Concept - AI assisted commenting in report
Adobe, 2023
Overview
This initiative was a Garage Week project, is akin to a hackathon endeavor. Coordinated by myself and another design manager, our focus is on distinct products that delve into data and analytics from various perspectives. Despite serving different user groups, the primary users of these products find it crucial to collaborate regularly for data insights and decision-making. Recognizing this need, we've seized the Garage Week opportunity to explore design solutions aimed at using AI as a collaborator to enable smooth communication between Analysts & Marketers across products.
Team & Role
Team: Me + 1 Design manager + 1 Junior designer
Role: Recruit internal user, User research, User story, Wireframes, UI/UX Design
Timeline: 1 week
The Problem
A marketer wants to get some data & insights to answer certain questions, but getting support from data team is hard because they need to prove that this request is worth working on to the data team while they are lack of data evidence. How can we give them a way to easily get answers to some simple data questions and make the process of getting data support more transparent so this marketer doesn’t feel frustrated when data team cannot help with her ticket?
Despite the tight one-week timeline from kick-off to final showcase for this project, a dedicated application of the user-centered design process was ensured.
Collaborating with an internal marketer who generously agreed to assist, I conducted a comprehensive user interview. This involved delving into her current processes, identifying pain points in requesting reports from the data team, and gathering feedback on the initial user story. A joint effort was then undertaken to analyze, synthesize insights, and determine the core problem to address in this project.
The design phase started with a focus on three key screens, progressing to wireframes that effectively conveyed the user journey. Subsequently, a design review involving stakeholders from both the design and product teams was conducted. Further validation came through additional user testing to confirm the usability and conceptual integrity before finalizing high-fidelity screens. With positive feedback received, the project was presented to design leadership, paving the way for discussions on the next steps in the project.
Design Process
Key Insights From Research
“It’s kind of like a chicken and egg conundrum where like we we need to have data support but in order to get data support and resourcing on a particular project we need to prove the viability of the of the the launch that we’re running and without the data support we can’t really do that.”
Jumping between tools for communication and data.
Requesting a new report is a time-consuming process due to high communication effort.
Opaque process with uncertain timeline.
User need data to prove the impact of the request to have those data partners prioritize their different requests.
Lack of self serve data tool.
Existing advanced analysis tool is not user friendly to business users.
Data format of returned report is not predictable and not editable.
Not easy to share report & insights from the existing products.
Analysts are occupied with simple requests when complex requests are deprioritized.
Brainstorming & Wireframes
Solution highlights
Leverage AI for simple analysis without expert's help.
To free up analysts’ time on the small simple tasks, AI assist takes the requests from marketers to answer simple data analysis questions.
2. AI assisted ticket creation without leaving the context.
To shorten the time on back and force communications and jumping between multiple tools, AI detects the potential tasks in the dialog and provide quick action to create the workfront ticket without leaving the context.
3. Add new chart to the existing report.
Whenever there is a new chart provided by the analysts or AI assit, the user can add the chart as a new widget to the report for the record and make it visible to other users.
4. Share the data with insights.
Users have full control of the insights they would like highlight from the data, and share them with the stakeholders when needed.
Reflection
This endeavor served as a valuable exercise in implementing user-centered design within a lean design process. Despite being completed in just one week, the results exceeded expectations. Engaging with a single user yielded crucial insights, and the rapid iteration based on user feedback was remarkable. The experience not only bolstered my confidence in applying this process but also highlighted its efficacy in addressing other identified challenges. Moving forward, I plan to conduct more workshops, concentrating on one problem at a time, with the aim of delivering comprehensive solutions to complex issues. I eagerly anticipate navigating through this process to provide impactful solutions in the future.