Product Strategy
Provider Payments UX Audit
Payments decides whether a business gets paid and whether it stays. I audited every payment product a provider touches, used AI to do the heavy lifting, and packaged the work as a reusable skill for the whole product design team.
Year :
2026
Industry :
SaaS · Beauty & wellness · Payments
Client :
Booksy
Project Duration :
2 months
Problem :
Activation is leaking
Providers were signing up and not reaching their first booking. The numbers were stark: 63% of providers get zero bookings in their first two weeks, and 60% of consumers drop off in the booking flow on the other side. 81.5% of bookings are rebookings, so a provider’s own customers are most of what keeps them active. A provider who never gets set up never gets those rebookings either.
Each product made sense on its own
Verification gated most payment features without making that clear. Settings and Payments & Checkout were mixed together, so providers couldn’t tell where anything lived. The same “off” state was drawn three different ways, and “Accepted” on web was “Active” on mobile. Nothing adapted to where a provider actually was: verified, unverified, or halfway through looked and behaved the same.
Nobody owned the path
The activation journey cut across at least three squads, and no one owned the whole of it. Each team optimised its own screen, so the journey as a whole kept getting worse.
Solution :
How I built it
1. Mapped every flow in Figma. I laid out the payments flows screen by screen, from verification to checkout, so every finding had a fixed place in the journey before any analysis started.
2. Set up Claude to scan every screen. Frames were pulled from each recording at set intervals and laid out as contact sheets, alongside the screenshots. Claude scanned all of them, so nothing depended on what I happened to notice on a first watch.
3. Analysed each screen, with my expertise on top. Claude produced a first-pass analysis of every screen and step: what the provider has to do, where the flow goes unclear, and where the UI hides the state a provider needs. I reviewed each pass, kept what held up, rejected what didn’t, and added what only product and payments experience would catch, such as when a fee is disclosed and how a badge can contradict the screen behind it.
4. Reconciled every finding against the source. Each finding was checked against the annotations already on the Figma canvas and against fresh high-resolution screenshots. Timestamps and pin positions were read from what was on screen, never guessed.
A digestible audit, not another document
A static document gets summarised by AI and forgotten. So I built the output as an interactive report around the evidence:
Every finding sits next to the recording or screenshot it refers to. Click a finding and the video jumps to that second.
Findings follow the order the provider experiences them, so the journey reads top to bottom.
Nine themes, five strategic priorities, and a products-by-themes matrix that shows where each squad’s work lands.
Each reader gets a different entry point. Leaders start at the strategic priorities, squads jump to their own row in the matrix, and anyone can open a single flow and watch it.
AI at scale, with people making the calls
A 276-finding audit across 12 products would take a small team months to assemble by hand. Using AI for throughput let me:
Ship a usable artifact fast. A repeatable pipeline encodes the recordings, embeds every asset in one self-contained file, and rebuilds with a single command. Changes that used to mean a full rebuild now take one step.
Keep judgement where it belongs. Which findings mattered, how to frame them for leadership, and which themes became priorities were my calls.
Delivery :
From audit to roadmap
The audit was the means, not the deliverable. Delivery meant getting it in front of the right people and turning findings into decisions they could act on.
Output
The walkthrough with financial services leadership produced a north-star direction for provider payments and three initiatives proposed for Q4: navigation, discoverability, and an activation and KYC revamp. Each one is traceable to specific findings in the report.
Build milestones
✅ Provider journey mapped end to end across 12 payment products
✅ 276 findings organised into 9 themes and 5 strategic priorities
✅ Interactive report with annotated recordings and screenshots
✅ Products-by-themes matrix for squad-level ownership
✅ Figma UX audit skill, packaged for the whole product design team to reuse
✅ Walked through with financial services leadership, with a proposed Q4 roadmap
Craft decisions
Recordings that blurred. Compressing the videos to fit one page first made the text unreadable. I stopped chasing lower bitrates and encoded at the size the page actually displays. That kept the text sharp inside a single file.
Evidence over opinion. Every finding links to a recording or screenshot. I would rather defend 276 findings I can show than 40 I can only describe.
Success Analysis :
Decisions, not just insight
The test of an audit is whether it changes what gets built. Three things to point to so far:
Three initiatives recorded as Q4 priorities in the financial services leadership meeting: navigation, discoverability, and an activation and KYC revamp.
One shared evidence base. The same recordings and findings are now the reference point for the activation discussion, rather than competing anecdotes.
A reusable foundation. The pipeline, evidence standard and report format carry over to the next audit, which starts from a working system instead of a blank page.
Primary goal: payments adoption
Grow the share of providers who turn payment features on, not only fix the drop-off after. Targets are set per market, since Poland and the US differ.
Secondary goal: fewer payments support tickets
Reduce payment-related CS tickets by a target percentage, once the baseline is agreed.
AI leverage
The structure is what makes this reusable. Findings are tagged to themes and products, so new recordings can be checked against what is already known instead of starting over. The same approach that built this audit can keep it current.
Next measures
Activation rate: share of new providers reaching their first booking, per market
Payment feature adoption within the first 30 days
Payment-related CS tickets per month, against the agreed baseline
Squad adoption: how many squads pull findings from the matrix into their planning
More Projects
Product Strategy
Provider Payments UX Audit
Payments decides whether a business gets paid and whether it stays. I audited every payment product a provider touches, used AI to do the heavy lifting, and packaged the work as a reusable skill for the whole product design team.
Year :
2026
Industry :
SaaS · Beauty & wellness · Payments
Client :
Booksy
Project Duration :
2 months
Problem :
Activation is leaking
Providers were signing up and not reaching their first booking. The numbers were stark: 63% of providers get zero bookings in their first two weeks, and 60% of consumers drop off in the booking flow on the other side. 81.5% of bookings are rebookings, so a provider’s own customers are most of what keeps them active. A provider who never gets set up never gets those rebookings either.
Each product made sense on its own
Verification gated most payment features without making that clear. Settings and Payments & Checkout were mixed together, so providers couldn’t tell where anything lived. The same “off” state was drawn three different ways, and “Accepted” on web was “Active” on mobile. Nothing adapted to where a provider actually was: verified, unverified, or halfway through looked and behaved the same.
Nobody owned the path
The activation journey cut across at least three squads, and no one owned the whole of it. Each team optimised its own screen, so the journey as a whole kept getting worse.
Solution :
How I built it
1. Mapped every flow in Figma. I laid out the payments flows screen by screen, from verification to checkout, so every finding had a fixed place in the journey before any analysis started.
2. Set up Claude to scan every screen. Frames were pulled from each recording at set intervals and laid out as contact sheets, alongside the screenshots. Claude scanned all of them, so nothing depended on what I happened to notice on a first watch.
3. Analysed each screen, with my expertise on top. Claude produced a first-pass analysis of every screen and step: what the provider has to do, where the flow goes unclear, and where the UI hides the state a provider needs. I reviewed each pass, kept what held up, rejected what didn’t, and added what only product and payments experience would catch, such as when a fee is disclosed and how a badge can contradict the screen behind it.
4. Reconciled every finding against the source. Each finding was checked against the annotations already on the Figma canvas and against fresh high-resolution screenshots. Timestamps and pin positions were read from what was on screen, never guessed.
A digestible audit, not another document
A static document gets summarised by AI and forgotten. So I built the output as an interactive report around the evidence:
Every finding sits next to the recording or screenshot it refers to. Click a finding and the video jumps to that second.
Findings follow the order the provider experiences them, so the journey reads top to bottom.
Nine themes, five strategic priorities, and a products-by-themes matrix that shows where each squad’s work lands.
Each reader gets a different entry point. Leaders start at the strategic priorities, squads jump to their own row in the matrix, and anyone can open a single flow and watch it.
AI at scale, with people making the calls
A 276-finding audit across 12 products would take a small team months to assemble by hand. Using AI for throughput let me:
Ship a usable artifact fast. A repeatable pipeline encodes the recordings, embeds every asset in one self-contained file, and rebuilds with a single command. Changes that used to mean a full rebuild now take one step.
Keep judgement where it belongs. Which findings mattered, how to frame them for leadership, and which themes became priorities were my calls.
Delivery :
From audit to roadmap
The audit was the means, not the deliverable. Delivery meant getting it in front of the right people and turning findings into decisions they could act on.
Output
The walkthrough with financial services leadership produced a north-star direction for provider payments and three initiatives proposed for Q4: navigation, discoverability, and an activation and KYC revamp. Each one is traceable to specific findings in the report.
Build milestones
✅ Provider journey mapped end to end across 12 payment products
✅ 276 findings organised into 9 themes and 5 strategic priorities
✅ Interactive report with annotated recordings and screenshots
✅ Products-by-themes matrix for squad-level ownership
✅ Figma UX audit skill, packaged for the whole product design team to reuse
✅ Walked through with financial services leadership, with a proposed Q4 roadmap
Craft decisions
Recordings that blurred. Compressing the videos to fit one page first made the text unreadable. I stopped chasing lower bitrates and encoded at the size the page actually displays. That kept the text sharp inside a single file.
Evidence over opinion. Every finding links to a recording or screenshot. I would rather defend 276 findings I can show than 40 I can only describe.
Success Analysis :
Decisions, not just insight
The test of an audit is whether it changes what gets built. Three things to point to so far:
Three initiatives recorded as Q4 priorities in the financial services leadership meeting: navigation, discoverability, and an activation and KYC revamp.
One shared evidence base. The same recordings and findings are now the reference point for the activation discussion, rather than competing anecdotes.
A reusable foundation. The pipeline, evidence standard and report format carry over to the next audit, which starts from a working system instead of a blank page.
Primary goal: payments adoption
Grow the share of providers who turn payment features on, not only fix the drop-off after. Targets are set per market, since Poland and the US differ.
Secondary goal: fewer payments support tickets
Reduce payment-related CS tickets by a target percentage, once the baseline is agreed.
AI leverage
The structure is what makes this reusable. Findings are tagged to themes and products, so new recordings can be checked against what is already known instead of starting over. The same approach that built this audit can keep it current.
Next measures
Activation rate: share of new providers reaching their first booking, per market
Payment feature adoption within the first 30 days
Payment-related CS tickets per month, against the agreed baseline
Squad adoption: how many squads pull findings from the matrix into their planning
More Projects
Product Strategy
Provider Payments UX Audit
Payments decides whether a business gets paid and whether it stays. I audited every payment product a provider touches, used AI to do the heavy lifting, and packaged the work as a reusable skill for the whole product design team.
Year :
2026
Industry :
SaaS · Beauty & wellness · Payments
Client :
Booksy
Project Duration :
2 months
Problem :
Activation is leaking
Providers were signing up and not reaching their first booking. The numbers were stark: 63% of providers get zero bookings in their first two weeks, and 60% of consumers drop off in the booking flow on the other side. 81.5% of bookings are rebookings, so a provider’s own customers are most of what keeps them active. A provider who never gets set up never gets those rebookings either.
Each product made sense on its own
Verification gated most payment features without making that clear. Settings and Payments & Checkout were mixed together, so providers couldn’t tell where anything lived. The same “off” state was drawn three different ways, and “Accepted” on web was “Active” on mobile. Nothing adapted to where a provider actually was: verified, unverified, or halfway through looked and behaved the same.
Nobody owned the path
The activation journey cut across at least three squads, and no one owned the whole of it. Each team optimised its own screen, so the journey as a whole kept getting worse.
Solution :
How I built it
1. Mapped every flow in Figma. I laid out the payments flows screen by screen, from verification to checkout, so every finding had a fixed place in the journey before any analysis started.
2. Set up Claude to scan every screen. Frames were pulled from each recording at set intervals and laid out as contact sheets, alongside the screenshots. Claude scanned all of them, so nothing depended on what I happened to notice on a first watch.
3. Analysed each screen, with my expertise on top. Claude produced a first-pass analysis of every screen and step: what the provider has to do, where the flow goes unclear, and where the UI hides the state a provider needs. I reviewed each pass, kept what held up, rejected what didn’t, and added what only product and payments experience would catch, such as when a fee is disclosed and how a badge can contradict the screen behind it.
4. Reconciled every finding against the source. Each finding was checked against the annotations already on the Figma canvas and against fresh high-resolution screenshots. Timestamps and pin positions were read from what was on screen, never guessed.
A digestible audit, not another document
A static document gets summarised by AI and forgotten. So I built the output as an interactive report around the evidence:
Every finding sits next to the recording or screenshot it refers to. Click a finding and the video jumps to that second.
Findings follow the order the provider experiences them, so the journey reads top to bottom.
Nine themes, five strategic priorities, and a products-by-themes matrix that shows where each squad’s work lands.
Each reader gets a different entry point. Leaders start at the strategic priorities, squads jump to their own row in the matrix, and anyone can open a single flow and watch it.
AI at scale, with people making the calls
A 276-finding audit across 12 products would take a small team months to assemble by hand. Using AI for throughput let me:
Ship a usable artifact fast. A repeatable pipeline encodes the recordings, embeds every asset in one self-contained file, and rebuilds with a single command. Changes that used to mean a full rebuild now take one step.
Keep judgement where it belongs. Which findings mattered, how to frame them for leadership, and which themes became priorities were my calls.
Delivery :
From audit to roadmap
The audit was the means, not the deliverable. Delivery meant getting it in front of the right people and turning findings into decisions they could act on.
Output
The walkthrough with financial services leadership produced a north-star direction for provider payments and three initiatives proposed for Q4: navigation, discoverability, and an activation and KYC revamp. Each one is traceable to specific findings in the report.
Build milestones
✅ Provider journey mapped end to end across 12 payment products
✅ 276 findings organised into 9 themes and 5 strategic priorities
✅ Interactive report with annotated recordings and screenshots
✅ Products-by-themes matrix for squad-level ownership
✅ Figma UX audit skill, packaged for the whole product design team to reuse
✅ Walked through with financial services leadership, with a proposed Q4 roadmap
Craft decisions
Recordings that blurred. Compressing the videos to fit one page first made the text unreadable. I stopped chasing lower bitrates and encoded at the size the page actually displays. That kept the text sharp inside a single file.
Evidence over opinion. Every finding links to a recording or screenshot. I would rather defend 276 findings I can show than 40 I can only describe.
Success Analysis :
Decisions, not just insight
The test of an audit is whether it changes what gets built. Three things to point to so far:
Three initiatives recorded as Q4 priorities in the financial services leadership meeting: navigation, discoverability, and an activation and KYC revamp.
One shared evidence base. The same recordings and findings are now the reference point for the activation discussion, rather than competing anecdotes.
A reusable foundation. The pipeline, evidence standard and report format carry over to the next audit, which starts from a working system instead of a blank page.
Primary goal: payments adoption
Grow the share of providers who turn payment features on, not only fix the drop-off after. Targets are set per market, since Poland and the US differ.
Secondary goal: fewer payments support tickets
Reduce payment-related CS tickets by a target percentage, once the baseline is agreed.
AI leverage
The structure is what makes this reusable. Findings are tagged to themes and products, so new recordings can be checked against what is already known instead of starting over. The same approach that built this audit can keep it current.
Next measures
Activation rate: share of new providers reaching their first booking, per market
Payment feature adoption within the first 30 days
Payment-related CS tickets per month, against the agreed baseline
Squad adoption: how many squads pull findings from the matrix into their planning