Recommendations Design
Background recommendations & search
This piece is about some ideas that I had around content recommendations, and how a user might experience and interact with those. I wanted to think of these outside of the typical design flows you would find in an application.
It came about because of an idea that I wanted to develop further, that is the idea of a knowledge engine - including the concept of recommendations for team knowledge and communications.
But the goal was to have this kind of set of information made available without it being intrusive, distracting and/or annoying. Simple thing to achieve huh?
Design 1:
For this mock up I thought of a simple unobstrusive option for the user, taking inspiration from the product designer Maggie Appleton. When the user is typing content - for example in a google document, the browser application could 'listen', run a background search based on the meaning of the text, and then allow the user to choose to interact with the recommended items, by type, as and when they prefer.
Design 2:
This design is part of the same flow - by selecting an option the user could generate a pop up modal that contains related document snippets, links and people that are also working on similar items as needed. The user could then easily insert a backlink within their own document as a reference or citation.
This is closer to the original idea of a plugin extension to google docs and notion, but does not sit as a sidebar. I am a little unenthusiastic about style of navigation. That is because I feel this additional step makes it very forgettable by the majority of users. As the saying goes - out of sight, out of mind. But I would be happy to be proved wrong on this type of design.
Design 3:
This representation is intended as an exploration and visualisation canvas related to knowledge - again using the 'type' node design from the first design, but folds out to interactive visual representations of the associated data when it is possible.
It could show something like any of the following related to a highlighted text:
- a timeline representation of documents centered on a particular topic (eg: a project)
- the depth and breadth of knowledge related to a particular topic
- the connections between people and the topic highlighted.
Summary
These are examples of a few potential ways that users could have a subtle yet useful way to access and collate related knowledge data from different sources within existing workflows.
Internal document and knowledge search for productivity will be one of the biggest opportunities over the next few years. As mentioned on a16z blog around generative AI - the statement was that who owns the workflow will win the market. But with so many specialist products, perhaps having an overlay, that exists as a natural part of those workflows and connecting different systems together is an interesting proposition.
