Cora
Retail Media Network
AI Agent
Cora
Retail Media Network
AI Agent
Cora
Retail Media Network
AI Agent
Cora
Cora
Cora was an internal RFP project I developed for Accenture. It is an agentic AI platform that helps retail media managers and clients create, run, and manage the full ad cycle from start to finish with minimal human intervention.
Cora was an internal RFP project I developed for Accenture. It is an agentic AI platform that helps retail media managers and clients create, run, and manage the full ad cycle from start to finish with minimal human intervention.
Role
Role
I collaborated with two other designers on the team. As the UX Lead, I was responsible for conducting research, identifying user pain points, and solving underlying architectural issues for creating and optimize the cora.
I collaborated with two other designers on the team. As the UX Lead, I was responsible for conducting research, identifying user pain points, and solving underlying architectural issues for creating and optimize the cora.
Goal
Goal
My goal was to create and optimize an AI agent platform that empowers human agents with intelligent suggestions and recommendations, allowing them to autonomously build and manage ad campaigns from launch to completion.
My goal was to create and optimize an AI agent platform that empowers human agents with intelligent suggestions and recommendations, allowing them to autonomously build and manage ad campaigns from launch to completion.
Duration
Duration
4 weeks
Problem statement
Retail media managers encounter an overwhelming manual workflow that leaves them flying blind, effectively stripping them of the time and data needed to execute high-level strategy or scale effectively.
Retail media managers encounter an overwhelming manual workflow that leaves them flying blind, effectively stripping them of the time and data needed to execute high-level strategy or scale effectively.
Metrics
Human Intervention
Human Intervention
80%
80%
Delay Time
Delay Time
40%
40%
Note - You can find the Impact Metrics (result) at the end of case study
Lets study the user first
Lets study the user first
User Persona
User Persona
Insights
Insights
Current process
So Basically
So Basically
The Retail media cycle involve too many step and it takes to much time to do it right.
A persistent lack of clarity on why a campaign succeeded or failed, making it impossible to scale.
Zero visibility into live optimization, leaving managers to guess what’s actually driving results.
Media manager have to closely monitor bids, audiences, pacing daily.
The Retail media cycle involve too many step and it takes to much time to do it right.
A persistent lack of clarity on why a campaign succeeded or failed, making it impossible to scale.
Zero visibility into live optimization, leaving managers to guess what’s actually driving results.
Media manager have to closely monitor bids, audiences, pacing daily.
Potential Solution
An AI Campaign Orchestrator that automates the entire lifecycle from market research and creative generation to approval tracking and real-time performance diagnostics.
An AI Campaign Orchestrator that automates the entire lifecycle from market research and creative generation to approval tracking and real-time performance diagnostics.
An AI Teammate that Does the Heavy Lifting
We’re building an AI Campaign Orchestrator to act as a smart assistant for the team. It doesn't just "help" it takes over the tedious stuff:
automatically scans market trends
generates ad ideas and visuals
tracks all approvals and feedback
tells you exactly why a KPI is down and how to pivot.
An AI Teammate that Does the Heavy Lifting
We’re building an AI Campaign Orchestrator to act as a smart assistant for the team. It doesn't just "help" it takes over the tedious stuff:
automatically scans market trends
generates ad ideas and visuals
tracks all approvals and feedback
tells you exactly why a KPI is down and how to pivot.
AI World
AI World
Bench marking
Bench marking
Trying understand more about AI Agents
Agents versus Chatbots
Agents versus Chatbots
Agents
The seasoned chef (agent) possesses years of culinary experience and intuition. Armed with a general understanding of your preferences and a brief description of available ingredients, they can whip up a delicious meal that caters to your needs. The exact steps might vary each time, and each version of the dish might have subtle differences, but the overall outcome is consistently satisfying. Likewise, an agent can adapt its approach based on the context and intent of the user's input, resulting in a successful interaction.
The seasoned chef (agent) possesses years of culinary experience and intuition. Armed with a general understanding of your preferences and a brief description of available ingredients, they can whip up a delicious meal that caters to your needs. The exact steps might vary each time, and each version of the dish might have subtle differences, but the overall outcome is consistently satisfying. Likewise, an agent can adapt its approach based on the context and intent of the user's input, resulting in a successful interaction.
The seasoned chef (agent) possesses years of culinary experience and intuition. Armed with a general understanding of your preferences and a brief description of available ingredients, they can whip up a delicious meal that caters to your needs. The exact steps might vary each time, and each version of the dish might have subtle differences, but the overall outcome is consistently satisfying. Likewise, an agent can adapt its approach based on the context and intent of the user's input, resulting in a successful interaction.
Chatbots
The novice cook (chatbot) relies heavily on a detailed recipe, complete with precise measurements, step-by-step instructions, and specific cooking times. Any deviation from the recipe results in a culinary disaster. Similarly, a chatbot must function within the confines of its pre-programmed decision tree.
The novice cook (chatbot) relies heavily on a detailed recipe, complete with precise measurements, step-by-step instructions, and specific cooking times. Any deviation from the recipe results in a culinary disaster. Similarly, a chatbot must function within the confines of its pre-programmed decision tree.
The novice cook (chatbot) relies heavily on a detailed recipe, complete with precise measurements, step-by-step instructions, and specific cooking times. Any deviation from the recipe results in a culinary disaster. Similarly, a chatbot must function within the confines of its pre-programmed decision tree.
AAA Washington case study
AAA Washington case study is a perfect example how to use AI as agents to solve problem with low stress level. AAA Washington realized that when someone’s car breaks down, they don’t want to wait on hold, they just want to know help is coming.
By using Salesforce’s AI agents, they’ve automated the 'stressful' part of the process. Now, an AI agent can instantly dispatch a tow truck and send real-time updates to the driver’s phone. This clears the lines so that when a member truly needs to hear a human voice for a complex emergency, a real person is actually available to answer.
AAA Washington case study is a perfect example how to use AI as agents to solve problem with low stress level. AAA Washington realized that when someone’s car breaks down, they don’t want to wait on hold, they just want to know help is coming.
By using Salesforce’s AI agents, they’ve automated the 'stressful' part of the process. Now, an AI agent can instantly dispatch a tow truck and send real-time updates to the driver’s phone. This clears the lines so that when a member truly needs to hear a human voice for a complex emergency, a real person is actually available to answer.
AAA Washington case study is a perfect example how to use AI as agents to solve problem with low stress level. AAA Washington realized that when someone’s car breaks down, they don’t want to wait on hold, they just want to know help is coming.
By using Salesforce’s AI agents, they’ve automated the 'stressful' part of the process. Now, an AI agent can instantly dispatch a tow truck and send real-time updates to the driver’s phone. This clears the lines so that when a member truly needs to hear a human voice for a complex emergency, a real person is actually available to answer.
Microsoft Copilot
Think of it this way: if Copilot is your personal assistant who helps you write emails and summarize meetings, AI Agents are your specialized teammates who actually go out and get the work done.
In the Microsoft world, they’ve broken these "digital coworkers" down into a few main types based on how much they can do on their own:
Reactive Agents: These are "if this, then that" helpers that follow strict, set rules to handle repetitive tasks instantly.
Goal-Based & Utility Agents: These are smarter assistants that weigh different options to find the most efficient way to reach a specific target you’ve set.
Autonomous Agents: These are high-level problem solvers that can plan, adapt, and work with other systems to handle complex, multi-step projects on their own.
Think of it this way: if Copilot is your personal assistant who helps you write emails and summarize meetings, AI Agents are your specialized teammates who actually go out and get the work done.
In the Microsoft world, they’ve broken these "digital coworkers" down into a few main types based on how much they can do on their own:
Reactive Agents: These are "if this, then that" helpers that follow strict, set rules to handle repetitive tasks instantly.
Goal-Based & Utility Agents: These are smarter assistants that weigh different options to find the most efficient way to reach a specific target you’ve set.
Autonomous Agents: These are high-level problem solvers that can plan, adapt, and work with other systems to handle complex, multi-step projects on their own.
Think of it this way: if Copilot is your personal assistant who helps you write emails and summarize meetings, AI Agents are your specialized teammates who actually go out and get the work done.
In the Microsoft world, they’ve broken these "digital coworkers" down into a few main types based on how much they can do on their own:
Reactive Agents: These are "if this, then that" helpers that follow strict, set rules to handle repetitive tasks instantly.
Goal-Based & Utility Agents: These are smarter assistants that weigh different options to find the most efficient way to reach a specific target you’ve set.
Autonomous Agents: These are high-level problem solvers that can plan, adapt, and work with other systems to handle complex, multi-step projects on their own.
Takeaways
After diving into Agentic AI casestudy and painpoints that we already have, I realized how much of the "busy work" we can actually hand off to a smart assistant. Whether it’s scanning the market, chasing approvals, or even drafting creatives that match a specific style, AI can handle the repetitive stuff. Now I will try to mapped out the key areas where we can let AI take the lead, so we can get back to the work that actually requires a human touch.
After diving into Agentic AI casestudy and painpoints that we already have, I realized how much of the "busy work" we can actually hand off to a smart assistant. Whether it’s scanning the market, chasing approvals, or even drafting creatives that match a specific style, AI can handle the repetitive stuff. Now I will try to mapped out the key areas where we can let AI take the lead, so we can get back to the work that actually requires a human touch.
After diving into Agentic AI casestudy and painpoints that we already have, I realized how much of the "busy work" we can actually hand off to a smart assistant. Whether it’s scanning the market, chasing approvals, or even drafting creatives that match a specific style, AI can handle the repetitive stuff. Now I will try to mapped out the key areas where we can let AI take the lead, so we can get back to the work that actually requires a human touch.
Proposed user flow
AI intervention
Human intervention
Wireframes
Following several brainstorming sessions and initial wireframe iterations, we aligned on these final concepts.
Lets check the product now
Upload the Documents and Cura will create a summary for confirmation before proceeding with the Campign proposal.
Upload the Documents and Cura will create a summary for confirmation before proceeding with the Campign proposal.
Confirm and with in few mins, the whole proposal will be ready with budget details, audience details, visuals and KPIs. You can iterate as much as you want at this step.
Confirm and with in few mins, the whole proposal will be ready with budget details, audience details, visuals and KPIs. You can iterate as much as you want at this step.
Wants to make changes talk to cura, or just click on edit for particular sections.
Wants to make changes talk to cura, or just click on edit for particular sections.
Cura is smart so as you make changes cura will provide you with additional market-driven recommendations.
Cura is smart so as you make changes cura will provide you with additional market-driven recommendations.
Once satisfied, Cura will ask for your permission to send the proposal for feedback and approval. It also provides an estimated response time and helps keep everyone accountable.
Once satisfied, Cura will ask for your permission to send the proposal for feedback and approval. It also provides an estimated response time and helps keep everyone accountable.
Once the proposal is reviewed, approvals can give feedback on specific sections and send for launch once the changes are implemented.
Once the proposal is reviewed, approvals can give feedback on specific sections and send for launch once the changes are implemented.
Post-launch, Cura provides a performance overview and suggestions to optimize your campaigns.
Post-launch, Cura provides a performance overview and suggestions to optimize your campaigns.
Conlusion
We successfully built Cura, a smart AI partner that strips away the heavy manual labor from the ad campaign lifecycle. By replacing hours of tedious work with data-backed recommendations and market-driven insights, Cura ensures every campaign is built on a solid, human-centric foundation. It handles the heavy lifting from scanning the market and generating designs to managing approvals and providing post-launch optimization allowing retail managers to move from "flying blind" to focusing on high-level strategy. This shift gives them the time and clarity needed to make the critical decisions that drive true campaign success.
We successfully built Cura, a smart AI partner that strips away the heavy manual labor from the ad campaign lifecycle. By replacing hours of tedious work with data-backed recommendations and market-driven insights, Cura ensures every campaign is built on a solid, human-centric foundation. It handles the heavy lifting from scanning the market and generating designs to managing approvals and providing post-launch optimization allowing retail managers to move from "flying blind" to focusing on high-level strategy. This shift gives them the time and clarity needed to make the critical decisions that drive true campaign success.
Metrics
Human Intervention
Human Intervention
<10%-15%
<10%-15%
Delay Time
Delay Time
<10%
<10%
Take away
Building Cura was a rare opportunity to dive into the world of AI with total creative freedom. Without being held back by technical limitations, I could focus entirely on the "what if" pushing the boundaries of how an Agentic AI could truly simplify a user’s life.
This freedom allowed me to design a human-centric experience that turns complex, weeks-long manual grinds into seconds of strategic clarity. It wasn't just about building another tool; it was about imagining a partner that gives retail managers their time back for the work that actually matters.
Building Cura was a rare opportunity to dive into the world of AI with total creative freedom. Without being held back by technical limitations, I could focus entirely on the "what if" pushing the boundaries of how an Agentic AI could truly simplify a user’s life.
This freedom allowed me to design a human-centric experience that turns complex, weeks-long manual grinds into seconds of strategic clarity. It wasn't just about building another tool; it was about imagining a partner that gives retail managers their time back for the work that actually matters.
Contact
Let's start creating together
ujjwal.design23@gmail.com
Copied
2026 Ujjwal Kumar Singh I Portfolio I All rights reserved
Contact
Let's start creating together
ujjwal.design23@gmail.com
Copied
2026 Ujjwal Kumar Singh I Portfolio I All rights reserved
Contact
Let's start creating together
ujjwal.design23@gmail.com
Copied
2026 Ujjwal Kumar Singh I Portfolio I All rights reserved