AI Content Generation for IDFC Bank
Designing an AI system that helps employees generate content for marketing communication faster
June,2024
Project Overview
With the bank increasing its customer base rapidly, the bank wanted effective ways to deal with communication towards the customer.
Communication could be in the form of Whatsapp messages, E-mails, Google Ads, Facebook Ads and much more.
The problem statement
To equip IDFC FIRST Bank’s employees with something that could help them speed up the communication process due to the uncertain and fast nature of it.
Employees faced delays in accessing critical information due to fragmented and outdated systems.
Navigation across existing tools was unintuitive and inefficient.
Understanding the Problem
Outdated Content:
Important documents and information were often outdated, creating confusion and errors.
Time Wastage:
Employees spent excessive time locating the right data due to fragmented storage systems.
Inefficient Creation Process:
Manual content creation was time-consuming and error-prone.
Secondary Research
Amid the continuous advancements in artificial intelligence, ChatGPT, Claude AI and Google Bard, are some of the top conversational models that have developed for the everyday users.
Design Process
Brainstorming: Identified features aligned with user needs.
MoSCoW Prioritization: Categorized features into Must-Haves, Should-Haves, Could-Haves, and Won’t-Haves.
Wireframing and Prototyping: Designed and tested multiple iterations to refine usability.
Project Goals
Automate Content Creation
Leverage generative AI to produce accurate, relevant, and updated content effortlessly.
Personalised Employee Experience
Deliver a tailored interface and content recommendations based on individual roles.
Enhance Information Retrieval
Simplify how employees find, access, and use resources.
Ideation Process
For this particular project, we conducted an ideation workshop with the stakeholders. We used MoSCoW prioritization method to ideate and go ahead with our process.
Sketches
Early Sketches & Ideations
Low-Fidelity Wireframes
Exploration And Variations
Visual Language
ABCDEFGHIJKLMNOPQRSTUVWXYZ
abcdefghijklmnopqrstuvwxyz()&@#[]
1234567890?!/.:,;
Inter is our typeface.
Inter
Aa
Inter is a typeface carefully crafted & designed for computer screens. Inter features a tall x-height to aid in readability. It also has many OpenType features that can be used to tailor functionality - especially in the use of numbers and other special characters that we use throughout our application.
PRIMARY RED
#9B1E26
#F0DADC
BACKGROUND 1
#F6E5E5
BACKGROUND 2
#6E0A0F
SECONDARY RED
TINT 10
#FCF1F2
TINT 40
#FAB7BC
TINT 30
#FFE3E6
In-Tile Editing
AI-generated variants were designed as editable text boxes that can be reused as variables in restricted spaces. These tiles provide enough space for content while embedding essential features.
Each tile includes options to regenerate, edit, and favourite. Selecting a variant changes its state to "selected," highlighted by a colour change. In edit mode, the tile adopts a more saturated hue for clarity.
Dynamic Filter Bar
The variants page featured a filter bar for sorting options, but excessive action icons cluttered the variant tiles. To address this, the filter bar was made dynamic.
When no variant is selected, the bar displays general sorting options. Once a variant is selected, it updates to show specific actions like copy, feedback, and favorite, streamlining the user experience.
Collapsible
Navigation Panel
To maximise space for the main content, the left-side navigation was designed to be collapsible. This feature ensures that while navigation remains easily accessible, it can be hidden when not in use, allowing the generated content to take centre stage.
By saving screen real estate, this approach improves focus on the primary content while maintaining seamless access to other sections when needed.
Version History
Tabs were implemented to keep track of version history, making it easier to manage and review variations. Each tab represents a specific version, showing when it was created and how many variations exist.
This structure allows users to swiftly navigate between versions, providing a clear overview and seamless switching for better comparison and organization.
Key Learnings
User-Centered Design is Crucial
Understanding user needs and pain points through research was essential in creating solutions like collapsible navigation and dynamic filters that truly improved the experience.
Clarity and Simplicity Matter
Features like tabs for version history and collapsible navigation emphasised how streamlined designs can enhance usability without overwhelming users.
Efficiency is Key
By automating content generation and providing tools like version tracking, we significantly reduced user workload and improved task efficiency.
Iterative Feedback Improves Outcomes
Continuous testing and user feedback were critical in refining features, ensuring that every solution addressed real-world challenges.
Thank You