Designing an AI-Powered Document Intelligence Workflow
Team
Chief Product Officer (Stakeholder & Product Direction)
Engineering Team
Product Design Team
Outcome
Shipped to production
Adopted within the product ecosystem
Integrated into multiple workflows and products
My Role
Role : Product designer
Duration : 2 months
Responsibilities : End-to-end process
Workflow Design | Information Architecture | UX Design | UI Design | Design Handoff
Insurance professionals frequently work with large, information-dense documents that can span hundreds or even thousands of pages. Reviewing these submissions manually is time-consuming, making it difficult to quickly identify the information needed to support risk assessment and decision-making.
To address this challenge, I designed an AI-powered document intelligence workflow that enables users to upload documents, generate tailored summaries, refine outputs through natural-language interactions, and reuse successful summarization strategies across future workflows.
Rather than treating AI as a one-time summarization tool, the experience was designed as an iterative workspace where users could continuously refine outputs, save reusable templates, and manage historical summaries in a single environment.
The solution was successfully shipped and later adopted across multiple products, extending AI-assisted document analysis into existing enterprise workflows.
(Note : Due to confidentiality and intellectual property restrictions, certain product details, visuals, workflows, and business information have been modified or generalized. This case study focuses on the design process, decision-making, and problem-solving approach rather than proprietary implementation details.)
Outcome
Shipped to production
Adopted within the product ecosystem
Integrated into multiple workflows and products
Team
Chief Product Officer (Stakeholder & Product Direction)
Engineering Team
Product Design Team
My Role
Role : Product designer
Duration : 2 months
Responsibilities : End-to-end process
Workflow Design | Information Architecture | UX Design | UI Design | Design Handoff
Insurance professionals frequently work with large, information-dense documents that can span hundreds or even thousands of pages. Reviewing these submissions manually is time-consuming, making it difficult to quickly identify the information needed to support risk assessment and decision-making.
To address this challenge, I designed an AI-powered document intelligence workflow that enables users to upload documents, generate tailored summaries, refine outputs through natural-language interactions, and reuse successful summarization strategies across future workflows.
Rather than treating AI as a one-time summarization tool, the experience was designed as an iterative workspace where users could continuously refine outputs, save reusable templates, and manage historical summaries in a single environment.
The solution was successfully shipped and later adopted across multiple products, extending AI-assisted document analysis into existing enterprise workflows.
(Note : Due to confidentiality and intellectual property restrictions, certain product details, visuals, workflows, and business information have been modified or generalized. This case study focuses on the design process, decision-making, and problem-solving approach rather than proprietary implementation details.)
Designing an AI-Powered Document Intelligence Workflow
Insurance professionals regularly receive highly detailed submissions, reports, and supporting documents that can span hundreds or thousands of pages.
The challenge wasn't simply generating a shorter version of a document. Users needed a way to quickly understand large amounts of information, refine summaries based on changing requirements, and reuse successful summarization strategies across future workflows.
At the same time, direct access to end users was limited. Product requirements and domain insights were gathered through internal product leadership and client-facing stakeholders.
This created three key challenges:
The Challenge
The Problem
When exploring how AI could support document-heavy workflows, several directions were possible. The solution could have focused on search, extraction, risk analysis, or conversational assistance.
Rather than starting with advanced capabilities, I focused on where users encountered friction first.
Before users can search for specific information, analyze risks, or ask meaningful questions, they need to understand what a document contains.
This led me to identify summarization as the highest-leverage point in the workflow. By helping users quickly build context, summarization could reduce manual review effort while creating a foundation for future AI capabilities.
Instead of treating summarization as a standalone feature, I saw an opportunity to design a workspace that supported understanding, refinement, and reuse throughout the document review process.
The Opportunity
This principle shaped the interface into three dedicated workspaces, each supporting a different stage of the summarization process.
TRANSLATING THE DECISION INTO THE INTERFACE
Instead of treating summaries as chat messages, I separated instructions from outputs.
The chat became a workspace for refinement while the summary existed as a living document that continuously updated.
This decision fundamentally changed the structure of the interface. Rather than displaying summaries inside a conversation, I designed two dedicated workspaces with distinct responsibilities.
DECISION
Multiple long summaries clutter the conversation.
Hard to locate the latest version.
Difficult to compare and track changes.
Excessive scrolling breaks focus.
EXPLORATION
Traditional chat interfaces work well for conversations but become difficult to manage when long content is repeatedly refined.
Every new summary version adds another large block of text, making it harder to locate the latest output.
DESIGN CHALLENGE
Design Decision #1 - Separating Instructions from Output
Users wouldn't generate a summary once and move on.
They would continuously refine, regenerate, and reshape the output until it matched their needs.
One insight quickly became clear
Key Insight
❌ Difficult to navigate
❌ Excessive scrolling
❌ Multiple summary versions
The interface shown throughout this case study has been intentionally redesigned and sanitized to protect confidential company information. Visuals are representative of the design decisions and user experience, not the production interface.
Workspace : Keeps previous sessions and reusable work
accessible without interrupting the current task.
Live Summary Area : Displays only the latest version, preventing users
from scrolling through multiple generations.
AI Assistant : Used exclusively for instructions and iterative
refinement.
While summarization solved the initial problem of understanding large documents, generating a summary alone wasn't enough. Through stakeholder discussions, it became clear that users would need to continuously refine outputs, revisit previous work, and apply successful summarization approaches across multiple documents.
This shifted the challenge from designing a summarization feature to designing an end-to-end workflow.
Exploring the workflow
Design Decision #2 - Reducing Prompt Friction Through Direct Manipulation
DESIGN CHALLENGE
EXPLORATION
DECISION
TRANSLATING THE DECISION INTO THE INTERFACE
Users first had to locate the correct paragraph, remember its position, and describe exactly what needed to change. As summaries became longer, writing prompts became slower.
Instead of asking users to describe where content should change, I introduced direct interactions that allowed them to highlight any part of the summary and immediately perform contextual actions such as Improve Wording, Deepen Explanation, or Remove Text.
The Turning Point
“If users wanted to rephrase two lines from paragraph 11, why should they have to remember the paragraph number and type a long instruction?”
Prompt based editing
❌ High cognitive load
Direct Manipulation
Lower Cognitive Load
Users interact directly with the generated summary instead of describing where edits should occur.
Design Decision #3 - From Manual Templates to Smart Template Extraction
THE TURNING POINT
During stakeholder discussions, a recurring question emerged:
“Underwriters often create similar summaries. How can we make those workflows reusable without asking them to recreate everything from scratch?”
DESIGN CHALLENGE
Underwriters often follow similar summarization strategies across multiple documents. While reusable templates seemed like the obvious solution, manually creating them introduced another task into an already document-heavy workflow.
DECISION
Instead of asking users to manually define reusable templates, I designed a workflow where AI automatically extracted summarization parameters from the conversation. Users could then review, modify, remove, or add parameters before saving the template keeping them in control while reducing repetitive work..
DESIGN OUTCOME
Rather than asking users to recreate successful prompting strategies, the system transforms one successful AI interaction into a reusable workflow.
This shifts templates from being manually authored artifacts to AI-assisted assets that users can review, customize, and reuse with confidence.
EXPLORATION
TRANSLATING THE DECISION INTO THE INTERFACE
Manual Template Creation
Users manually create a reusable template.
File Template
Gives users complete control
Gives users complete control
Requires users to know exactly what to write
Hard to maintain and keep consistent
Template Name
Instructions
• Focus on financial risks
• Ignore legal clauses
• Maximum 300 words
• Mention policy limits
Executive Summary
Save Template
Decide parameters/
fileds
Picking preset fields
Save
I have to manually create everything?
Reuse Conversation History
Reuse previously generated summaries and conversations from the workspace.
Conversation List
Executive Summary
Quarterly Report
Research Notes
Client Overview
Risk Assessment
Policy Review
Claims Analysis
Can you summarize this report focusing on financial risks?
Sure. Here is a summary...
Go a little deeper into
the technical findings.
Which conversation had the summary style I need?
Already available in the workspace
Contains lots of unrelated conversation
Users still need to find the useful instructions
Difficult to reuse consistently
Smart Template Extraction
AI extracts the summarization strategy from the conversation and converts it into a reusable template.
Extracted Parameters
Summary Type
Word Limit
Focus Areas
Exclude / Skip
Tone / Style
Professional Summary
Maximum 700 words
• dbbdvsdvhsbd
• jsdjsds
• jjjasgasgagsga
Legal Boilerplate
Professional & Concise
AI Conversation
The extracted parameters look good.
Captures the actual summarization strategy
Structured and reusable across documents
Users can review and edit extracted fields
Users edit extracted values
Remove unnecessary parameters
Add custom parameters
Reflection
This project changed the way I think about designing AI products.
One of the biggest constraints was that I never had direct access to end users.
Instead, product direction and feedback came through discussions with the manager and lead, who regularly worked with clients and demonstrated the product. That meant I had to ask better questions, understand the reasoning behind requirements, and continuously validate design decisions with stakeholders and engineering throughout the process.
It also reinforced that designing AI experiences goes far beyond generating an output. The real challenge was designing everything around the AI, how users refine results, reuse successful workflows, stay in control of AI-generated content, and integrate those interactions into their existing way of working.
If I could continue evolving this product, I would love to observe real users interacting with it. Seeing where they hesitate, what they trust, and how they naturally adapt the workflow would help uncover opportunities that stakeholder feedback alone can't fully reveal.
This project taught me that good product design isn't about finding the perfect solution on the first attempt, it's about continuously asking better questions, exploring alternatives, and refining ideas until they genuinely make people's work easier.
Impact
The workflow was successfully shipped to production and later integrated into multiple products across the platform, making AI-assisted summation available within existing enterprise workflows.
The design introduced reusable summation templates, direct summary editing, and a dedicated workspace for managing summaries, helping transform summation from a one-time AI interaction into a reusable workflow that fits naturally into users' day-to-day processes.
While I don't have access to usage metrics, seeing the workflow adopted across multiple products demonstrated that the design solved a broader need beyond its initial use case and established a foundation for future AI capabilities within the platform.
“This project reminded me that great product design isn't about having every answer from the start. It's about staying curious, asking better questions, embracing constraints, and continuously refining ideas alongside the people building the product. As an early-career designer, that mindset has become one of the most valuable things I've taken away from this experience.”
✨
~Farha
The Problem
The Challenge
Insurance professionals regularly receive highly detailed submissions, reports, and supporting documents that can span hundreds or thousands of pages.
The challenge wasn't simply generating a shorter version of a document. Users needed a way to quickly understand large amounts of information, refine summaries based on changing requirements, and reuse successful summarization strategies across future workflows.
At the same time, direct access to end users was limited. Product requirements and domain insights were gathered through internal product leadership and client-facing stakeholders.
This created three key challenges:
The Opportunity
When exploring how AI could support document-heavy workflows, several directions were possible. The solution could have focused on search, extraction, risk analysis, or conversational assistance.
Rather than starting with advanced capabilities, I focused on where users encountered friction first.
Before users can search for specific information, analyze risks, or ask meaningful questions, they need to understand what a document contains.
This led me to identify summarization as the highest-leverage point in the workflow. By helping users quickly build context, summarization could reduce manual review effort while creating a foundation for future AI capabilities.
Instead of treating summarization as a standalone feature, I saw an opportunity to design a workspace that supported understanding, refinement, and reuse throughout the document review process.
Exploring the workflow
While summarization solved the initial problem of understanding large documents, generating a summary alone wasn't enough. Through stakeholder discussions, it became clear that users would need to continuously refine outputs, revisit previous work, and apply successful summarization approaches across multiple documents.
This shifted the challenge from designing a summarization feature to designing an end-to-end workflow.
Key Insight
One insight quickly became clear
Users wouldn't generate a summary once and move on.
They would continuously refine, regenerate, and reshape the output until it matched their needs.
Design Decision #1 - Separating Instructions from Output
DESIGN CHALLENGE
Traditional chat interfaces work well for conversations but become difficult to manage when long content is repeatedly refined.
Every new summary version adds another large block of text, making it harder to locate the latest output.
DECISION
Instead of treating summaries as chat messages, I separated instructions from outputs.
The chat became a workspace for refinement while the summary existed as a living document that continuously updated.
This decision fundamentally changed the structure of the interface. Rather than displaying summaries inside a conversation, I designed two dedicated workspaces with distinct responsibilities.
TRANSLATING THE DECISION INTO THE INTERFACE
This principle shaped the interface into three dedicated workspaces, each supporting a different stage of the summarization process.
EXPLORATION
Multiple long summaries clutter the conversation.
Hard to locate the latest version.
Excessive scrolling breaks focus.
Difficult to compare and track changes.
❌ Difficult to navigate
❌ Excessive scrolling
❌ Multiple summary versions
The interface shown throughout this case study has been intentionally redesigned and sanitized to protect confidential company information. Visuals are representative of the design decisions and user experience, not the production interface.
Workspace : Keeps previous sessions and reusable work
accessible without interrupting the current task.
Live Summary Area : Displays only the latest version, preventing users
from scrolling through multiple generations.
AI Assistant : Used exclusively for instructions and iterative
refinement.
Design Decision #2 - Reducing Prompt Friction Through Direct Manipulation
TRANSLATING THE DECISION INTO THE INTERFACE
DESIGN CHALLENGE
Users first had to locate the correct paragraph, remember its position, and describe exactly what needed to change. As summaries became longer, writing prompts became slower.
DECISION
Instead of asking users to describe where content should change, I introduced direct interactions that allowed them to highlight any part of the summary and immediately perform contextual actions such as Improve Wording, Deepen Explanation, or Remove Text.
THE TURNING POINT
“If users wanted to rephrase two lines from paragraph 11, why should they have to remember the paragraph number and type a long instruction?”
EXPLORATION
Prompt based editing
❌ High cognitive load
Direct Manipulation
Lower Cognitive Load
Users interact directly with the generated summary instead of describing where edits should occur.
Design Decision #3 - From Manual Templates to Smart Template Extraction
THE TURNING POINT
During stakeholder discussions, a recurring question emerged:
“Underwriters often create similar summaries. How can we make those workflows reusable without asking them to recreate everything from scratch?”
DESIGN CHALLENGE
Underwriters often follow similar summarization strategies across multiple documents. While reusable templates seemed like the obvious solution, manually creating them introduced another task into an already document-heavy workflow.
DECISION
Instead of asking users to manually define reusable templates, I designed a workflow where AI automatically extracted summarization parameters from the conversation. Users could then review, modify, remove, or add parameters before saving the template keeping them in control while reducing repetitive work..
DESIGN OUTCOME
Rather than asking users to recreate successful prompting strategies, the system transforms one successful AI interaction into a reusable workflow.
This shifts templates from being manually authored artifacts to AI-assisted assets that users can review, customize, and reuse with confidence.
TRANSLATING THE DECISION INTO THE INTERFACE
EXPLORATION
Manual Template Creation
Users manually create a reusable template.
File Template
Gives users complete control
Gives users complete control
Requires users to know exactly what to write
Hard to maintain and keep consistent
Template Name
Instructions
• Focus on financial risks
• Ignore legal clauses
• Maximum 300 words
• Mention policy limits
Executive Summary
Save Template
Decide parameters/fileds
Picking preset fields
Save
I have to manually create everything?
Reuse Conversation History
Reuse previously generated summaries and conversations from the workspace.
Conversation List
Executive Summary
Quarterly Report
Research Notes
Client Overview
Risk Assessment
Policy Review
Claims Analysis
Can you summarize this report focusing on financial risks?
Sure. Here is a summary...
Go a little deeper into
the technical findings.
Which conversation had the summary style I need?
Already available in the workspace
Contains lots of unrelated conversation
Users still need to find the useful instructions
Difficult to reuse consistently
Smart Template Extraction
AI extracts the summarization strategy from the conversation and converts it into a reusable template.
Extracted Parameters
Summary Type
Word Limit
Focus Areas
Exclude / Skip
Tone / Style
Professional Summary
Maximum 700 words
• dbbdvsdvhsbd
• jsdjsds
• jjjasgasgagsga
Legal Boilerplate
Professional & Concise
AI Conversation
The extracted parameters look good.
Captures the actual summarization strategy
Structured and reusable across documents
Users can review and edit extracted fields
Users edit extracted values
Add custom parameters
Remove unnecessary parameters
Reflection
This project changed the way I think about designing AI products.
One of the biggest constraints was that I never had direct access to end users.
Instead, product direction and feedback came through discussions with the manager and lead, who regularly worked with clients and demonstrated the product. That meant I had to ask better questions, understand the reasoning behind requirements, and continuously validate design decisions with stakeholders and engineering throughout the process.
It also reinforced that designing AI experiences goes far beyond generating an output. The real challenge was designing everything around the AI, how users refine results, reuse successful workflows, stay in control of AI-generated content, and integrate those interactions into their existing way of working.
If I could continue evolving this product, I would love to observe real users interacting with it. Seeing where they hesitate, what they trust, and how they naturally adapt the workflow would help uncover opportunities that stakeholder feedback alone can't fully reveal.
This project taught me that good product design isn't about finding the perfect solution on the first attempt, it's about continuously asking better questions, exploring alternatives, and refining ideas until they genuinely make people's work easier.
Impact
The workflow was successfully shipped to production and later integrated into multiple products across the platform, making AI-assisted summation available within existing enterprise workflows.
The design introduced reusable summation templates, direct summary editing, and a dedicated workspace for managing summaries, helping transform summation from a one-time AI interaction into a reusable workflow that fits naturally into users' day-to-day processes.
While I don't have access to usage metrics, seeing the workflow adopted across multiple products demonstrated that the design solved a broader need beyond its initial use case and established a foundation for future AI capabilities within the platform.
✨
“This project reminded me that great product design isn't about having every answer from the start. It's about staying curious, asking better questions, embracing constraints, and continuously refining ideas alongside the people building the product. As an early-career designer, that mindset has become one of the most valuable things I've taken away from this experience.”
~Farha
Reflection
This project changed the way I think about designing AI products.
One of the biggest constraints was that I never had direct access to end users.
Instead, product direction and feedback came through discussions with the manager and lead, who regularly worked with clients and demonstrated the product. That meant I had to ask better questions, understand the reasoning behind requirements, and continuously validate design decisions with stakeholders and engineering throughout the process.
It also reinforced that designing AI experiences goes far beyond generating an output. The real challenge was designing everything around the AI, how users refine results, reuse successful workflows, stay in control of AI-generated content, and integrate those interactions into their existing way of working.
If I could continue evolving this product, I would love to observe real users interacting with it. Seeing where they hesitate, what they trust, and how they naturally adapt the workflow would help uncover opportunities that stakeholder feedback alone can't fully reveal.
This project taught me that good product design isn't about finding the perfect solution on the first attempt, it's about continuously asking better questions, exploring alternatives, and refining ideas until they genuinely make people's work easier.
Impact
The workflow was successfully shipped to production and later integrated into multiple products across the platform, making AI-assisted summation available within existing enterprise workflows.
The design introduced reusable summation templates, direct summary editing, and a dedicated workspace for managing summaries, helping transform summation from a one-time AI interaction into a reusable workflow that fits naturally into users' day-to-day processes.
While I don't have access to usage metrics, seeing the workflow adopted across multiple products demonstrated that the design solved a broader need beyond its initial use case and established a foundation for future AI capabilities within the platform.
✨
“This project reminded me that great product design isn't about having every answer from the start. It's about staying curious, asking better questions, embracing constraints, and continuously refining ideas alongside the people building the product. As an early-career designer, that mindset has become one of the most valuable things I've taken away from this experience.”
~Farha
My Role
Role : Product designer
Duration : 2 months
Responsibilities : End-to-end process
Workflow Design | Information Architecture | UX Design | UI Design | Design Handoff
Team
Chief Product Officer (Stakeholder & Product Direction)
Engineering Team
Product Design Team
Shipped to production
Adopted within the product ecosystem
Integrated into multiple workflows and products
Outcome
Insurance professionals frequently work with large, information-dense documents that can span hundreds or even thousands of pages. Reviewing these submissions manually is time-consuming, making it difficult to quickly identify the information needed to support risk assessment and decision-making.
To address this challenge, I designed an AI-powered document intelligence workflow that enables users to upload documents, generate tailored summaries, refine outputs through natural-language interactions, and reuse successful summarization strategies across future workflows.
Rather than treating AI as a one-time summarization tool, the experience was designed as an iterative workspace where users could continuously refine outputs, save reusable templates, and manage historical summaries in a single environment.
The solution was successfully shipped and later adopted across multiple products, extending AI-assisted document analysis into existing enterprise workflows.
Designing an AI-Powered Document Intelligence Workflow
(Note : Due to confidentiality and intellectual property restrictions, certain product details, visuals, workflows, and business information have been modified or generalized. This case study focuses on the design process, decision-making, and problem-solving approach rather than proprietary implementation details.)
The Problem
The Challenge
Insurance professionals regularly receive highly detailed submissions, reports, and supporting documents that can span hundreds or thousands of pages.
The challenge wasn't simply generating a shorter version of a document. Users needed a way to quickly understand large amounts of information, refine summaries based on changing requirements, and reuse successful summarization strategies across future workflows.
At the same time, direct access to end users was limited. Product requirements and domain insights were gathered through internal product leadership and client-facing stakeholders.
This created three key challenges:
The Opportunity
When exploring how AI could support document-heavy workflows, several directions were possible. The solution could have focused on search, extraction, risk analysis, or conversational assistance.
Rather than starting with advanced capabilities, I focused on where users encountered friction first.
Before users can search for specific information, analyze risks, or ask meaningful questions, they need to understand what a document contains.
This led me to identify summarization as the highest-leverage point in the workflow. By helping users quickly build context, summarization could reduce manual review effort while creating a foundation for future AI capabilities.
Instead of treating summarization as a standalone feature, I saw an opportunity to design a workspace that supported understanding, refinement, and reuse throughout the document review process.
Exploring the workflow
While summarization solved the initial problem of understanding large documents, generating a summary alone wasn't enough. Through stakeholder discussions, it became clear that users would need to continuously refine outputs, revisit previous work, and apply successful summarization approaches across multiple documents.
This shifted the challenge from designing a summarization feature to designing an end-to-end workflow.
DECISION
Instead of treating summaries as chat messages, I separated instructions from outputs.
The chat became a workspace for refinement while the summary existed as a living document that continuously updated.
This decision fundamentally changed the structure of the interface. Rather than displaying summaries inside a conversation, I designed two dedicated workspaces with distinct responsibilities.
EXPLORATION
Multiple long summaries clutter the conversation.
Hard to locate the latest version.
Excessive scrolling breaks focus.
Difficult to compare and track changes.
❌ Difficult to navigate
❌ Excessive scrolling
❌ Multiple summary versions
Key Insight
One insight quickly became clear
Users wouldn't generate a summary once and move on.
They would continuously refine, regenerate, and reshape the output until it matched their needs.
Design Decision #1 - Separating Instructions from Output
DESIGN CHALLENGE
Traditional chat interfaces work well for conversations but become difficult to manage when long content is repeatedly refined.
Every new summary version adds another large block of text, making it harder to locate the latest output.
TRANSLATING THE DECISION INTO THE INTERFACE
This principle shaped the interface into three dedicated workspaces, each supporting a different stage of the summarization process.
The interface shown throughout this case study has been intentionally redesigned and sanitized to protect confidential company information. Visuals are representative of the design decisions and user experience, not the production interface.
1
2
3
1
Workspace : Keeps previous sessions and reusable work accessible without interrupting the current task.
2
Live Summary Area : Displays only the latest version, preventing users
from scrolling through multiple generations.
3
AI Assistant : Used exclusively for instructions and iterative refinement.
Design Decision #2 - Reducing Prompt Friction Through Direct Manipulation
DESIGN CHALLENGE
Users first had to locate the correct paragraph, remember its position, and describe exactly what needed to change. As summaries became longer, writing prompts became slower.
DECISION
Instead of asking users to describe where content should change, I introduced direct interactions that allowed them to highlight any part of the summary and immediately perform contextual actions such as Improve Wording, Deepen Explanation, or Remove Text.
THE TURNING POINT
“If users wanted to rephrase two lines from paragraph 11, why should they have to remember the paragraph number and type a long instruction?”
TRANSLATING THE DECISION INTO THE INTERFACE
EXPLORATION
Prompt based editing
❌ High cognitive load
Direct Manipulation
Lower Cognitive Load
Users interact directly with the generated summary instead of describing where edits should occur.
Design Decision #3 - From Manual Templates to Smart Template Extraction
THE TURNING POINT
During stakeholder discussions, a recurring question emerged:
“Underwriters often create similar summaries. How can we make those workflows reusable without asking them to recreate everything from scratch?”
DESIGN CHALLENGE
Underwriters often follow similar summarization strategies across multiple documents. While reusable templates seemed like the obvious solution, manually creating them introduced another task into an already document-heavy workflow.
DECISION
Instead of asking users to manually define reusable templates, I designed a workflow where AI automatically extracted summarization parameters from the conversation. Users could then review, modify, remove, or add parameters before saving the template keeping them in control while reducing repetitive work..
DESIGN OUTCOME
Rather than asking users to recreate successful prompting strategies, the system transforms one successful AI interaction into a reusable workflow.
This shifts templates from being manually authored artifacts to AI-assisted assets that users can review, customize, and reuse with confidence.
TRANSLATING THE DECISION INTO THE INTERFACE
EXPLORATION
Manual Template Creation
Users manually create a reusable template.
File Template
Gives users complete control
Gives users complete control
Requires users to know exactly what to write
Hard to maintain and keep consistent
Template Name
Instructions
• Focus on financial risks
• Ignore legal clauses
• Maximum 300 words
• Mention policy limits
Executive Summary
Save Template
Decide parameters/fileds
Picking preset fields
Save
I have to manually create everything?
Reuse Conversation History
Reuse previously generated summaries and conversations from the workspace.
Conversation List
Executive Summary
Quarterly Report
Research Notes
Client Overview
Risk Assessment
Policy Review
Claims Analysis
Can you summarize this report focusing on financial risks?
Sure. Here is a summary...
Go a little deeper into
the technical findings.
Which conversation had the summary style I need?
Already available in the workspace
Contains lots of unrelated conversation
Users still need to find the useful instructions
Difficult to reuse consistently
Smart Template Extraction
AI extracts the summarization strategy from the conversation and converts it into a reusable template.
Extracted Parameters
Summary Type
Word Limit
Focus Areas
Exclude / Skip
Tone / Style
Professional Summary
Maximum 700 words
• dbbdvsdvhsbd
• jsdjsds
• jjjasgasgagsga
Legal Boilerplate
Professional & Concise
AI Conversation
The extracted parameters look good.
Captures the actual summarization strategy
Structured and reusable across documents
Users can review and edit extracted fields
Users can edit extracted values
Add custom parameters
Remove unnecessary parameters