Parlai is a new type of todo list for sales professionals, one that literally writes itself. Parlai applies machine learning to email and CMS data to surface the most valuable contacts to target.
Imagine you work in sales and you are trying to stay on top of your overflowing inbox. Buried under happy hour emails from co-workers, weekly status reports, and reply-all madness, an email from an important lead lies forgotten, still waiting for your reply. How do you filter the important emails from the noise?
The target users of Parlai were individual sales representatives who used email as a means to communicate with customers. They are busy, often on the go, and highly motivated by anything that will help them reach their sales targets.
As part of the founding team, I was the only designer at Parlai and owned the entire design process from ideation, to user research, to visual design. I worked directly with the the VP of Product and the CEO to figure out product strategy and talk to customers.
By combining the information in our users' inboxes with the context around opportunities and accounts from Salesforce, we believed we could begin to surface the meaningful overlap between the two. If we could then layer on user behavior, the behavior of the sales targets, and the resulting changes in Salesforce, we could begin to build a better model for email management.
An early round of ideation around the problem centered around the idea of reminders - surfacing things that users may have forgotten to do in their email. These reminders would be centralized in a main feed, but also automatically grouped by lead or account.
The feedback we got from customers was that the reminders were interesting, but not all of them were relevant. Also, if the users didn't do them right away, it was easy to forget about them. Even if they snoozed them to postpone them, they got sick of snoozing them and seeing the same reminder over an over again.
How could we motivate users to complete the loop on reminders that were useful, and not get bothered by ones that weren't?
We re-focused the product around the idea of a task list, where the user can get the satisfaction of checking off items that were accomplished. Even better, because Parlai was integrated with the users' inboxes, we could automatically check off the item for them!
This is how we explained Parlai to users as they onboarded:
Here's how we explained it to potential customers:
Our adjustments to the product were gaining traction and leading to more regular usage. One user told a story about how he had forgotton to reply to a customer about meeting up for lunch, and how Parlai had helped him remember in time to still schedule the meeting.
However, users didn't love the fact that Parlai was yet one more browser tab to keep open. To help users could easily view their task list and inbox at the same time, we created the browser extension version of Parlai.
Until this point, it had been fastest for our development team to develop and iterate within a desktop app. However, now that we had a product direction that was gaining traction with the users, it seemed like the right time to test out mobile. Since sales representatives are often on the go and triaging emails on their phones, a mobile app seemed better suited to their jobs.
Mobile also gave us the opportunity to layer on some fun interactions to try to make the product even more sticky. Taking inspiration from popular dating apps, the user got "matched" with their recommendations and could either do it on the spot, swipe left to dismiss, or swipe right to save for later.
If you can't get enough, here's some more work.