James Martins Byte

BYTE BUSINESS
BYTE BUSINESS

EXECUTIVE SUMMARY

I spearheaded the end-to-end product redesign of Byte, transitioning it from a fragmented tracking tool into a unified financial co-pilot. Our design decisions unlocked 300& Month-over-Month (Mom) growth by capturing underserved SME segments across Africa.

METADATA

My Role:

Team:

Timeline:

Lead Product Designer (end-to-end UX/UI, plus a good chunk of the research)

1 PM, 3 Engineers, 1 UX Researcher, Me

Sep 2020 – 2023, continuous iteration

BYTE BUSINESS
BYTE BUSINESS

EXECUTIVE SUMMARY

METADATA

My Role:

Team:

Timeline:

Lead Product Designer (end-to-end UX/UI, plus a good chunk of the research)

1 PM, 3 Engineers, 1 UX Researcher, Me

Sep 2020 – 2023, continuous iteration

BYTE BUSINESS
BYTE BUSINESS

EXECUTIVE SUMMARY

METADATA

My Role:

Team:

Timeline:

Lead Product Designer (end-to-end UX/UI, plus a good chunk of the research)

1 PM, 3 Engineers, 1 UX Researcher, Me

Sep 2020 – 2023, continuous iteration

SCROLL

The Project in 3 Stats (My Real Business Impact)

10x

Onboarding leap

5 to 50% Conversion

300%

Month over Month

Growth post-launch

90% +

AI Auto-Category

accuracy on launch

The Problem: Infrastrucutre not literacy
The assumption we walked in with that merchants needed educating on their finances turned out to be wrong. What they actually lacked was one place to see it all. A typical day for a merchant looked like: cash in a drawer, a mobile money app for transfers, WhatsApp for supplier invoices, and a notebook (or nothing) tying it together.

Too many front doors.

Merchants were bouncing between 3–6 separate apps daily just to piece together what came in and what went out.

Truth, delayed.

By the time cash and transfers got manually reconciled, the numbers were already stale useful for looking back, useless for deciding right now.

A business that lives in two places.

Sales happened in person, in cash or by card, but the record of them only existed once someone typed it in later if they got around to it.

"I need my business money separate from my personal money so I can actually tell if I'm profitable."

Fola Adebayo

Fashion Brand Owner

"I need my business money separate from my personal money so I can actually tell if I'm profitable."

Fola Adebayo

Fashion Brand Owner

"I need my business money separate from my personal money so I can actually tell if I'm profitable."

Fola Adebayo

Fashion Brand Owner

Three (3) Decisions that mattered

Decision 1:
Delaying trust instead of front-loading it

Before

After

The Friction:

Our first onboarding flow asked for bank-account linking before users saw any value a reasonable assumption on paper, a 95% drop-off in practice.

The Design Solution:

We flipped the order. Let people explore the dashboard and log one transaction manually first, and push bank-linking to later, once they'd gotten something out of the app.

The Result:

Conversion went from 5% to 50% within two months. The lesson wasn't "reduce friction" in the abstract it was where the friction sits that matters more than how much of it there is.

Decision 2:
Making the AI feel like a coach, not a compliance officer

Perform Daily Business Task ] ──> [ Open App ] ──> [ Select Module ] ──> [ Fill Multi-Field Form ] 🔴 HIGH FRICTION POINT

The Problem: The Manual Data Entry Trap

Manual categorization is the kind of task that feels reasonable to ask for once and unreasonable to ask for every day forcing it kills daily use.

The Reality of Market Merchants:

A busy vendor in Lagos doesn't have time to sit down and fill out forms to track sales, manually tally inventory, or send individual WhatsApp reminders to customers who owe them money.

The Risk:

If logging sales, inventory, and invoices requires heavy manual effort, the records stay incomplete. Incomplete data destroys the accuracy of business analytics and debt recovery.

[ POS Sale / Automated Sync ] ──> [ AI Auto-Updates Inventory & Sales ] ──> [ Proactive Revenue Action Prompt ] 🟢 ZERO MANUAL FORM-FILLING

The Redesign: AI as an Ecosystem Co-Pilot

Instead of treating each tool (Invoicing, Inventory, Sales) as a separate silo with heavy forms, we used AI to connect the dots across the entire suite automatically.

We turned Byte into a active co-pilot that automates data entry, predicts inventory shortages, and actively recovers money owed to the merchant.

Instead of treating each tool (Invoicing, Inventory, Sales) as a separate silo with heavy forms, we used AI to connect the dots across the entire suite automatically.

We turned Byte into a active co-pilot that automates data entry, predicts inventory shortages, and actively recovers money owed to the merchant.

Automated Sales & Inventory Syncing:

When a sale is made (via POS or transfer), AI automatically deducts the items from Inventory, logs the revenue under Sales Records, and updates Real-time Analytics with 90%+ accuracy on launch. Zero manual tallying required.

Smart Debt Recoverry & Invoicing:

Instead of a merchant manually tracking unpaid debts, the AI monitors outstanding invoices and prompts the merchant with one-tap recovery actions:

💡 "Fola still owes ₦45,000 for order #108 (3 days overdue). Tap here to send a polite WhatsApp payment link reminder."

Decision 3:
Bringing the offline half of the business online

The Friction:

Desktop and mobile apps are great, but most of our merchants' actual sales happened in person cash and card, on-site while the app only ever saw what got logged after the fact.

The Design Solution:

I co-led the interface design for a physical POS terminal that syncs with the mobile app in real time, closing the gap between offline sales and digital records.

The Result:

t's a less common thing for a product designer to have shipped, and it directly moved the needle on retention merchants stopped treating Byte as a side chore and started treating it as the record of their business.

What didn't work? Early on, I noticed merchants were logging transactions but rarely opening analytics so I assumed they needed more insight, and designed a full visual dashboard with weekly margin graphs, cash-flow projections, and expense breakdowns.

On-site usability testing killed it fast: merchants ignored the graphs entirely. One put it bluntly: "I don't have time to study stock market lines while handling three customers." The density wasn't clarity, it was analysis paralysis they wanted an answer, not a chart to interpret. I scrapped the graphs and replaced them with the conversational notification cards that became the "Business Coach" model the same shift behind the delivery-cost example above.

Gallery Showcase

What I (We) Learned & What’s Next?

Data Driven Design Works

Utilizing behavioral analytics, A/B testing, and direct user user interviews allowed us to to continously iterate based on what users did, not just what they said.

Trust is Fragile

Micro-copy changes (like changing “Connect Bank” to “Securely Read Statements”) significantly improved user confidence during syncing.

Data Driven Design Works

Utilizing behavioral analytics, A/B testing, and direct user user interviews allowed us to to continously iterate based on what users did, not just what they said.

Trust is Fragile

Micro-copy changes (like changing “Connect Bank” to “Securely Read Statements”) significantly improved user confidence during syncing.

Data Driven Design Works

Utilizing behavioral analytics, A/B testing, and direct user user interviews allowed us to to continously iterate based on what users did, not just what they said.

Trust is Fragile

Micro-copy changes (like changing “Connect Bank” to “Securely Read Statements”) significantly improved user confidence during syncing.

Ready to ship? Let’s talk numbers and systems.

TO GET IN TOUCH

Contact Me

BOOK A CALL

SET UP A DAY TO TALK ABOUT YOUR DESIGN NEEDS

CLICK TO COPY

CREATED BY

JAMES MARTINs

JAMES MARTINS

2026

ALL BUILT IN

FRAMER

10:32 PM

August 7

Abuja, Nigeria

Available for work

Check X

Instagram

Behance

Dribbble

The Project in 3 Stats (My Real Business Impact)

The Project in 3 Stats (My Real Business Impact)

The Project in 3 Stats (My Real Business Impact)

Too many front doors.

Too many front doors.

Merchants were bouncing between 3–6 separate apps daily just to piece together what came in and what went out.

Merchants were bouncing between 3–6 separate apps daily just to piece together what came in and what went out.

Truth, delayed.

Truth, delayed.

By the time cash and transfers got manually reconciled, the numbers were already stale useful for looking back, useless for deciding right now.

By the time cash and transfers got manually reconciled, the numbers were already stale useful for looking back, useless for deciding right now.

A business that lives in two places.

A business that lives in two places.

Sales happened in person, in cash or by card, but the record of them only existed once someone typed it in later if they got around to it.

Sales happened in person, in cash or by card, but the record of them only existed once someone typed it in later if they got around to it.

Three (3) Decisions that mattered

Three (3) Decisions that mattered

Three (3) Decisions that mattered

Decision 1: Delaying trust instead of front-loading it

Decision 1: Delaying trust instead of front-loading it

Decision 1: Delaying trust instead of front-loading it

The Friction:

The Friction:

Our first onboarding flow asked for bank-account linking before users saw any value a reasonable assumption on paper, a 95% drop-off in practice.

Our first onboarding flow asked for bank-account linking before users saw any value a reasonable assumption on paper, a 95% drop-off in practice.

The Design Solution:

The Design Solution:

We flipped the order. Let people explore the dashboard and log one transaction manually first, and push bank-linking to later, once they'd gotten something out of the app.

We flipped the order. Let people explore the dashboard and log one transaction manually first, and push bank-linking to later, once they'd gotten something out of the app.

The Result:

The Result:

Conversion went from 5% to 50% within two months. The lesson wasn't "reduce friction" in the abstract it was where the friction sits that matters more than how much of it there is.

Conversion went from 5% to 50% within two months. The lesson wasn't "reduce friction" in the abstract it was where the friction sits that matters more than how much of it there is.

Decision 2: Making the AI feel like a coach, not a compliance officer

Decision 2: Making the AI feel like a coach, not a compliance officer

Decision 2: Making the AI feel like a coach, not a compliance officer

The Problem: The Manual Data Entry Trap

The Problem: The Manual Data Entry Trap

Manual categorization is the kind of task that feels reasonable to ask for once and unreasonable to ask for every day forcing it kills daily use.

Manual categorization is the kind of task that feels reasonable to ask for once and unreasonable to ask for every day forcing it kills daily use.

The Reality of Market Merchants:

The Reality of Market Merchants:

A busy vendor in Lagos doesn't have time to sit down and fill out forms to track sales, manually tally inventory, or send individual WhatsApp reminders to customers who owe them money.

A busy vendor in Lagos doesn't have time to sit down and fill out forms to track sales, manually tally inventory, or send individual WhatsApp reminders to customers who owe them money.

The Risk:

The Risk:

If logging sales, inventory, and invoices requires heavy manual effort, the records stay incomplete. Incomplete data destroys the accuracy of business analytics and debt recovery.

If logging sales, inventory, and invoices requires heavy manual effort, the records stay incomplete. Incomplete data destroys the accuracy of business analytics and debt recovery.

The Redesign: AI as an Ecosystem Co-Pilot

The Redesign: AI as an Ecosystem Co-Pilot

Instead of treating each tool (Invoicing, Inventory, Sales) as a separate silo with heavy forms, we used AI to connect the dots across the entire suite automatically.

We turned Byte into a active co-pilot that automates data entry, predicts inventory shortages, and actively recovers money owed to the merchant.

Instead of treating each tool (Invoicing, Inventory, Sales) as a separate silo with heavy forms, we used AI to connect the dots across the entire suite automatically.

We turned Byte into a active co-pilot that automates data entry, predicts inventory shortages, and actively recovers money owed to the merchant.

Automated Sales & Inventory Syncing:

Automated Sales & Inventory Syncing:

When a sale is made (via POS or transfer), AI automatically deducts the items from Inventory, logs the revenue under Sales Records, and updates Real-time Analytics with 90%+ accuracy on launch. Zero manual tallying required.

When a sale is made (via POS or transfer), AI automatically deducts the items from Inventory, logs the revenue under Sales Records, and updates Real-time Analytics with 90%+ accuracy on launch. Zero manual tallying required.

Smart Debt Recoverry & Invoicing:

Smart Debt Recoverry & Invoicing:

Instead of a merchant manually tracking unpaid debts, the AI monitors outstanding invoices and prompts the merchant with one-tap recovery actions:

Instead of a merchant manually tracking unpaid debts, the AI monitors outstanding invoices and prompts the merchant with one-tap recovery actions:

💡 "Fola still owes ₦45,000 for order #108 (3 days overdue). Tap here to send a polite WhatsApp payment link reminder."

Decision 3: Bringing the offline half of the business online

Decision 3: Bringing the offline half of the business online

Decision 3: Bringing the offline half of the business online

The Friction:

The Friction:

Desktop and mobile apps are great, but most of our merchants' actual sales happened in person cash and card, on-site while the app only ever saw what got logged after the fact.

Desktop and mobile apps are great, but most of our merchants' actual sales happened in person cash and card, on-site while the app only ever saw what got logged after the fact.

The Design Solution:

The Design Solution:

I co-led the interface design for a physical POS terminal that syncs with the mobile app in real time, closing the gap between offline sales and digital records.

I co-led the interface design for a physical POS terminal that syncs with the mobile app in real time, closing the gap between offline sales and digital records.

The Result:

The Result:

t's a less common thing for a product designer to have shipped, and it directly moved the needle on retention merchants stopped treating Byte as a side chore and started treating it as the record of their business.

t's a less common thing for a product designer to have shipped, and it directly moved the needle on retention merchants stopped treating Byte as a side chore and started treating it as the record of their business.

Gallery Showcase

Gallery Showcase

Decision 1:
Delaying trust instead of front-loading it

Decision 1:
Delaying trust instead of front-loading it

Decision 1:
Delaying trust instead of front-loading it

Decision 2:
Making the AI feel like a coach, not a compliance officer

Decision 2:
Making the AI feel like a coach, not a compliance officer

Decision 2:
Making the AI feel like a coach, not a compliance officer

Decision 3:
Bringing the offline half of the business online

Decision 3:
Bringing the offline half of the business online

Decision 3:
Bringing the offline half of the business online