Anusha Maradani

Anusha maradani

Home

Home

Projects

Projects

View Resume

View Resume

Sigmoid

Becton & Dickinson (BD)

AIden: AI Content Tool for BD Marketing

AIden: AI Content Tool for BD Marketing

Enterprise UX

Product Design

Gen AI Tool

Duration

3 Months


My Role

Product Strategy, UX & UI,

Design System Development

Context

To design a scalable, user-friendly generative AI tool that enables BD’s global marketing teams to create medically accurate, brand-consistent content — faster and with fewer resources.

Project Brief

Project Brief

AIden is a generative AI tool designed for internal use by BD’s global marketing teams. It’s part of BD’s larger digital transformation effort to improve content creation efficiency, brand consistency, and cost-effectiveness, while ensuring scientific and regulatory accuracy in all client-facing messaging.

When BD’s global marketing teams started experimenting with Aiden, their first version — a simple Microsoft Power Apps tool — could generate basic email content, but not at the quality or speed the teams needed.

Built on Streamlit, it followed a linear flow with limited control over tone or structure, often producing off-brand outputs and slowing down marketers instead of helping them.

My role was to reimagine this tool into a scalable, intelligent system — one that aligned with BD’s brand voice, regulatory requirements, and daily marketing workflows.

AIden is a generative AI tool designed for internal use by BD’s global marketing teams. It’s part of BD’s larger digital transformation effort to improve content creation efficiency, brand consistency, and cost-effectiveness, while ensuring scientific and regulatory accuracy in all client-facing messaging.

When BD’s global marketing teams started experimenting with Aiden, their first version — a simple Microsoft Power Apps tool — could generate basic email content, but not at the quality or speed the teams needed.

Built on Streamlit, it followed a linear flow with limited control over tone or structure, often producing off-brand outputs and slowing down marketers instead of helping them.

My role was to reimagine this tool into a scalable, intelligent system — one that aligned with BD’s brand voice, regulatory requirements, and daily marketing workflows.

AIden is a generative AI tool designed for internal use by BD’s global marketing teams. It’s part of BD’s larger digital transformation effort to improve content creation efficiency, brand consistency, and cost-effectiveness, while ensuring scientific and regulatory accuracy in all client-facing messaging.

When BD’s global marketing teams started experimenting with Aiden, their first version — a simple Microsoft Power Apps tool — could generate basic email content, but not at the quality or speed the teams needed.

Built on Streamlit, it followed a linear flow with limited control over tone or structure, often producing off-brand outputs and slowing down marketers instead of helping them.

My role was to reimagine this tool into a scalable, intelligent system — one that aligned with BD’s brand voice, regulatory requirements, and daily marketing workflows.

Start Generating

GENERATE

Start Generating

GENERATE

Understanding the people

process and problems

Understanding the

people process and

problems

Understanding the people process and problems

I began by studying the existing tool and mapping how different roles across BD’s marketing department — from campaign managers to content specialists — collaborated to produce client-facing emails and creative assets. Through interviews and workflow mapping, I identified key pain points and ranked them by impact, persistence and frequency to understand where inefficiencies occurred most often.


The following swimlane diagram visualizes the typical content creation workflow — from product release to final approval — and highlights where AIden could drive the most impact. I identified three key user types and mapped out their unique pain points and expectations.

I began by studying the existing tool and mapping how different roles across BD’s marketing department — from campaign managers to content specialists — collaborated to produce client-facing emails and creative assets. Through interviews and workflow mapping, I identified key pain points and ranked them by impact, persistence and frequency to understand where inefficiencies occurred most often.


The following swimlane diagram visualizes the typical content creation workflow — from product release to final approval — and highlights where AIden could drive the most impact. I identified three key user types and mapped out their unique pain points and expectations.

I began by studying the existing tool and mapping how different roles across BD’s marketing department — from campaign managers to content specialists — collaborated to produce client-facing emails and creative assets. Through interviews and workflow mapping, I identified key pain points and ranked them by impact, persistence and frequency to understand where inefficiencies occurred most often.


The following swimlane diagram visualizes the typical content creation workflow — from product release to final approval — and highlights where AIden could drive the most impact. I identified three key user types and mapped out their unique pain points and expectations.

1

“As an Account Manager, I frequently need to write product launch emails for different client segments. I want to quickly generate accurate, on-brand email content that I can tweak slightly — without starting from scratch every time — so I can respond to client needs faster while staying compliant”

2

As a Campaign Owner, I manage multiple campaigns across different geographies. I need a tool that allows me to create creatives with consistent messaging and visuals, using approved templates and prompts, so I can maintain alignment across teams while reducing time spent on revisions

3

As a Content Specialist, I create marketing copy and visuals for different BD campaigns. I need a tool that helps me structure my prompts clearly, maintain tone consistency, and generate brand-safe outputs without feeling restricted by rigid templates.

Pain Points

Rewriting similar prompts from scratch every time

Outputs vary in tone and accuracy depending on how the prompt is written

No way to save or reuse effective prompts

Pain Points

Inability to track or share high-performing prompts

Repetition of content creation across teams

Unclear workflows — users unsure how to structure prompt + creative combinations

Pain Points

 Users writing free-form prompts, resulting in inconsistent creative messages

Current tool is too rigid and unintuitive

Lack of guardrails to enforce brand-safe outputs

🤖 Generate & Iterate

drafts

🤖 Source trace

Product Trigger

Channel Planning

Context Setup

Content Drafting

Review & Approval

Execution

Product Team

Marketing

Brand

Compliance

🤖 Output backed

by specific source

✍️ Provide

Product Brief

💻 Support

with positioning

🛠️Share relevant

information

🛠️ Finding relevant

information

✍️ Select

channels +tone

✍️ Identify

buyer persona

✍️ Align on brand

tone

💻 provide

visual guidance 

🛠️  Provide

disclaimers

🛠️ Review

technical accuracy

✅ Approve product

message   

🛠️   Review for

consistency 

✅ Approve

messaging style 

✅  Final approval

from lead 

✅  Sign-off for

compliance

🤖 upload & recall

context informaion


🤖 Set tone through prompt


💻 Create copy or

content

✍️ Manual/Strategic

💻 Tool-based (non-AI)

🛠️ Hybrid/manual workflow

🤖 Aiden

Prioritizing Pain Points

Prioritizing Pain Points

The priority matrix visualizes key user pain points based on their impact, frequency, and persistence. It helped us identify which issues to address first in the design of AIden for maximum user and business value.

The priority matrix visualizes key user pain points based on their impact, frequency, and persistence. It helped us identify which issues to address first in the design of AIden for maximum user and business value.

The priority matrix visualizes key user pain points based on their impact, frequency, and persistence. It helped us identify which issues to address first in the design of AIden for maximum user and business value.

Low Impact | Very Frequent

Low Impact | Rarely Occurs

High Impact | Rarely Occurs

High Impact | Very Frequent

Impact

Frequency

Inconsistent output

Complex or Unclear

Workflows

No Way to Save or Reuse Prompts

 No dark mode or visual theme options

No ability to edit or replace

uploaded context documents mid-flow

No Visibility into AI Response Source References

 No multi-user collaboration / co-editing features

Rigid, Limiting Experience

No Guardrails for Brand-Safe Output

No History of Previously Uploaded Context Documents

Frequency Scale

1

Rarely

2

Infrequent

3

Sometimes

4

Often

5

Very Frequent

Impact Scale

1

Negligible

2

Minor

3

Moderate

4

High

5

Critical

Persistence Scale

Short-lived

Intermittent

Recurrent

Ongoing

Persistant

Foundation of Aiden

Foundation of Aiden

My research into BD’s people, processes, and pain points shaped the foundation of Aiden.

From this understanding, three core product goals emerged—

My research into BD’s people, processes, and pain points shaped the foundation of Aiden.

From this understanding, three core product goals emerged—

My research into BD’s people, processes, and pain points shaped the foundation of Aiden. From this understanding, three core product goals emerged—

Prodcutivity

Prodcutivity

Prodcutivity

Faster ideas, multiple ideas & greater creative efficience

Scalability

Flexible system that can grow beyond email creation


Gaurdrails

Aligned with BD’s brand and compliance standards through structured, guided creation.

Persistant | High Impact | Rarely Occurs

Structuring Inputs for

Brand-Safe Output

Structuring Inputs for

Brand-Safe Output

Earlier, user prompts were typed directly into a single chat field, resulting in long, unstructured messages and inconsistent AI responses.

To address this, I designed a two-way input system that combines structure for brand safety and flexibility for iterative refinement.

1.Structured Prompt: The Anchor
 A structured input form defines clear fields for Task, Email Structure, and Temperature. This anchors every generation to a brand-safe framework — preventing subjective tone or off-brief phrasing.

2.Conversational Refinement: The Improvisation Layer.

Once the structured prompt generates a draft, users can continue in the chat interface — refining tone, phrasing, or structure while the base prompt stays active.

This preserves creative flow but within defined boundaries.

🎯 A clear input structure helped standardize output — reducing tone drift, improving brand alignment, and solving for inconsistent content generation

Persistant | High Impact | Rarely Occurs

Improving visibility and

traceability of the output

Improving visibility

and traceability of

the output

Improving visibility and

traceability of the output

Earlier, users had no way to review or trace how AI-generated content was produced — once an output was generated, it disappeared from view.To address this, I introduced a Chat History and Traceability system — enabling users to view previous prompts and outputs, reuse them, or build upon them directly within the chat interface.

1.Chat History: The Recall Layer
 Users can revisit any past version, compare tone variations, and reuse effective outputs and prompts

2.Traceablity Guardrails

Each output is linked back to its Structured Prompt (Task, Email Structure, Temperature), Context documents and the source reference from the uploaded context documents

🎯 Visibility into prompt history and document references — strengthened productivity and reinforced guardrails

Refined User flow for Aiden

Refined User flow

for Aiden

A redesigned experience that merges structured inputs with conversational refinement — enabling faster creation, reusable prompts, and full visibility into every output

🎬 Start

🏁 Final Output

🚀 Onboarding

📚 Prompt Library

🕓 Chat history

📥 Provide Inputs

📄 Upload

context documents

⚡️ Structured Prompt

💻💬️ Conversational

refinement

Generate Output

Set Task

Set Email Structure

Set Tone Temperature

📝  Select Content

and Platform

🌐 Platform

Email

LinkedIn

Display Ads

Website

 🗂️ Content

Text

Creative

🚦 Is the output

satisfactory

YES

NO

Load/create a prompt template

View or reuse past interactions

Impact

Results of the redesign

Improving visibility

and traceability of

the output

🕒  Reduction in campaign setup Time

structured templates and approval workflows reduced the average campaign creation time from 3 days to 1 day, improving efficiency by nearly 65%.

📈 Improved quality consistency

by standardizing prompts and introducing content guardrails, brand compliance errors dropped by 30%, and teams produced more on-brand marketing assets.

💼 Increased marketing productivity

Centralized access to templates and ai-generated suggestions helped marketing teams create 2x more campaigns per cycle while reducing dependency on third-party agencies and cutting related costs by approximately 25%.

🔒

While I’ve shared key parts of my design process, some visuals and final assets are under NDA and can’t be shown publicly.