Growth Design Case Study · 2026

AI Workflow
Automations

A project to help users start Jotform Workflows with ready-made use cases and integrations.

Elif Özdemir, UI/UX DesignerAcquisition Hackers

Goal: let users start from ready-made use cases

Before
jotform.com/ai/workflow-builderBefore
jotform.com/ai/workflow-builder, full page

One general builder page for every task.

Goal

A landing page for each use case and integration. The user finds a ready-made automation, clicks the CTA and starts building.

The problem: users had little to start from

Workflows had only one general builder page. A user who wanted to automate a task, for example “text me when a new email arrives”, landed there and saw a prompt field.

There were only a few example prompts, and they might not match what the user needed. So most users had to figure out the workflow alone, and automation looked hard.

Challenge

There are hundreds of use case and integration pairs. A page for each one can't be designed by hand.

How competitors present an integration page

n8n.io/integrations/clicksend-smsCompetitor
n8n ClickSend SMS integration page, full page

n8n: long and technical. It doesn't show the workflow working.

make.com/templates/3132-send-an-sms…Competitor
make.com template page, full page

Make: clean but static. Two app icons, no steps in between.

What I took from this: show the workflow working, for the user's own task.
No competitor had a one-to-one match for what we built: an animated, use-case-specific hero for each integration.

If users see the workflow working…

…they will understand faster what it does, and they will start building.

And if AI can make these visuals (hero animations, thumbnails and other page visuals) for each use case, I can cover hundreds of pages without designing each one by hand.

I designed 3 Lottie hero animations for every use case

Each one shows a workflow in a different way and can fit any use case. I tested them with 7 participants in a think-aloud test.

7

participants · think-aloud test
2 use AI tools often · 3 a few times · 2 never

What I asked
1What does this automation do?
2Can you see where AI comes in?
3What else do you think it can do?

What participants said about the 3 heroes

5/7

picked Variation 1 as the clearest

4/7

found Variation 1 the most AI-powered

6.4/7

think the product can handle many automations

6.1/7

feel they could build another workflow alone

What worked
“It shows exactly what is going to happen.”on Variation 1
“Variation three helps me understand there are many other things it can do.”on Variation 3
What didn't
“Variation two is very simplistic.”on Variation 2
“I would like to see the connections between the apps.”on Variation 3

After the test: the final hero direction

Variation 1 gave the clearest direction. The plan was to use one animation on all landings, so the trigger and action got general names.

An animation skill for use-case heroes

Then I saw that AI could turn the trigger and action parts into a skill. So I shaped them for each use case, and a Codex skill now creates a Lottie hero for any use case.

skill.md · excerpt
skill:   use-case-hero-animation
built with: Codex
input:   use case (apps + event + details)
rules:   workflow node · glow #8A38F5 → #0075E3
         use-case details on nodes · verb from action
output:  hero-animation.lottie.json
email-to-sms-workflow.lottie.jsonSkill output

Integration hero: a simple 2‑step workflow

When the prompt field was removed, the integration hero needed a new design. My solution was a simple 2‑step workflow. Small animations show each step clearly.

1
0.25s

Trigger starts

A badge appears

1.0s

Data moves

The line glows

Sending
1.6s

Action runs

It shows what it does

Sent
6.2s

Done

Then it starts again

Thank you, Burak Şahin

for his help with the small animations.

A skill that makes a thumbnail for every use case

Every use case needs a thumbnail in the directory where it is listed. Making each one by hand would take too much time. So I set the visual style first. Then Codex creates a thumbnail for each use case. For the Podo images, I connected the Magnific MCP to Codex. I can reuse this skill later on features pages.

skill.md · excerpt
skill:   use-case-thumbnail
built with: Codex + Magnific MCP
input:   use case (trigger app → action app)
rules:   Jotform visual style · app icons · brand colors
         one clear trigger → action story
output:  thumbnail.png
Thumbnail made by the skill

Skill output: a thumbnail

Thumbnails in the AI Workflow Automation Tools directory

In the directory: jotform.com/ai/workflows/tools

Benefits: an animated visual for each use case

Each landing has four Benefits cards, and each card has its own animation for that use case. Example: Gmail to SMS. Hover a card to play it.

New email alerts

Sender-based routing

Message details

Follow-up tracking

Where it sits on the landing page
Thank you, Burak Şahin

Burak created these animations with Claude and turned them into a skill. The skill made the Benefits visuals for all landings.

How it works: one layout, use-case icons

The steps and the layout stay the same. Only the app icons and names change for each use case.

Where it sits on the landing page
How it works, Google Calendar to Slack
How it works, Gmail to SMS
How it works, Asana to Google Calendar

Discover other integrations without leaving the page

This section lets users discover other integrations for the same apps. If they find one they want, Use Automation takes them straight to the builder, without opening its page.

In the reviewed sessions, some users found other automations this way.

Where it sits on the landing page
Discover other automations section on the Gmail to SMS landing

A wizard: set up the workflow on the landing

Google Calendar to Slack landing with the wizard

Before, the CTA opened the Workflow Builder. It is powerful, but there are many choices on the first visit.

With the wizard, the user connects both accounts and sets up the trigger and action on the same page.

Seen in real sessions
Gmail → Google Sheets · wizard: the user reached Enable Automation
Excel → Google Sheets · wizard: the user reached Enable Automation
Gmail → SMS · the user finished setup and found Test Run

Real user sessions, reviewed as a team

We watched JotStory recordings of real users as a team. Design issues and what we could fix, we fixed. The rest went to the related teams.

6,586sessions reviewed
369with written notes
6reviewers
Nov 12latest JotStory report

On mobile, the animation was below the fold

Users saw the CTA but not the animation. They didn't understand what the page offered, so they didn't click.

→

Design fix

Mobile hero redesigned: the animation and the CTA are both visible above the fold.

Fixed

Users try to click the hero

On desktop and mobile, many users clicked the workflow nodes again and again before they found the CTA.

→

Design note

The animated hero stays, as requested. This issue is still open.

Open

Users explore other use cases

Some users moved to other landings through the breadcrumb and Discover other automations, or watched the how-to video.

→

Design note

The page structure supports discovery. It stays the same on all landings.

Working

Prompt modal and sign-in in the builder

The prompt modal made users leave. Sign-up modals repeated, Excel and Sheets sign-in failed, SMS codes didn't arrive.

→

Action

Prompt removed from the builder. The rest shared with the related teams.

In progress

Loading problems and bugs

Benefits not loading, a JS conflict, a missing template, a slow builder (25–40 s), a stuck button in the wizard.

→

Action

Benefits, JS conflict and missing template fixed. The rest shared with the teams.

Mostly fixed

A/B/C test: three ways to start

Still running · no clear winner yet

Which one gets more workflows? A: integration landing · B: wizard · C: template page. Running on Microsoft Excel to Google Sheets, the page with the most traffic.

A: Integration landing
AIntegration landing2,666 users
6.4%Control

170 of 2,666 users created a workflow

B: Wizard on the landing
BWizard on the landing2,665 users
5.1%−21% vs. A

135 of 2,665 users created a workflow

C: Template page
CTemplate page2,666 users
5.5%−14% vs. A

146 of 2,666 users created a workflow

7,997 users so far · Early data, the gaps are not final.
Across all landings: 5 integration landings run this A/B/C test. The other 90 run an A/B test: landing vs. template.

How I measure success

User impact

Do users start a workflow?

CTA click rate · workflows created · workflows turned on

Discovery

Do users see more use cases?

How far they scroll · clicks to other use case pages

Design efficiency

Does it save design time?

Time per landing · pages per month

Results so far

Design efficiency
A few days → a few seconds

Time to make the visuals for one landing: by hand vs. with the skills

95

integration landings

made with the skills

25

use case landings

made with the skills

A/B/C test

Users who created a workflow

AIntegration landing
6.4%
BWizard
5.1%
CTemplate page
5.5%

No clear winner yetThe test is still running.

A/B/C numbers: 7,997 users so far on the Microsoft Excel to Google Sheets page.

From visit to sign-up

How users moved through the landings from Aug 22 to Nov 10, 2026.

39,289
Sessions
38,492 users · 46,100 page views
→63.9%
25,100
Engaged sessions
63.9% engagement rate
→5.9%
1,473
Workflow clones
clone approvals from the landings
→45.9%
676
Sign-ups
639 social · 37 standard
3.7%

of sessions led to a workflow clone

1.7%

of sessions led to a sign-up

57s

average session duration

GA4, unsampled, 81 days. Steps are event and session counts, so the rates between steps are approximate.

What worked

What worked

Clarity first

The clear trigger → action hierarchy won the user test. 5 of 7 participants picked it as the clearest.

A skill-based design system

All 120 live landings were built with the skills. Each section follows its own rules, from hero motion to use-case-specific Benefits interfaces.

Consistent section structure

Hero and How it works share one structure across landings. Only icons and content change, so users see a familiar pattern on every page.

Adapting to a changing scope

In 6 months, one landing grew into four entry points. The system was flexible enough to support prompt, integration, template and wizard.

Small tests, clear decisions

Seven think-aloud sessions were enough to choose a direction for the hero.

Watching real sessions

Session reviews showed where users got stuck, so the team could find the issues and fix them.

What didn't

What didn't work

The animation looked clickable

Users clicked the workflow nodes again and again. The hero reads as UI, not as a demo. This is still open.

Key content below the fold on mobile

The animation was below the fold, so users saw the CTA without understanding the page. I redesigned it.

A generic prompt field

A few example prompts did not match what users needed. On integration landings, a clear 2-step workflow worked better.

The handoff to the builder

After the CTA, many users met modals and sign-in steps in the builder. This part was outside the landing, but it affected the results.

What I'd do differently

What I'd do differently

Design mobile and desktop together

Not one after the other, so the whole design works on both.

Set motion rules at the start

What the animation explains, where it points and what is clickable.

Lock the system before scaling

Agree on the rules before generating many pages.

Plan the full journey

Stay in closer contact with the product teams, so the landing and the builder steps work as one journey.

Test the final design again

Check the iterated hero with users before using it on every page.

What the team can reuse

01

Build a skill for repeated design work

When a visual repeats on many pages, write it once as a skill: style rules, good examples and a clear input. Anyone can then generate visuals that follow the design language. The heroes, thumbnails, Benefits and How it works visuals were made this way.

02

Share skills, then keep curating them

A skill one designer builds can serve the whole team. The Thumbnail skill made the visuals for all landings in this project, and it can be adapted for features pages next. When the output is off, fix the rule, not each file.

03

Test visuals on every device

Check animations and visuals on desktop, tablet and mobile before going live. What works on one screen can hide the message on another.

04

Real sessions can show what we miss

Session recordings let us check our design assumptions with real users. They can reveal paths we didn't expect, and some of them may turn into the next design improvements.

Hero animation skillThumbnail skillBenefits skillready to share with the team

The final landing: Gmail to SMS

Full page · scroll to explore, hover the Benefits cards
jotform.com/ai/workflows/automations/gmail-to-smsLive
Gmail to SMS landing, full page1

Thank you

Elif Özdemir, UI/UX DesignerAcquisition Hackers