OVERVIEW
Making compliance invisible → 70% faster Certificate of Analysis document creation
How I designed a template builder that makes compliance structural, giving QA teams flexibility without the fear of making errors or breaking rules.
Problem
QA teams spent ~45 min per COA, manually copying data and relying on memory to ensure compliance. Errors were inevitable.
Solution
A COA builder feature designed as a core part of the platform with a three-tier permission model - locked, partially locked, and free sections, so compliance is structural, not manual.
Impact
70% Faster COA creation
~85% Fewer errors
89% Template reuse
< Back
EAT SAFE VERIFIED
EAT SAFE VERIFIED
/
/
SHIPPED 2022
SHIPPED 2022
Making compliance invisible → 70% faster Certificate of Analysis document creation
Making compliance invisible → 70% faster Certificate of Analysis document creation
How I designed a template builder that makes compliance structural, giving QA teams flexibility without the fear of making errors or breaking rules.
How I designed a template builder that makes compliance structural, giving QA teams flexibility without the fear of making errors or breaking rules.
UX/UI
UX/UI
Systems Design
Systems Design
B2B SaaS
B2B SaaS
0 → 1
0 → 1
TL;DR
TL;DR
Problem
Problem
QA teams spent ~45 min per COA, manually copying data and relying on memory to ensure compliance. Errors were inevitable.
QA teams spent ~45 min per COA, manually copying data and relying on memory to ensure compliance. Errors were inevitable.
Solution
Solution
A COA builder feature designed as a core part of the platform with a three-tier permission model - locked, partially locked, and free sections, so compliance is structural, not manual.
A COA builder feature designed as a core part of the platform with a three-tier permission model - locked, partially locked, and free sections, so compliance is structural, not manual.
Impact
Impact
70%
70%
Faster COA creation
Faster COA creation
~85%
~85%
Fewer errors
Fewer errors
89%
89%
Template Reuse
Template Reuse
OVERVIEW
OVERVIEW
DISCOVERY
The problem hiding in plain sight
The problem hiding in plain sight
When I joined EatSafeVerified, the brief for my first project seemed straightforward:
Help QA teams create Certificates of Analysis faster.
But I noticed something strange. These critical documents, the ones our entire product was supposed to support, were being assembled outside the platform. In Word. From templates saved in personal drives.
I needed to understand why. So I interviewed five QA managers across different company sizes and product types. Despite their differences, their workflows were almost identical.
When I joined EatSafeVerified, the brief for my first project seemed straightforward:
Help QA teams create Certificates of Analysis faster.
But I noticed something strange. These critical documents, the ones our entire product was supposed to support, were being assembled outside the platform. In Word. From templates saved in personal drives.
I needed to understand why. So I interviewed five QA managers across different company sizes and product types. Despite their differences, their workflows were almost identical.
What I saw
What I saw
Here’s a typical walkthrough. A QA Manager at a mid-sized organic supplier has been building the same Certificate of Analysis for forty-three minutes. Exports test results from the lab system. Finds the right template buried somewhere in a personal drive. Copies values by hand. Cross-references FDA requirements. Double-checks everything...triple-checks the regulatory language. Generate PDF. Hope everything was right and nothing was missed.
Here’s a typical walkthrough. A QA Manager at a mid-sized organic supplier has been building the same Certificate of Analysis for forty-three minutes. Exports test results from the lab system. Finds the right template buried somewhere in a personal drive. Copies values by hand. Cross-references FDA requirements. Double-checks everything...triple-checks the regulatory language. Generate PDF. Hope everything was right and nothing was missed.
~45 minutes, 6 App switches
~45 minutes, 6 App switches
Every step was manual. Every step was a chance for error.
Every step was manual. Every step was a chance for error.
"I know what a correct COA looks like... I just don't have the time to rebuild it every time."
"I know what a correct COA looks like... I just don't have the time to rebuild it every time."
- QA Manager, during research
- QA Manager, during research
This was the daily reality for most QA teams across our user base, and beyond. And in a zero-tolerance compliance domain, this was a problem that needed a solution.
This was the daily reality for most QA teams across our user base, and beyond. And in a zero-tolerance compliance domain, this was a problem that needed a solution.
But wait, what is ESV? And what is a COA!?
But wait, what is ESV? And what is a COA!?
EatSafeVerified (ESV) is a B2B food safety platform used by suppliers, manufacturers, and QA teams to manage the entire workflow - from sample submission and lab test tracking to document generation and audit preparation. Clients range from small organic farms to large-scale food manufacturers.
EatSafeVerified (ESV) is a B2B food safety platform used by suppliers, manufacturers, and QA teams to manage the entire workflow - from sample submission and lab test tracking to document generation and audit preparation. Clients range from small organic farms to large-scale food manufacturers.
A Certificate of Analysis (COA) is the regulatory backbone of food safety. They're official documents showing test results, product details, and compliance statements...it’s essentially a product's safety passport. A COA is what auditors ask for during FDA inspections. A missing field or outdated regulatory language can invalidate an entire product batch, delay shipments worth hundreds of thousands of dollars, or worse, trigger a recall.
A Certificate of Analysis (COA) is the regulatory backbone of food safety. They're official documents showing test results, product details, and compliance statements...it’s essentially a product's safety passport. A COA is what auditors ask for during FDA inspections. A missing field or outdated regulatory language can invalidate an entire product batch, delay shipments worth hundreds of thousands of dollars, or worse, trigger a recall.
My role in this project
My role in this project
I joined ESV as a Product Designer around February 2022. After a few days of onboarding and a few weeks of properly getting to know the platform top-bottom, I, along with Abhay, were tasked with owning this problem end-to-end. That meant framing the system problem, defining design strategy, making decisions about how compliance logic should work, and partnering with engineering and compliance teams to ship a solution.
I joined ESV as a Product Designer around February 2022. After a few days of onboarding and a few weeks of properly getting to know the platform top-bottom, I, along with Abhay, were tasked with owning this problem end-to-end. That meant framing the system problem, defining design strategy, making decisions about how compliance logic should work, and partnering with engineering and compliance teams to ship a solution.
Role
Role
Product Designer
Product Designer
Team
Team
2 Product Designers, 2 Engineers, 1 PM
2 Product Designers, 2 Engineers, 1 PM
Timeline
Timeline
~6 months | Apr - Oct 2022
~6 months | Apr - Oct 2022
Skills
Skills
Problem definition & UX strategy, User research & workflow audits, Information architecture & interaction design, Prototyping, usability testing, iteration, Cross-functional collaboration with engineering, QA, and SMEs
Problem definition & UX strategy, User research & workflow audits, Information architecture & interaction design, Prototyping, usability testing, iteration, Cross-functional collaboration with engineering, QA, and SMEs
The initial brief was about speed, "Help QA teams create COAs faster." But after watching those workflows, I realised we were solving the wrong problem.
The issue wasn't speed. The issue was that correctness depended on people being careful. And in regulated domains, that's not a workflow, it's a liability.
The initial brief was about speed, "Help QA teams create COAs faster." But after watching those workflows, I realised we were solving the wrong problem.
The issue wasn't speed. The issue was that correctness depended on people being careful. And in regulated domains, that's not a workflow, it's a liability.
DISCOVERY
RESEARCH
Getting uncomfortable with the truth
Getting uncomfortable with the truth
Getting to know our users meant asking questions that made us get to the root of the problem...
Getting to know our users meant asking questions that made us get to the root of the problem...
"What’s your workflow like generally creating a COA?"
"What’s your workflow like generally creating a COA?"
"Walk me through the last time you caught a compliance error."
"Walk me through the last time you caught a compliance error."
"What one thing you don’t like about this process?”
"What one thing you don’t like about this process?”
I spent a week interviewing and shadowing QA specialists.
I spent a week interviewing and shadowing QA specialists.
🗣️
5 QA MANAGERS
5 QA MANAGERS
In-depth Interviews
In-depth Interviews
👀
2 WEEKS SHADOWING
2 WEEKS SHADOWING
Contextual Inquiry
Contextual Inquiry
📊
USAGE ANALYTICS
USAGE ANALYTICS
Quantitative Data
Quantitative Data
‘The Templates were archaeological artifacts’
‘The Templates were archaeological artifacts’
Here's what struck me: those personal drive templates weren't just used for convenience. They were years of institutional knowledge encoded in formatting decisions and consistency, field placements, and layout choices that nobody fully remembered anymore.
Teams avoided rebuilding them because the invisible expertise would be lost. Every template was an artifact of lessons learned the hard way.
Here's what struck me: those personal drive templates weren't just used for convenience. They were years of institutional knowledge encoded in formatting decisions and consistency, field placements, and layout choices that nobody fully remembered anymore.
Teams avoided rebuilding them because the invisible expertise would be lost. Every template was an artifact of lessons learned the hard way.
I realised we weren't dealing with a documentation problem. We were dealing with a responsibility problem. The platform was delegating its most critical responsibility, compliance , to human vigilance.
I realised we weren't dealing with a documentation problem. We were dealing with a responsibility problem. The platform was delegating its most critical responsibility, compliance , to human vigilance.
What the numbers revealed
What the numbers revealed
Digging into usage data with the team revealed patterns that matched what I was seeing:
Digging into usage data with the team revealed patterns that matched what I was seeing:
6-8
6-8
Compliance errors per month
Compliance errors per month
Compliance errors per month
60%
60%
From outdated templates
From outdated templates
From outdated templates
0%
0%
Direct template reuse
Direct template reuse
Direct template reuse
100%
100%
Manual review before sending
Manual review before sending
Manual review before sending
Most compliance errors didn't come from users being careless.
Most compliance errors didn't come from users being careless.
They came from users carefully following outdated templates.
They came from users carefully following outdated templates.
The error existed before the work began.
The error existed before the work began.
RESEARCH
FRAMING
The Flexibility vs Safety standoff
The Flexibility vs Safety standoff
There was this tension that kept turning up during research...
There was this tension that kept turning up during research...
QA teams genuinely needed flexibility. Customisation wasn't just for looks and aesthetics... it was part of client trust and professional credibility. Different clients wanted different layouts, branding, and presentation styles.
QA teams genuinely needed flexibility. Customisation wasn't just for looks and aesthetics... it was part of client trust and professional credibility. Different clients wanted different layouts, branding, and presentation styles.
But any flexibility that allowed required fields or regulatory language to be altered introduced unacceptable risk.
But any flexibility that allowed required fields or regulatory language to be altered introduced unacceptable risk.
Too much freedom
Too much freedom
Missing mandatory fields
Missing mandatory fields
Outdated regulatory language
Outdated regulatory language
Transcription errors
Transcription errors
Audit failures
Audit failures
Too much restriction
Too much restriction
Rigid, one-size-fits-all output
Rigid, one-size-fits-all output
Workarounds outside the system
Workarounds outside the system
Unhappy clients
Unhappy clients
Eventually back to personal templates
Eventually back to personal templates
Both approaches depended on vigilance. And any design in this context that requires people to be careful has already failed.
Both approaches depended on vigilance. And any design in this context that requires people to be careful has already failed.
The question that led to the solution
The question that led to the solution
What if we stopped asking "how do we help users avoid mistakes" and started asking "how do we make mistakes structurally impossible"?
What if we stopped asking "how do we help users avoid mistakes" and started asking "how do we make mistakes structurally impossible"?
FRAMING
SOLUTION
What if documents were systems ??
What if documents were systems ??
The solution emerged from a shift in perspective.
The solution emerged from a shift in perspective.
COAs aren't really documents. They're assemblies of parts - each with its own rules about what can change.
COAs aren't really documents. They're assemblies of parts - each with its own rules about what can change.
I worked with engineering to decompose COAs into modular sections : branding, product metadata, test results, regulatory text, signatures, each with explicit permissions:
I worked with engineering to decompose COAs into modular sections : branding, product metadata, test results, regulatory text, signatures, each with explicit permissions:
Information Architecture
Information Architecture
Certificate of Analysis
Certificate of Analysis
🔒
LOCKED
Test results
Regulatory statements
Regulatory statements
Product name
System-controlled. Can reposition, can't edit content.
🔐
PARTIALLY LOCKED
Product metadata
Signature block
Contact details
Some fields locked, others editable. Layout flexible.
✏️
FREE
Company branding
Custom sections
Text boxes
Full control. Add, edit, delete, style as needed.
💡 The Key Insight
💡 The Key Insight
Compliance logic was enforced structurally. Invalid states simply can’t be created.
If the system let you do something, it was safe.
If it didn't, there was a reason.
Compliance logic was enforced structurally. Invalid states simply can’t be created.
If the system let you do something, it was safe.
If it didn't, there was a reason.
The system owns correctness. Users own customisation.
The system owns correctness. Users own customisation.
and finally...The COA Builder
and finally...The COA Builder
The workflow became straightforward and simple:
The workflow became straightforward and simple:
1 Build template
1 Build template
One time setup
One time setup
2 Select template
2 Select template
From Library
From Library
3 Apply Product
3 Apply Product
Auto-populate data
Auto-populate data
4 Generate COA
4 Generate COA
Validated output
Validated output
Drag-and-droppable data fields
Drag-and-droppable data fields
Users can drag data fields from the right panel directly onto sections. These fields auto-populate when a product is applied - no manual entry required.
Users can drag data fields from the right panel directly onto sections. These fields auto-populate when a product is applied - no manual entry required.
Freedom to customize
Freedom to customize
Easily customise headers, logo, text and placement to match your company or brand.
Easily customise headers, logo, text and placement to match your company or brand.
Straightforward end-to-end process
Straightforward end-to-end process
Create a template, select a product, and apply the template to the product to create a detailed COA.
Create a template, select a product, and apply the template to the product to create a detailed COA.
SOLUTION
TESTING
This wasn’t the final answer...
This wasn’t the final answer...
Naturally, the first design had problems.
Naturally, the first design had problems.
Early usability testing revealed a problem: users couldn't tell what was editable vs. locked until they tried to interact with it. This created constant uncertainty.
Early usability testing revealed a problem: users couldn't tell what was editable vs. locked until they tried to interact with it. This created constant uncertainty.
"I'm not sure if this is a bug or if I'm not supposed to do this."
"I'm not sure if this is a bug or if I'm not supposed to do this."
- Participant during Round 1 testing
- Participant during Round 1 testing
Our first solution? Tooltips , explaining each permission level.
Our first solution? Tooltips , explaining each permission level.
They were completely useless because nobody hovers long enough everytime to read tooltips.
They were completely useless because nobody hovers long enough everytime to read tooltips.
What actually worked...
What actually worked...
Persistent visual hierarchy
Persistent visual hierarchy
Lock badges on required sections, visible at all times. Distinct colours by permission level. No guessing.
Lock badges on required sections, visible at all times. Distinct colours by permission level. No guessing.
Silent Prevention
Silent Prevention
Cursor becomes a lock icon when hovering over locked content. Grabber over moveable content.
Cursor becomes a lock icon when hovering over locked content. Grabber over moveable content.
By Round 3 testing, hesitation had almost completely disappeared. Users moved decisively. When the system prevented an action, they accepted it as a structural boundary, not a failure.
By Round 3 testing, hesitation had almost completely disappeared. Users moved decisively. When the system prevented an action, they accepted it as a structural boundary, not a failure.
✨ The moment we knew it worked!
A QA specialist tried to delete a required section, saw it snap back, and immediately said
" Oh, right...can't remove that."
No frustration. No confusion. The interaction taught her the rule clearly.
A QA specialist tried to delete a required section, saw it snap back, and immediately said
" Oh, right...can't remove that."
No frustration. No confusion. The interaction taught her the rule clearly.
TESTING
IMPACT
The numbers ( and what they don’t tell you )
The numbers
( and what they don’t tell you )
70%
70%
70%
Faster creation time
Faster creation time
Faster creation time
~85%
~85%
~85%
Fewer compliance errors
Fewer compliance errors
Fewer compliance errors
89%
89%
89%
Template reuse
Template reuse
Template reuse
But the numbers weren’t the real story...
But the numbers weren’t the real story...
🧠 The behavioural shifts were!
Teams stopped double checking
Teams stopped double checking
In the first weeks after launch, QA teams still manually reviewed every generated COA - old habits. By week four, that behaviour had almost completely disappeared. They'd internalised that even if they wanted to create an invalid COA, the system wouldn't let them.
In the first weeks after launch, QA teams still manually reviewed every generated COA - old habits. By week four, that behaviour had almost completely disappeared. They'd internalised that even if they wanted to create an invalid COA, the system wouldn't let them.
Expertise shifted focus
Expertise shifted focus
QA managers stopped spending cognitive energy on compliance mechanics. Instead, they focused on what actually required human judgment: interpreting borderline test results, advising clients on edge cases, analysing patterns across products.
QA managers stopped spending cognitive energy on compliance mechanics. Instead, they focused on what actually required human judgment: interpreting borderline test results, advising clients on edge cases, analysing patterns across products.
Onboarding transformed
Onboarding transformed
Before : 2-3 weeks of shadowing before new hires could create COAs independently.
Before : 2-3 weeks of shadowing before new hires could create COAs independently.
After : New team members productive on day one. The system encoded the institutional knowledge.
After : New team members productive on day one. The system encoded the institutional knowledge.