Akooda is an AI-powered platform that helps organizations unlock company-wide insights while collecting critical interactions, conversations, and tasks that happen inside, outside, and in between the SaaS platforms to give users a full picture of the efforts (and gaps) across the organization.
About The Project
Project Scope
Develop an AI-powered personalized summary tool that analyzes interactions, incorporates sentiment, and delivers trustworthy, contextually relevant answers derived from the source of truth in seconds.
Process
In an agile environment, this project was organized into two-week sprints. To launch the feature quickly, teams had to work in parallel, which proved challenging as new discoveries and insights frequently emerged throughout the process.
Challenges
What are we solving?
-
Summarizing the entire Interactions Drawer can take over a minute.
-
Balancing answer length with content quality due to token limitations.
-
Incorporating sentiment analysis.
-
Ensuring the credibility of answers based on sources.
-
AI feature should be easily discoverable, but seamlessly integrated without being obtrusive.
How are we solving?
-
Summarizing interactions per page (20 at a time) saved time and tokens, while keeping answers concise, easy to read, and maintaining a high-quality context from the sources.
-
By considering the user’s role and their involvement in projects and interactions, we provided personalized sentiment analysis to ensure content relevance to their daily work.
Research Popular AI Tools
We’ve reviewed the most popular AI tools and identified common features and patterns that users typically expect in these services.
Feature Location
We started by focusing on two areas of our platform for this feature:
Organization Profile Page
Interactions Drawer
*** accessible throughout every page in the platform
Bird-eye view of a project, topic, or account they are involved in.
01 Comprehensive Overview
Highlights the status of various topics and provides updates on upcoming or ongoing action items.
02 Personalized Analysis
Get the next steps and important insights based on specific interactions.
03 Key Insights
Ask open-ended questions in search bar.
04 Ask Anything (Future)
User Feedback & Pain Points
As a small startup, we have the advantage of engaging with our clients on a bi-weekly basis. During these meetings, we asked clients to share their experiences using Akooda and suggest areas for improvement. What users wished they could do:
In the first phase, we chose to focus on options 1, 2, 3 as a baseline, aiming to deliver the best user experience in a short timeframe.
Feature location - Interactions Drawer
AI Branding & Components Explorations
Hi-Fi Mock-Ups
To verify these designs, I created a few quick Hi-Fi mock-ups to test ease of use, clarity, visibility and the space the answer and data takes altogether.
Organization Profile Page
Top Card
Interactions Drawer
Table Tabs
Organization Profile Page
Side Card
Organization Profile Page
New Tab
Interactions Drawer
Switch
Interactions Drawer
Button
Interactions Drawer
Toggle Buttons
Chosen Designs
After gathering internal feedback from the team and from a few of our loyal clients, we decided to continue with these 2 directions:
Organization Profile Page
Interactions Drawer
Interactions Drawer - AI Answer Refinements
Deliverables
After 2 months of work, we launched Akooda AI as a BETA feature.
Organization Profile Page
Slack Bot
Interactions Drawer
At-a-glance to the next AI feature designed for Akooda
We're developing a Chrome extension to make it easy for our users to access insights and data directly from their browser. By integrating with their workspace, we streamline their workflow without the need to switch platforms.
Exploration
Feature location - Organization Profile Page
Key Learnings & Next Steps
During the BETA phase, we gathered feedback from users through the feedback form and bi-weekly meetings with our champion clients. Here’s some of the feedback we received:
With the BETA being very successful, the team is constantly improving the AI features in the platform to give users the best and seamless experience as possible.
Users wish to be able to scan the text quickly and get to insights fast.
01 Bullet Points
Users were still interested in asking open ended questions to get to the answers faster.
02 Open-Ended Questions
To get a better context, users asked to include hyperlinks to relevant sources within the textual answer.
03 Add Hyperlinks
User wish to share their answers with their colleagues in different apps such as Slack, email, etc.
04 Share Answers