Scaling Content Sources
Led UX strategy for Loopio’s SharePoint integration under tight delivery timelines. Navigated critical architectural trade-offs impacting performance, trust, and scalability. Early GA validated key risks, enabling a course correction toward a pre-indexed, scalable solution aligned with Automated Answers.
January 19, 2026
Scaling Content Sources
Led UX strategy for Loopio’s SharePoint integration under tight delivery timelines. Navigated critical architectural trade-offs impacting performance, trust, and scalability. Early GA validated key risks, enabling a course correction toward a pre-indexed, scalable solution aligned with Automated Answers.
January 19, 2026
CLIENT
Loopio
Role
Product Designer
CLIENT
Loopio
Role
Product Designer
CLIENT
Loopio
Role
Product Designer
Overview
Understanding the problem
Customers relied heavily on SharePoint as a primary content source. To trust AI-generated answers, they needed the experience to be fast, accurate, and transparent.
Early discovery revealed that the core challenge was architectural rather than purely UI-driven:
Should Loopio fetch SharePoint files dynamically during answer generation, or pre-index content upfront?
This decision directly impacted user trust, system performance, and the long-term scalability of the product.
Key risks identified early:
Slow answer generation would reduce adoption.
Inconsistent file visibility would erode trust in AI outputs.
Lack of bulk generation would limit enterprise value.
Early hypothesis and recommendation
Based on discovery and early collaboration with engineering, I strongly recommended a pre-indexing approach.
This was not a design preference, but a product and trust decision with long-term implications.
Rationale:
Faster and more predictable answer generation.
Improved accuracy through structured, indexed content.
Support for bulk answer generation.
Clear visibility, logging, and editability of synced files.
Ability to use a native Loopio file picker instead of a one-way Microsoft iframe.
To validate this, I partnered with the PM and Tech Lead to run a full proof of concept, testing both technical feasibility and UX implications.
Overview
Understanding the problem
Customers relied heavily on SharePoint as a primary content source. To trust AI-generated answers, they needed the experience to be fast, accurate, and transparent.
Early discovery revealed that the core challenge was architectural rather than purely UI-driven:
Should Loopio fetch SharePoint files dynamically during answer generation, or pre-index content upfront?
This decision directly impacted user trust, system performance, and the long-term scalability of the product.
Key risks identified early:
Slow answer generation would reduce adoption.
Inconsistent file visibility would erode trust in AI outputs.
Lack of bulk generation would limit enterprise value.
Early hypothesis and recommendation
Based on discovery and early collaboration with engineering, I strongly recommended a pre-indexing approach.
This was not a design preference, but a product and trust decision with long-term implications.
Rationale:
Faster and more predictable answer generation.
Improved accuracy through structured, indexed content.
Support for bulk answer generation.
Clear visibility, logging, and editability of synced files.
Ability to use a native Loopio file picker instead of a one-way Microsoft iframe.
To validate this, I partnered with the PM and Tech Lead to run a full proof of concept, testing both technical feasibility and UX implications.
Overview
Understanding the problem
Customers relied heavily on SharePoint as a primary content source. To trust AI-generated answers, they needed the experience to be fast, accurate, and transparent.
Early discovery revealed that the core challenge was architectural rather than purely UI-driven:
Should Loopio fetch SharePoint files dynamically during answer generation, or pre-index content upfront?
This decision directly impacted user trust, system performance, and the long-term scalability of the product.
Key risks identified early:
Slow answer generation would reduce adoption.
Inconsistent file visibility would erode trust in AI outputs.
Lack of bulk generation would limit enterprise value.
Early hypothesis and recommendation
Based on discovery and early collaboration with engineering, I strongly recommended a pre-indexing approach.
This was not a design preference, but a product and trust decision with long-term implications.
Rationale:
Faster and more predictable answer generation.
Improved accuracy through structured, indexed content.
Support for bulk answer generation.
Clear visibility, logging, and editability of synced files.
Ability to use a native Loopio file picker instead of a one-way Microsoft iframe.
To validate this, I partnered with the PM and Tech Lead to run a full proof of concept, testing both technical feasibility and UX implications.
Constraints
Leadership decision and constraints
We presented findings, hypotheses, and trade-offs to the Senior Leadership Team.
Despite the advantages of pre-indexing, the decision was made to proceed with an on-the-fly approach to reduce time to market, driven by sales pressure and upcoming enterprise renewals.
Once the decision was final, my role shifted from advocating for the ideal solution to designing guardrails within a constrained architecture. The focus became reducing user confusion, managing expectations, and minimising trust risks while meeting business urgency.
This was a defining moment in balancing product quality with delivery constraints.
Discovery and validation under constraints
Even within architectural limitations, I ensured we followed a rigorous validation process:
Conducted customer calls to understand SharePoint structures and workflows.
Walked customers through proposed concepts to validate mental models.
Ran unmoderated usability tests using Great Question.
Partnered closely with engineering through UAT, Alpha, and Beta phases.
Key insights included:
Users expected clear visibility into which files were used during answer generation.
Enterprise SharePoint structures varied widely, making predictability critical.
Several improvements were deprioritised due to timelines. Throughout, I worked with the tripod to continuously reassess risk and document known gaps.
Constraints
Leadership decision and constraints
We presented findings, hypotheses, and trade-offs to the Senior Leadership Team.
Despite the advantages of pre-indexing, the decision was made to proceed with an on-the-fly approach to reduce time to market, driven by sales pressure and upcoming enterprise renewals.
Once the decision was final, my role shifted from advocating for the ideal solution to designing guardrails within a constrained architecture. The focus became reducing user confusion, managing expectations, and minimising trust risks while meeting business urgency.
This was a defining moment in balancing product quality with delivery constraints.
Discovery and validation under constraints
Even within architectural limitations, I ensured we followed a rigorous validation process:
Conducted customer calls to understand SharePoint structures and workflows.
Walked customers through proposed concepts to validate mental models.
Ran unmoderated usability tests using Great Question.
Partnered closely with engineering through UAT, Alpha, and Beta phases.
Key insights included:
Users expected clear visibility into which files were used during answer generation.
Enterprise SharePoint structures varied widely, making predictability critical.
Several improvements were deprioritised due to timelines. Throughout, I worked with the tripod to continuously reassess risk and document known gaps.
Constraints
Leadership decision and constraints
We presented findings, hypotheses, and trade-offs to the Senior Leadership Team.
Despite the advantages of pre-indexing, the decision was made to proceed with an on-the-fly approach to reduce time to market, driven by sales pressure and upcoming enterprise renewals.
Once the decision was final, my role shifted from advocating for the ideal solution to designing guardrails within a constrained architecture. The focus became reducing user confusion, managing expectations, and minimising trust risks while meeting business urgency.
This was a defining moment in balancing product quality with delivery constraints.
Discovery and validation under constraints
Even within architectural limitations, I ensured we followed a rigorous validation process:
Conducted customer calls to understand SharePoint structures and workflows.
Walked customers through proposed concepts to validate mental models.
Ran unmoderated usability tests using Great Question.
Partnered closely with engineering through UAT, Alpha, and Beta phases.
Key insights included:
Users expected clear visibility into which files were used during answer generation.
Enterprise SharePoint structures varied widely, making predictability critical.
Several improvements were deprioritised due to timelines. Throughout, I worked with the tripod to continuously reassess risk and document known gaps.
Insights
GA outcome and learning
Post-GA, engagement with the SharePoint integration was lower than expected, providing clear validation of earlier risks.
Customer feedback consistently highlighted:
Slow answer generation.
Lack of confidence in which files were being used.
Missing enterprise workflows such as bulk generation.
Rather than treating this as a failure, I used this feedback to validate earlier hypotheses and build a strong evidence-based case for revisiting the architecture.
Course correction: returning to pre-indexing
After completing work on another content source, I re-engaged with the PM and Tech Lead to revisit SharePoint.
This time:
The pre-indexing approach was approved.
Scope expanded to support Automated Answers.
Collaboration deepened with the Library team to manage dependencies and risk.
I created demo walkthrough videos and shared them with CX for early feedback.
Senior leadership alignment was secured using GA learnings and real customer evidence.
The focus shifted from optimising a constrained solution to building a scalable foundation.
Insights
GA outcome and learning
Post-GA, engagement with the SharePoint integration was lower than expected, providing clear validation of earlier risks.
Customer feedback consistently highlighted:
Slow answer generation.
Lack of confidence in which files were being used.
Missing enterprise workflows such as bulk generation.
Rather than treating this as a failure, I used this feedback to validate earlier hypotheses and build a strong evidence-based case for revisiting the architecture.
Course correction: returning to pre-indexing
After completing work on another content source, I re-engaged with the PM and Tech Lead to revisit SharePoint.
This time:
The pre-indexing approach was approved.
Scope expanded to support Automated Answers.
Collaboration deepened with the Library team to manage dependencies and risk.
I created demo walkthrough videos and shared them with CX for early feedback.
Senior leadership alignment was secured using GA learnings and real customer evidence.
The focus shifted from optimising a constrained solution to building a scalable foundation.
Insights
GA outcome and learning
Post-GA, engagement with the SharePoint integration was lower than expected, providing clear validation of earlier risks.
Customer feedback consistently highlighted:
Slow answer generation.
Lack of confidence in which files were being used.
Missing enterprise workflows such as bulk generation.
Rather than treating this as a failure, I used this feedback to validate earlier hypotheses and build a strong evidence-based case for revisiting the architecture.
Course correction: returning to pre-indexing
After completing work on another content source, I re-engaged with the PM and Tech Lead to revisit SharePoint.
This time:
The pre-indexing approach was approved.
Scope expanded to support Automated Answers.
Collaboration deepened with the Library team to manage dependencies and risk.
I created demo walkthrough videos and shared them with CX for early feedback.
Senior leadership alignment was secured using GA learnings and real customer evidence.
The focus shifted from optimising a constrained solution to building a scalable foundation.
Decisions
Final solution
The redesigned SharePoint experience:
Pre-indexes content upfront, significantly reducing generation time.
Improves answer accuracy and reliability.
Supports bulk answer generation for enterprise workflows.
Provides clear visibility into synced and used content.
Aligns SharePoint with Loopio’s broader Automated Answers and content integration strategy.
All major design decisions were reviewed within the tripod before broader rollout.
Impact and success signals
Early signals show:
Directionally faster answer generation.
Increased confidence in AI-generated responses.
Better alignment across content integrations.
Reduced friction for enterprise workflows.
Fewer CX escalations related to content visibility.
Long-term success is measured through adoption, generation speed, accuracy feedback, and support volume.
Decisions
Final solution
The redesigned SharePoint experience:
Pre-indexes content upfront, significantly reducing generation time.
Improves answer accuracy and reliability.
Supports bulk answer generation for enterprise workflows.
Provides clear visibility into synced and used content.
Aligns SharePoint with Loopio’s broader Automated Answers and content integration strategy.
All major design decisions were reviewed within the tripod before broader rollout.
Impact and success signals
Early signals show:
Directionally faster answer generation.
Increased confidence in AI-generated responses.
Better alignment across content integrations.
Reduced friction for enterprise workflows.
Fewer CX escalations related to content visibility.
Long-term success is measured through adoption, generation speed, accuracy feedback, and support volume.
Decisions
Final solution
The redesigned SharePoint experience:
Pre-indexes content upfront, significantly reducing generation time.
Improves answer accuracy and reliability.
Supports bulk answer generation for enterprise workflows.
Provides clear visibility into synced and used content.
Aligns SharePoint with Loopio’s broader Automated Answers and content integration strategy.
All major design decisions were reviewed within the tripod before broader rollout.
Impact and success signals
Early signals show:
Directionally faster answer generation.
Increased confidence in AI-generated responses.
Better alignment across content integrations.
Reduced friction for enterprise workflows.
Fewer CX escalations related to content visibility.
Long-term success is measured through adoption, generation speed, accuracy feedback, and support volume.
My role
Ownership
Owned UX strategy and end-to-end user flows.
Framed architectural hypotheses and supported them with POC and customer evidence.
Drove expectation-setting and trust-focused design decisions.
Actively partnered with PM and Tech Lead to navigate trade-offs.
Influenced leadership decisions through data and user feedback.
Helped course-correct the product toward a scalable, future-ready architecture.
My role
Ownership
Owned UX strategy and end-to-end user flows.
Framed architectural hypotheses and supported them with POC and customer evidence.
Drove expectation-setting and trust-focused design decisions.
Actively partnered with PM and Tech Lead to navigate trade-offs.
Influenced leadership decisions through data and user feedback.
Helped course-correct the product toward a scalable, future-ready architecture.
My role
Ownership
Owned UX strategy and end-to-end user flows.
Framed architectural hypotheses and supported them with POC and customer evidence.
Drove expectation-setting and trust-focused design decisions.
Actively partnered with PM and Tech Lead to navigate trade-offs.
Influenced leadership decisions through data and user feedback.
Helped course-correct the product toward a scalable, future-ready architecture.
