How might we help users with repetitive tasks?

So, what's the problem? The research and discovery phase

Our Chat product has had a lot of focus in recent months as we look to really grow (and sell) the product more. Our sales and customer success team in particular have really been knocking it out the park.

Usage of the platform is increasing month on month, but also, the complexity of usage has grown with it. Our customers have been expanding the product across multiple teams within the institutions, which has thrown up some weaknesses within the product. To better understand these challenges, we interviewed agents and customer success teams. Key pain points that emerged from these discussions were:

• Search efficiency: Agents found it difficult to search and retrieve past conversations efficiently.
• Repetitive actions: Agents had to manually close, assign, or tag conversations one by one, leading to significant time lost in managing multiple chats.

Ok, so what next? The ideation phase

The first port of call for us was again to get together as a team - myself, our product manager and lead engineer for the Chat product - and map out the expected user flow. This allowed us to identify any pitfalls early and ensure that all implications, from both users and engineering point of view, were considered.

Armed with knowledge gleaned from our customers and engineers I spent some time researching other products - search and actioning is a fairly well trodden path and I was curious to see how others were doing it, and looking for consistencies.

In general I indentified some obvious design patters:

• Search bar with filters: A persistent search bar at the top of the browser window, with easy-to-use filters for various criteria.
• Multi-select feature: Checkboxes beside each row in a table allowing users to select multiple results.
• Action dropdown: A dropdown menu enabling mass actions.

This research enabled me to quickly come up with some ideas...

The solution The design and iteration phase

The outcome of the various brainstorming sessions and customer calls was that the focus would be on 3 main areas - ease of search, filtering results and a new search results page.

Ease of search

In-keeping with our mantra of ensuring our customers have the most efficient working practices we introduced shortcut commands, CMD + K is introduced as a means of quickly searching. This coupled with the re-positioning of the search input means users can search for their data no matter where they are in the product.

Filtering results

A more robust searching experience. This included the ability to pre-filter your search results, for example you could search within Conversations and specifically search within Channels or from certain individuals. We will also surface recent or frequent searches to the user. I also love the randomised tips suggested to the user 😍.

Search results page

The search results now have their own dedicated page instead of the restrictive right hand panel view. This opens up so many possibilities when it comes to actioning search data!

The search results page allows users to:

• Break down results into individual areas such as Conversations, Agents, Contacts etc. The reason for this is that there are different actions you can perform for each individual area, so having a table of mixed results wouldn’t make much sense and only lead to confusion - it would also create more of a technical challenge and slow down delivery.
• View results in a table format for ease of scanning, filtering, sorting and actioning.
• Filter results, allowing the user to really drill down to the data they require.
• Select required results and perform mass actions.

Unfortunately, we did run into some stumbling blocks during development. We hadn't forseen that we may run into performance issues when it come to mass actions. The initial plan was for us to have the ability to mass action all search results, but we hadn't taken into account the sheer volume of data that some customers have. Therefore we had to make a call and reduce the mass actions from unlimited to 50 at a time. This isn't ideal but it is still a significant improvement for our users.

Final thoughts Outcome, reflection and takeaways

Outcome

As part of the implementation process we ensured analytics were installed on the new search bar and results page. We revisited the analytics after 3 months and were delighted with the results.

• Over 8x increase on users searching within the product. Users had mostly given up using the search facility due to it being ineffective. We've seen a huge upturn in usage with the more functional process.
• Users have reported tasks which could take days are now done in hours (sometimes less).
• The ability to manage conversations in bulk meant agents could focus more on delivering quality customer support rather than administrative tasks.

Takeaways

The main takeaways from the project were:

• Utilising known patterns can significantly speed up design and development time, and also ensures a known experience for users. As designers we always want to push new designs but often the answers are already out there and we shouldn't be scared to borrow from other products. Especially if it gets new features in the hands of our customers quicker.
• In hindsight, we should have spent more time speaking with the engineering team to probe them fpr any pitfalls that might occur, especially when it comes to the mass actions.