All work
Tailwind · B2B SaaS · 0-to-1

Tailwind Copilot

A marketing assistant that tells small business owners what to post and when, so growing their business stops feeling like guesswork.

Tailwind Copilot — the Hub home screen on a tablet
RoleSole Designer & Product Owner
TeamDevs, PMs, Executive Board
ScopeResearch, strategy, design, testing
Timeline0-to-1 in 8 months
+73%surge in user engagement
+30%conversion boost
+6%increase in market share
25%faster product development cycle
01Context

A copilot that turns "I don't know what to post" into a daily plan

Tailwind is a social media tool for small business owners who don't have a marketing team. It handles the basics, designing, scheduling, and publishing across channels, plus a hashtag finder, a shoppable bio link, and post-time suggestions.

Copilot was a new 0-to-1 product built inside it: an assistant that hands each owner a marketing plan tailored to their business, so they always know the next move.

Tailwind logo
02Problem

Small businesses have the passion. They don't have the marketing team.

Big companies hire people to figure out what sells. Small owners can't. They have the product and the drive, but not the know-how to get it in front of the right people.

"You're a mom with two toddlers trying to grow a homemade candle business while running the house alone."

That's who we designed for. For her, marketing is one more thing that never makes it to the top of the list.

Overwhelmed small business owner
  • No idea what to post
  • No time to design or schedule it
  • Needs results now, with no clear sense of what actually works
  • No plan, just guessing
03Goals

What we were aiming for

For users

Help people with zero marketing background grow their business, using a plan specific to them with steps they can actually follow.

For the business

Make Tailwind easier to discover, try, and adopt, and open it up to owners who are just starting to think about marketing at all.

04Solution

Tell people what to post, and when

Research turned up a gap nobody was filling: no tool told owners the specific thing to post on a specific day. Copilot does that. It builds a personalized plan that takes the guesswork out, so owners can spend their time making a good product instead of second-guessing their feed.

05How it works

From survey to published post

01

Personalize

A short survey asks about the owner's business and how much marketing they've done before, then builds a plan around the answers. An overview walks them through how it maps to their goals.

Copilot personalized plan overview
The onboarding survey becomes a plan with clear next steps.
02

Browse

The plan breaks down into daily tasks: a deck of prompt cards, each with a post idea, a short why, visuals, and a CTA. Scheduled prompts drop into a calendar that lays out the week and makes it easy to keep going.

The Hub home screen with the weekly posting schedule
The calendar turns chosen ideas into a posting schedule.
03

Execute

From a prompt, the owner can jump into design or go straight to publishing, with a mini prompt card guiding them. Once something's scheduled, a quick recap surfaces the next task so momentum doesn't stall.

The guided flow from prompt card through design to a scheduled post, ending on a success recap and progress tracker
From prompt card to scheduled post, ending on a success recap and progress tracker.
04

Track & adjust

A plan timeline lets owners see progress and change course once they've tried a few things. The plan stays theirs instead of a fixed script.

Plan customization modal where owners adjust how often each campaign type is recommended
Owners tune how often each campaign shows up, so the plan stays theirs.
06Design process

From planning to production ramp-up

Stage-gate process diagram: planning, concept development, system-level design, a design-test-build loop, and production ramp-up
The delivery cadence: planning through production ramp-up, with a tight design–test–build loop in the middle. Markers show the stages I was hands-on.

Macro flow

Macro user flow: personalization survey, marketing plan, assigned tasks, create and schedule, with progress tracking closing the loop
The macro flow: survey → plan → tasks → create and schedule, with progress tracking closing the habit loop. Each step is annotated with the goal it serves.

Early iterations

Copilot early UI iterations
Early on I explored a few directions, different levels of progressive disclosure, some light gamification, to figure out what actually helped people finish their tasks instead of just decorating them.
07Design challenges

Three problems worth getting right

Challenge 01

Two very different users

Interviews showed two patterns: people who want to do the one thing for today and be done, and people who want to run ahead and knock out a week at once. We'd built only for the first, one time-sensitive prompt at a time, to keep things calm. To serve the second without wrecking that calm, I added a way to work through several tasks in a sitting: a card deck for focus, an interactive timeline for the people who wanted to steer.

Card deck progressive disclosure design

Case 1: The card deck surfaces the single most important task and keeps everything else out of the way.

Interactive timeline with recap and momentum view

Case 2: The interactive timeline lets users browse ahead and pick what they want to work on.

Challenge 02

Making it feel like one product, not a bolt-on

We shipped Copilot first as a standalone feature, partly to see if anyone wanted it. That isolated version taught us a lot: how people actually used it, how fast they adopted it, where the friction was. Once we knew it had legs, I folded the best parts into the core homepage Hub, so every login opens on the most important task and a clear read on the week ahead.

Copilot standalone MVP action view zero state

The standalone MVP let us learn adoption and friction before committing.

Copilot integrated into the Hub results dashboard

Those learnings shaped how Copilot got built into the Hub.

Challenge 03

Showing just enough, not everything

A full marketing plan, funnel, campaigns, dozens of prompts, is a lot for someone doing this for the first time. Our early version listed prompts as plain titles, and people couldn't tell them apart or pick one; too many choices, not enough to go on. I switched to a card deck that shows one prompt at a time with just enough to make the call, and pushed the rest into a side panel that opens only when a card is active.

Early iteration showing multiple prompt titles

Early on, prompts were bare titles. Users couldn't tell them apart, and the choices piled up.

Card deck interface showing one prompt at a time

The card deck shows one prompt at a time with only what's needed to decide: continue, or move on.

08Results

Measured on a real rollout

We ran it as a two-week rollout to two-thirds of new sign-ups. Owners who got a personalized plan were 73% more likely to schedule their first post, and 30% more likely to upgrade.

A/B cohort activation funnels comparing owners who chose a marketing plan against those who skipped it
The cohort funnels behind the numbers: sign-ups with a plan scheduled their first post at roughly 1.6× the rate of those without.
09Next steps

Built to grow

The current design covers organic social (Facebook, Pinterest, Instagram) and email. The underlying system is set up to extend into paid social and other channels as Tailwind grows into them.

Next project
AI Customer Profile Prediction