Manychat
Redesigning Inbox:
Signal Over Surface
This case study shows how I reframed Manychat's Inbox as a prioritisation system, and the design decisions that reduced churn among the users who mattered most.
01 . Context and Problem
The Business Had a Signal Problem
The business had a retention crisis. Users had a prioritisation crisis. The team had neither framed clearly.
The retention gap
High-value sellers were churning at 15% — a segment driving ~60% of ARR. The team had a backlog of reactive fixes but no strategy anchored to the root cause.
Business goal
Grow Inbox activation and reduce churn among high-value users by 2026.
What users experienced
Sellers managing 100+ messages daily had no way to identify which conversations mattered. Every message looked the same, so high-intent buyers kept going cold.
Users goal
Spend less time scanning. Respond to the right conversations before they go cold.
02 . My role
My role beyond the pixels
At Manychat, I was the lead designer on Inbox end-to-end. I defined the problem framing after reanalysing existing data, ran 8 moderated user interviews across creators, SMBs and local businesses, and set the design principles that shaped every decision. I presented the vision to leadership, aligned the PM and engineering leads around a phased approach, and the work became a company-wide north star for AI adoption.
03 . The data
Numers never lie.
Before running a single interview, the data already told a clear story. This wasn't a UX problem, it was a revenue problem.
15%
Churn rate among high-value sellers driving ~60% of ARR
5 / 8
Interview participants who had considered cancelling Pro for WhatsApp Business native
100+
Messages per day managed with no priority signal — entirely by manual scanning
04 . Discovery
Talking to users changed everything.
I ran 8 moderated interviews across creators, infopreneurs, affiliates and local businesses. Every question was shaped by a decision we needed to make.
What we learned
- Intent was invisible: 3–5 threads opened before finding one worth replying to
- Delay kills conversion: late replies were the most common reason for missed sales
- Automation trusted for routing, not relationships
- Simpler always won: even if less powerful
05 . The Design
Making Priority Visible.
I explored and tested multiple triage models before landing on the solution, reducing cognitive load without disrupting familiar workflows. The Inbox went from a chat tool to a revenue system.
Briley Lowery
1min
Hey! I saw your sale ends soon...
92% likely to convert
AI Scoring
Filters need intent. A score doesn't. The right conversations surface on their own — no behaviour change required.
Suggested Replies
Speed without losing voice. Agents stay in control, the suggestion does the heavy lifting.
🚀 Engaged customers
5 people
Personal
1 person
0
Spam
Empty
🔥 High CR
3 people
Smart Folders
More views meant more scanning. I collapsed the inbox into three states, not to organise, but to reduce what agents need to hold in their head.
07 . What Success Looked Like
The Inbox Became a Company-Wide Bet
~22%
Faster time to first high-intent interaction vs control group
8%
Improvement in weekly retention among high-volume sellers
74%
Of sessions started in folders view, immediate triage adoption
Beyond the product metrics, the work led to something bigger: Inbox was established as a company-wide OKR
Neu Antoli
Manychat
Redesigning Inbox:
Signal Over Surface
This case study shows how I reframed Manychat's Inbox as a prioritisation system, and the design decisions that reduced churn among the users who mattered most.
01 . Context and Problem
The Business Had a Signal Problem
The business had a retention crisis. Users had a prioritisation crisis. The team had neither framed clearly.
The retention gap
High-value sellers were churning at 15% — a segment driving ~60% of ARR. The team had a backlog of reactive fixes but no strategy anchored to the root cause.
Business goal
Grow Inbox activation and reduce churn among high-value users by 2026.
What users experienced
Sellers managing 100+ messages daily had no way to identify which conversations mattered. Every message looked the same, so high-intent buyers kept going cold.
Users goal
Spend less time scanning. Respond to the right conversations before they go cold.
02 . My role
My role beyond the pixels
At Manychat, I was the lead designer on Inbox end-to-end. I defined the problem framing after reanalysing existing data, ran 8 moderated user interviews across creators, SMBs and local businesses, and set the design principles that shaped every decision. I presented the vision to leadership, aligned the PM and engineering leads around a phased approach, and the work became a company-wide north star for AI adoption.
03 . The data
Numers never lie.
Before running a single interview, the data already told a clear story. This wasn't a UX problem, it was a revenue problem.
15%
Churn rate among high-value sellers driving ~60% of ARR
5 / 8
Interview participants who had considered cancelling Pro for WhatsApp Business native
100+
Messages per day managed with no priority signal — entirely by manual scanning
04 . Discovery
Talking to users changed everything.
I ran 8 moderated interviews across creators, infopreneurs, affiliates and local businesses. Every question was shaped by a decision we needed to make.
What we learned
- Intent was invisible: 3–5 threads opened before finding one worth replying to
- Delay kills conversion: late replies were the most common reason for missed sales
- Automation trusted for routing, not relationships
- Simpler always won: even if less powerful
05 . The Design
Making Priority Visible.
I explored and tested multiple triage models before landing on the solution, reducing cognitive load without disrupting familiar workflows. The Inbox went from a chat tool to a revenue system.
Briley Lowery
1min
Hey! I saw your sale ends soon...
92% likely to convert
AI Scoring
Filters need intent. A score doesn't. The right conversations surface on their own — no behaviour change required.
🔥 High CR
3 people
🚀 Engaged customers
5 people
0
Spam
Empty
Smart Folders
More views meant more scanning. I collapsed the inbox into three states, not to organise, but to reduce what agents need to hold in their head.
Suggested Replies
Speed without losing voice. Agents stay in control, the suggestion does the heavy lifting.
07 . What Success Looked Like
The Inbox Became a Company-Wide Bet
~22%
Faster time to first high-intent interaction vs control group
8%
Improvement in weekly retention among high-volume sellers
74%
Of sessions started in folders view, immediate triage adoption
Beyond the product metrics, the work led to something bigger: Inbox was established as a company-wide OKR
Neu Antoli
Manychat
Redesigning Inbox:
Signal Over Surface
This case study shows how I reframed Manychat's Inbox as a prioritisation system, and the design decisions that reduced churn among the users who mattered most.
01 . Context and Problem
The Business Had a Signal Problem
The business had a retention crisis. Users had a prioritisation crisis. The team had neither framed clearly.
The retention gap
High-value sellers were churning at 15% — a segment driving ~60% of ARR. The team had a backlog of reactive fixes but no strategy anchored to the root cause.
Business goal
Grow Inbox activation and reduce churn among high-value users by 2026.
What users experienced
Sellers managing 100+ messages daily had no way to identify which conversations mattered. Every message looked the same, so high-intent buyers kept going cold.
Users goal
Spend less time scanning. Respond to the right conversations before they go cold.
02 . My role
My role beyond the pixels
Lead designer on Inbox end-to-end. I defined the problem framing after reanalyzing existing data, ran 8 moderated user interviews across creators, SMBs and local businesses, and set the design principles that shaped every decision. I presented the vision to leadership, aligned the PM and engineering leads around a phased approach, and the work became a company-wide north star for AI adoption.
03 . The data
Numers never lie.
Before running a single interview, the data already told a clear story. This wasn't a UX problem, it was a revenue problem.
15%
Churn rate among high-value sellers driving ~60% of ARR
5 / 8
Interview participants who had considered cancelling Pro for WhatsApp Business native
100+
Messages per day managed with no priority signal — entirely by manual scanning
04 . Discovery
Talking to users changed everything.
I ran 8 moderated interviews across creators, infopreneurs, affiliates and local businesses. Every question was shaped by a decision we needed to make.
What we learned
- Intent was invisible: 3–5 threads opened before finding one worth replying to
- Delay kills conversion: late replies were the most common reason for missed sales
- Automation trusted for routing, not relationships
- Simpler always won: even if less powerful
05 . The Design
Making Priority Visible.
I explored and tested multiple triage models before landing on the solution, reducing cognitive load without disrupting familiar workflows. The Inbox went from a chat tool to a revenue system.
Briley Lowery
1min
Hey! I saw your sale ends soon...
92% likely to convert
AI Scoring
Filters need intent. A score doesn't. The right conversations surface on their own — no behaviour change required.
🔥 High CR
3 people
🚀 Engaged customers
5 people
0
Spam
Empty
Smart Folders
More views meant more scanning. I collapsed the inbox into three states, not to organise, but to reduce what agents need to hold in their head.
Suggested Replies
Speed without losing voice. Agents stay in control, the suggestion does the heavy lifting.
06 . The system
Built to Scale Across Every Surface
Three features that work as one, triage, intent, and response speed operating together across mobile and desktop.
07 . What Success Looked Like
The Inbox Became a Company-Wide Bet
~22%
Faster time to first high-intent interaction vs control group
8%
Improvement in weekly retention among high-volume sellers
74%
Of sessions started in folders view, immediate triage adoption
Beyond the product metrics, the work led to something bigger: Inbox was established as a company-wide OKR
Neu Antoli