Brand and website copy for Zentra

Role
Product Designer, Co-op
Duration
May to August 2026
Tools
Figma, Framer
Live site
zentra.bot

Finding the words an AI financial platform for founders uses to explain itself, then building the brand, website, and design language around them.

  • Copy Design
  • UX Writing
  • Brand Identity
  • Marketing Website

Skip to solution

Zentra platform overview

Overview

An AI CFO for founders

An AI financial platform that turns a founder’s raw numbers into plain language insight on runway, burn rate, MRR, and cash flow. It acts as an AI CFO before a founder can afford a real one.

The product worked, but it had no brand

When Miydesign brought me onto Zentra, there wasn’t a brand yet. We had a handful of references and a product that hadn’t figured out how to describe itself.

The first job was the words: a voice that could make a fintech company sound less like a spreadsheet and more like a CFO in your pocket. The colours, type, and layout followed from that voice.

The problems the site had to solve

Founder path and partner path across the same site

Founders and partners shared one site

Founders came for speed and clarity. Partners came for incentive and credibility. Both had to be spoken to without splitting the site in two.

Competitor fintech pages, all reading the same way

Finance sites all sound the same

Cold, corporate, written for accountants. Zentra had to sound built for founders without losing the seriousness of the numbers underneath.

Two pages side by side, spacing and type not matching

Each new page started from scratch

No type scale, no colour tokens, no components, so every page drifted from the last.

My role

I designed six production pages and an early dashboard concept as part of the Miydesign team. The work ran from brand workshop through mobile implementation.

Research

I wrote the messaging before designing anything

I compared other brands’ homepage copy

FormalApproachable
Stripe
Ramp
Mercury
Zentra
Directional read from the benchmark, not measured data.
Company
How it sounds
What Zentra took
What Zentra avoided
Ramp
How it soundsPunchy and proof-heavy. “Time is money. Save both.” Customer counts and growth stats on every scroll.
What Zentra tookTwo-beat headlines. Pairing a claim with a number.
What Zentra avoidedController and finance-team vocabulary (procure-to-pay, multi-entity). Zentra’s reader is the founder doing the finance job.
Mercury
How it soundsWarm, premium, founder-first. “Radically different banking.”
What Zentra tookSpeaking to founders directly. Plain explanations of what each feature does for you.
What Zentra avoidedBig superlatives about being unlike anything before. A new product can’t back that up yet.
Linear
How it soundsSpare and declarative. Few adjectives, one idea per line.
What Zentra tookShort sentences. The confidence to state a position without hedging.
What Zentra avoidedCategory-coining lines (“A new species of product tool”) that need an established brand to land.
Stripe
How it soundsPrecise and infrastructural. Numbers do the persuading.
What Zentra tookEvery claim tied to a figure you can check. This fed directly into the accuracy checklist.
What Zentra avoidedDeveloper and enterprise depth. Zentra’s reader isn’t integrating an API.

What I found

The finance brands earn trust with proof: customer counts, uptime, processed volume. Linear earns it with restraint, short declarative lines and almost no adjectives.

None of them lead with the question a founder actually types, like “how long until I run out of cash?”. That left room for Zentra to open with the question itself.

What I did

I read the homepages and pricing pages of Ramp, Mercury, Linear, and Stripe and logged every headline, subhead, and CTA in one sheet. For each, I noted sentence length, how jargon was handled, and what proof backed each claim. Linear was a deliberate outlier: not a finance product, but the clearest benchmark for short, confident product copy.

What it changed

Zentra’s hero became a question instead of a claim, “Ask your finances anything.” Proof moved further down the page, and every number had to pass the accuracy checklist before it shipped.

Design

Four rules for how Zentra writes

Answer first

Lead with the number or the decision, then explain.

This
“You have 14 months of runway at your current burn.”
Not that
“Zentra leverages AI-powered analytics to deliver comprehensive runway insights.”

Use the words founders use

Write the question a founder would type, not the term an accountant would file it under.

This
“When will I run out of cash?”
Not that
“View your cash depletion forecast and liquidity projections.”

Keep the tone calm

Confidence comes from specifics. No exclamation marks, no “revolutionary.”

This
“Read-only, always. Zentra can’t move money or change your books.”
Not that
“Bank-grade, military-level security you can trust 100%!”

Only say what’s true

If we can’t verify it, we don’t publish it. If Zentra doesn’t know, it says so.

This
“If it can’t find the number, it says so instead of guessing.”
Not that
“Trusted by 10,000+ founders worldwide.”

Everything started from one positioning line

The brand sentence came first, and every visual choice answered to it. We tested four palette directions against it until one felt founder-facing without feeling unserious.

Why these colours

The founders came in with one firm request: purple. They wanted Zentra to feel trustworthy, and to them purple carried that. Most colour research ties trust to blue, so the job was finding a purple with blue’s steadiness that didn’t look like every bank’s navy.

Each brand in this space signals trust its own way: Stripe’s electric blurple, Mercury’s dark navy with a quiet indigo accent, Ramp’s single loud yellow. Stripe proved purple can work in finance, which made it the brand Zentra most needed to avoid resembling. Stripe uses blurple as an accent on links and buttons. Zentra uses a deeper violet as a full brand surface, in the hero and logo field.

Every text pairing below passes WCAG AA.

Aa
Primary#7C2FF0White text on primary, 5.96:1, AA
Aa
Accent#C49AF7Accent with deep purple #2D1A6E, 6.27:1, AA
Aa
Surface#F7F6FFPrimary text on surface, 5.56:1, AA
Aa
Text#26243EText on surface, 13.94:1, AAA

Colours taken from the live site. WCAG AA needs 4.5:1 for body text, AAA needs 7:1.

In context

Stripe#635BFFAccent on links and buttons
Mercury#5266EB on #15151EIndigo accent on dark navy
Ramp#FCFB70One loud colour
Zentra#7C2FF0Full brand surface

Approximate brand colours, sampled from each live site.

Each line answers one founder question

Hero headline

“Ask your finances anything.”
Its job
Explain what Zentra does by inviting the visitor to use it.
Why
A question invites someone to try it. A claim only asks them to believe it.

Subhead

“Freelancer, agency, consultant, small team — whoever you are, Zentra turns your numbers into plain-English answers. Just ask.”
Its job
Answer “is this for me?”
Why
It names the people who’d use it in their own words, and promises answers instead of dashboards.

Call to action

“Join waitlist”Free during beta · No credit card required
Its job
Remove the cost question at the moment of decision.
Why
“Free during beta” answers cost and sets the expectation that it’s early. The line under the button covers the card question.

Pricing descriptions

Launch
“Idea stage and MVP teams getting their first clear view of the numbers.”
Growth
“Startups in growth stage that need deeper planning and larger teams.”
Its job
Let someone pick a plan by their stage, not by counting features.
Why
Each description says who the plan is for, so there are only three stages to choose between instead of a feature grid to compare (Hick’s Law). The feature list underneath handles what’s in it.

The same voice across all six pages

Home, Pricing, Partner, Blog, Join Waitlist, and Log In. Each page has one job and a headline that states it, all drawn from a shared voice and component library.

Tabs fixed the broken mobile sections

A scroll-jacked platform section and a clipped pricing table both became tab patterns.

Platform section

On mobile, the scroll-jacked section became tabs, so every card is one tap away instead of hidden behind a long scroll.

Pricing

The three plans sit side by side on desktop, and under 768px they stack into one column while the comparison table becomes tabs, one plan at a time.

Targets

I checked the shipped copy against research

The hero was missing specifics

10 seconds

The window a page has to communicate its value before most visitors leave (Nielsen Norman Group)

3/4

Questions the shipped hero answered above the fold

Question
Where the shipped hero answers it
What is it?
Where the shipped hero answers it“Ask your finances anything.” and “Zentra turns your numbers into plain-English answers.”
Who is it for?
Where the shipped hero answers it“Freelancer, agency, consultant, small team”
What do I do next?
Where the shipped hero answers it“Join waitlist”, with “Free during beta · No credit card required” underneath
What does it cover?
Where the shipped hero answers itGap Not named. The hero never says which numbers Zentra tracks.

What I found

Nielsen Norman Group’s page-visit research puts the stay-or-leave decision in the first 10 seconds. The shipped hero answered what Zentra is, who it’s for, and what to do next without a scroll. It never named the metrics it covers, so a visitor had to keep reading to learn it tracks cash and runway.

What I did

Without founders available for a live test, I checked the hero against published research instead. Then I audited the version I shipped in August against the questions a five-second test asks: what is it, who is it for, what does it cover, and what do I do next.

What I’d change

The question-led headline and the named audience held up. The gap was specifics.

Since handoff, the live subhead names runway, burn, and MRR, the same fix this audit points to.

“Join waitlist” said exactly what would happen

[6/8]

Users in a Nielsen Norman Group study who clicked a generic “Get Started” expecting information and landed in a signup flow instead

Button copy
Sets the right expectation?
Notes
Join waitlist (shipped)
Sets the right expectation?Yes
NotesNames exactly what happens during a beta: you join the list, not the product.
Start free
Sets the right expectation?Not at handoff
NotesPromises immediate access the beta couldn’t give yet. It fits once there’s a free plan, which is why the live site uses it now.
Get started
Sets the right expectation?No
NotesReads as a path to more information, so people expecting details land in a form instead.
Learn more
Sets the right expectation?No
NotesToo generic to stand alone.
Ask your first question
Sets the right expectation?Partly
NotesPreviews the product, but the click joined the waitlist instead of opening a question box.

What I found

Nielsen Norman Group found that generic labels like “Get Started” set the wrong expectation: most users who clicked one expected information and landed in a signup flow. I scored the common alternatives against their guidance (specific, sincere, substantial, succinct). “Join waitlist” passed on the two that matter most for a beta, saying exactly what happens and being honest about access.

The pages and components that shipped

6

production pages shipped

24

components documented

8

weeks from brand workshop to handoff

The bar for the next version: 8 in 10 visitors can say what Zentra does, and what the button does, after a five-second look.

Reflection

Writing took most of the design time

Working live

Direct work with a founder and a small studio team on a sprint schedule, with decisions made in days rather than weeks.

Integrity

Catching fabricated claims taught me that accuracy is a process running through the whole project, not a final review.

What I’d do differently

Test the hero copy with real visitors before handoff, not after. The gap my audit found, no named metrics, is something a five-second test would have caught in May.

Visit live site

The site has continued to evolve since handoff in August 2026.

See next

All work →