Solving for conversation context
During the workday, we spend at least 2 hours reading through collaboration apps like Webex, Teams, and Slack just to get through our tasks for the day. One of the key areas that my team was suffering was keeping up with the number of discussions and information shared on these channels while being productive. Most of the applications that we work with are Figma, JIRA, Confluence, and Outlook. Of course, there are more applications but the conversations revolve more or less around these.
​
This micro project is a personal one where I am looking at ways to solve chaos around conversation and have a more streamlined information flow to ease out daily tasks.
Over sharing of infomation
With discussions being evident for every small topic, it was difficult to keep up with the conversation on different chats. Sometimes, it would lead to discussing the same topic over and again.
Multi-channel information
Even though JIRA and Figma were standard apps to follow as guidance. There was nothing that could be a considered source of truth. This itself created confusion on what was the latest change on a particular scope/story.
Information continuity
If a leave was taken or a meeting was missed, it was difficult to keep up with what happened. There would be notes but not helpful enough to maintain the continuity of conversations.
AI Aided Flows
Using AI as a tool to enhance productivity and provide context
With ChatGPT and other ML models in full swing, this was the perfect opportunity to reuse it to provide context and streamline the process. These models can also identify media assets and hence more beneficial for this kind of problem.
​
Below are 2 different suggestive ways of re-using ML to solve my productivity issues.
Q&A bot on the collaboration app
Solution 1
Pretty much with current situations, we were already asking other colleagues about updates and figuring out information from different channels. The solution here revolved around having a bot that would understand the context and fetch info to the person. Very similar to Chat GPT but with more links, media references, etc.
Let's take a situation here
​
' Before going on your vacation, you had an important meeting and once you are back, there are multi-thread conversations across multiple channels suggesting changes needed, meeting recordings, etc.'
​
This is usually an anxious situation and creates mental fatigue. Also to add, the whole exercise of getting back on track takes at least 2 business days. The flow below explains how this bot could take over.
Task-based cards on the app
Solution 2
From my learning, I have noticed that when information is made into a checklist, people can grasp that better. Many employees also follow this offline - they usually have a handwritten to-do list or put down something similar on their notepad.
The concept of this solution is to extend that with more context so that the person is more aware of whom to contact, the next steps, any media assets related, etc.
​
The flow below explains how these task cards are beneficial to the employee and can speed up their day-to-day work.
Pre-requisite
This concept considers applications like Outlook, JIRA, Figma, etc integrated with the app and allows the exchange of data within the applications.
Learnings from the exercise
1. If a lot of information is also presented to people it can feel fatigued and overwhelming. Hence, it needs to follow a structure.
2. With reading I found out that the ML models are not yet stable enough to have summarisation across applications and this could take a while to be on the market.
3. There is also situational information that becomes a priority - for example - what kind of information is the user looking for when they are asking for context is also something that needs to be explored.
4. Overall, when done in phases - this can certainly elevate productivity.