Building the Loopio Copilot Agent
TBD
March 4, 2026
Building the Loopio Copilot Agent
TBD
March 4, 2026
CLIENT
Loopio
Role
Product Designer
CLIENT
Loopio
Role
Product Designer
CLIENT
Loopio
Role
Product Designer
Context
Sales and proposal teams rely heavily on Loopio’s content library to answer RFPs and customer questions. However, retrieving the right information often requires switching between tools such as Loopio, Microsoft Teams, and documents.
As AI assistants began becoming embedded into productivity tools like Microsoft Copilot, an opportunity emerged: instead of asking users to come to Loopio, what if Loopio’s knowledge could come directly to where they were already working?
The Loopio Copilot Agent was designed to bring trusted library content into the Microsoft 365 ecosystem, allowing users to retrieve vetted answers using natural language directly within their workflow.
The goal was simple but ambitious: reduce context switching and make trusted answers accessible in seconds during live sales conversations or proposal work.
Context
Sales and proposal teams rely heavily on Loopio’s content library to answer RFPs and customer questions. However, retrieving the right information often requires switching between tools such as Loopio, Microsoft Teams, and documents.
As AI assistants began becoming embedded into productivity tools like Microsoft Copilot, an opportunity emerged: instead of asking users to come to Loopio, what if Loopio’s knowledge could come directly to where they were already working?
The Loopio Copilot Agent was designed to bring trusted library content into the Microsoft 365 ecosystem, allowing users to retrieve vetted answers using natural language directly within their workflow.
The goal was simple but ambitious: reduce context switching and make trusted answers accessible in seconds during live sales conversations or proposal work.
Context
Sales and proposal teams rely heavily on Loopio’s content library to answer RFPs and customer questions. However, retrieving the right information often requires switching between tools such as Loopio, Microsoft Teams, and documents.
As AI assistants began becoming embedded into productivity tools like Microsoft Copilot, an opportunity emerged: instead of asking users to come to Loopio, what if Loopio’s knowledge could come directly to where they were already working?
The Loopio Copilot Agent was designed to bring trusted library content into the Microsoft 365 ecosystem, allowing users to retrieve vetted answers using natural language directly within their workflow.
The goal was simple but ambitious: reduce context switching and make trusted answers accessible in seconds during live sales conversations or proposal work.
Insights
Early discovery conversations with customers and internal teams revealed a consistent workflow challenge.
Proposal teams spent a significant amount of time searching for previously written answers. While Loopio’s library already stored these answers, users still had to open the platform, navigate to the right section, and manually search for content.
This friction became even more visible for sales teams working in real-time environments such as Teams chats or customer calls.
Two key insights shaped the design direction:
Context switching slows teams down
Users frequently left their active workflow to search Loopio for answers.
Trust in answers matters more than speed alone
Users needed confidence that the answers surfaced by AI were accurate and sourced from vetted library entries.
These insights suggested that simply adding an AI chatbot would not be enough. The experience had to be tightly integrated with Loopio’s existing knowledge system while maintaining transparency about where answers came from.
Insights
Early discovery conversations with customers and internal teams revealed a consistent workflow challenge.
Proposal teams spent a significant amount of time searching for previously written answers. While Loopio’s library already stored these answers, users still had to open the platform, navigate to the right section, and manually search for content.
This friction became even more visible for sales teams working in real-time environments such as Teams chats or customer calls.
Two key insights shaped the design direction:
Context switching slows teams down
Users frequently left their active workflow to search Loopio for answers.
Trust in answers matters more than speed alone
Users needed confidence that the answers surfaced by AI were accurate and sourced from vetted library entries.
These insights suggested that simply adding an AI chatbot would not be enough. The experience had to be tightly integrated with Loopio’s existing knowledge system while maintaining transparency about where answers came from.
Insights
Early discovery conversations with customers and internal teams revealed a consistent workflow challenge.
Proposal teams spent a significant amount of time searching for previously written answers. While Loopio’s library already stored these answers, users still had to open the platform, navigate to the right section, and manually search for content.
This friction became even more visible for sales teams working in real-time environments such as Teams chats or customer calls.
Two key insights shaped the design direction:
Context switching slows teams down
Users frequently left their active workflow to search Loopio for answers.
Trust in answers matters more than speed alone
Users needed confidence that the answers surfaced by AI were accurate and sourced from vetted library entries.
These insights suggested that simply adding an AI chatbot would not be enough. The experience had to be tightly integrated with Loopio’s existing knowledge system while maintaining transparency about where answers came from.
Strategy
The product vision centered on positioning the Copilot Agent as a knowledge layer inside Microsoft 365, rather than a separate interface.
Three principles guided the design direction:
Bring knowledge to the workflow
Users should be able to retrieve Loopio answers directly inside Teams or Copilot chat without opening the Loopio platform.Make AI responses trustworthy
Responses needed to clearly reference library entries so users could verify information quickly.Support multiple moments of use
The agent needed to work across several scenarios:Searching the Loopio library
Creating new library entries
Answering product help questions
To validate these ideas, the team planned a staged rollout beginning with internal testing followed by closed customer beta program and then to observe real-world usage better an open beta.
Strategy
The product vision centered on positioning the Copilot Agent as a knowledge layer inside Microsoft 365, rather than a separate interface.
Three principles guided the design direction:
Bring knowledge to the workflow
Users should be able to retrieve Loopio answers directly inside Teams or Copilot chat without opening the Loopio platform.Make AI responses trustworthy
Responses needed to clearly reference library entries so users could verify information quickly.Support multiple moments of use
The agent needed to work across several scenarios:Searching the Loopio library
Creating new library entries
Answering product help questions
To validate these ideas, the team planned a staged rollout beginning with internal testing followed by closed customer beta program and then to observe real-world usage better an open beta.
Strategy
The product vision centered on positioning the Copilot Agent as a knowledge layer inside Microsoft 365, rather than a separate interface.
Three principles guided the design direction:
Bring knowledge to the workflow
Users should be able to retrieve Loopio answers directly inside Teams or Copilot chat without opening the Loopio platform.Make AI responses trustworthy
Responses needed to clearly reference library entries so users could verify information quickly.Support multiple moments of use
The agent needed to work across several scenarios:Searching the Loopio library
Creating new library entries
Answering product help questions
To validate these ideas, the team planned a staged rollout beginning with internal testing followed by closed customer beta program and then to observe real-world usage better an open beta.
Execution
My role focused on shaping how customers would interact with the agent and ensuring we could learn from real-world usage early in the process.
This involved several areas of work.
Designing the interaction model
Because the agent operated inside Microsoft Copilot, the experience needed to work within the constraints of conversational interfaces.
Key design considerations included:
How users discover what they can ask the agent
How answers are structured and referenced
How AI responses link back to Loopio library entries
How new content could be created directly from conversations
The goal was to ensure the agent felt like a natural extension of users’ existing workflows.
Designing the research and validation process
To understand how the agent performed in real environments, I designed a structured research approach during the beta phase.
This included:
Creating the customer interview discussion guide
Defining evaluation criteria for search accuracy, usability, and trust
Running feedback conversations with beta participants
The discussion guide focused on understanding how users interacted with the agent’s three primary capabilities:
Searching the Loopio library
Creating library entries
Accessing help center knowledge
These conversations helped us identify both moments of success and areas where the experience still needed improvement.
Structuring the beta program
The beta program was designed to test the agent in real customer workflows, particularly among enterprise teams already using Microsoft 365 extensively.
Participants included:
Proposal managers testing content accuracy
Sales engineers retrieving answers during customer conversations
IT administrators evaluating deployment and permissions
We defined clear success metrics to evaluate adoption, performance, and usability, including response accuracy, retrieval speed, and active usage rates.
Execution
My role focused on shaping how customers would interact with the agent and ensuring we could learn from real-world usage early in the process.
This involved several areas of work.
Designing the interaction model
Because the agent operated inside Microsoft Copilot, the experience needed to work within the constraints of conversational interfaces.
Key design considerations included:
How users discover what they can ask the agent
How answers are structured and referenced
How AI responses link back to Loopio library entries
How new content could be created directly from conversations
The goal was to ensure the agent felt like a natural extension of users’ existing workflows.
Designing the research and validation process
To understand how the agent performed in real environments, I designed a structured research approach during the beta phase.
This included:
Creating the customer interview discussion guide
Defining evaluation criteria for search accuracy, usability, and trust
Running feedback conversations with beta participants
The discussion guide focused on understanding how users interacted with the agent’s three primary capabilities:
Searching the Loopio library
Creating library entries
Accessing help center knowledge
These conversations helped us identify both moments of success and areas where the experience still needed improvement.
Structuring the beta program
The beta program was designed to test the agent in real customer workflows, particularly among enterprise teams already using Microsoft 365 extensively.
Participants included:
Proposal managers testing content accuracy
Sales engineers retrieving answers during customer conversations
IT administrators evaluating deployment and permissions
We defined clear success metrics to evaluate adoption, performance, and usability, including response accuracy, retrieval speed, and active usage rates.
Execution
My role focused on shaping how customers would interact with the agent and ensuring we could learn from real-world usage early in the process.
This involved several areas of work.
Designing the interaction model
Because the agent operated inside Microsoft Copilot, the experience needed to work within the constraints of conversational interfaces.
Key design considerations included:
How users discover what they can ask the agent
How answers are structured and referenced
How AI responses link back to Loopio library entries
How new content could be created directly from conversations
The goal was to ensure the agent felt like a natural extension of users’ existing workflows.
Designing the research and validation process
To understand how the agent performed in real environments, I designed a structured research approach during the beta phase.
This included:
Creating the customer interview discussion guide
Defining evaluation criteria for search accuracy, usability, and trust
Running feedback conversations with beta participants
The discussion guide focused on understanding how users interacted with the agent’s three primary capabilities:
Searching the Loopio library
Creating library entries
Accessing help center knowledge
These conversations helped us identify both moments of success and areas where the experience still needed improvement.
Structuring the beta program
The beta program was designed to test the agent in real customer workflows, particularly among enterprise teams already using Microsoft 365 extensively.
Participants included:
Proposal managers testing content accuracy
Sales engineers retrieving answers during customer conversations
IT administrators evaluating deployment and permissions
We defined clear success metrics to evaluate adoption, performance, and usability, including response accuracy, retrieval speed, and active usage rates.
Impact
The beta program validated several core assumptions about the Copilot Agent concept.
First, integrating Loopio knowledge directly into Microsoft tools proved valuable for teams that frequently needed answers during live conversations.
Second, AI-powered search significantly reduced the time required to locate relevant answers compared to traditional manual searches.
Finally, customer feedback highlighted the importance of transparency and trust in AI responses, reinforcing the need for clear source references and predictable answer formatting.
These insights helped shape the next iterations of the Copilot Agent experience and informed how Loopio approaches AI-powered knowledge retrieval within existing workflows.
(Validated this by running a UMUX Survey & 1:1 feedback calls)
Impact
The beta program validated several core assumptions about the Copilot Agent concept.
First, integrating Loopio knowledge directly into Microsoft tools proved valuable for teams that frequently needed answers during live conversations.
Second, AI-powered search significantly reduced the time required to locate relevant answers compared to traditional manual searches.
Finally, customer feedback highlighted the importance of transparency and trust in AI responses, reinforcing the need for clear source references and predictable answer formatting.
These insights helped shape the next iterations of the Copilot Agent experience and informed how Loopio approaches AI-powered knowledge retrieval within existing workflows.
(Validated this by running a UMUX Survey & 1:1 feedback calls)
Impact
The beta program validated several core assumptions about the Copilot Agent concept.
First, integrating Loopio knowledge directly into Microsoft tools proved valuable for teams that frequently needed answers during live conversations.
Second, AI-powered search significantly reduced the time required to locate relevant answers compared to traditional manual searches.
Finally, customer feedback highlighted the importance of transparency and trust in AI responses, reinforcing the need for clear source references and predictable answer formatting.
These insights helped shape the next iterations of the Copilot Agent experience and informed how Loopio approaches AI-powered knowledge retrieval within existing workflows.
(Validated this by running a UMUX Survey & 1:1 feedback calls)
