘on top’ of products is the low-hanging fruit of innovation, and no harm in getting that first. But to get to productivity growth levels of the golden age, we need products which have machine learning technology as the ingredient. Products need to be re-imagined, and not just improved.
And that’s where we need design-driven innovation.