1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

BACK

Overview
Problem
Solution
Impact
Fast Alignment
Design Decisions
Systems Considerations
Beta Insights
Learnings

ACTIVATION / GORGIAS • 2024

Onboarding Redesign

This project highlights

ROLE
Product Designer

TIMELINE
Nov 2023 - Feb 2024

COLLABORATORS
Lenaig L., Product Manager
2 Engineers

CONTEXT

Intro

Current onboarding

THE PROBLEM

In practice, however, its impact is limited by an engagement problem. Only 5% of all chat interactions are funneled through the assistant.

And without engagement, even the smartest recommendations and checkout nudges can't influence purchasing decisions.

Chat is for problems, not shopping

Most shoppers still only associate chat as a problem-solving channel for issues like returns, tracking, and complaints.


Without a clear need, they ignore the widget.

Chat is for problems, not shopping

Most shoppers still only associate chat as a problem-solving channel for issues like returns, tracking, and complaints.


Without a clear need, they ignore the widget.

It's hidden and passive

Shopping Assistant is tucked away in the chat widget, competing with product-related content that's far more visually engaging and relevant.

Starting a conversation takes effort

Shoppers must notice the widget, decide to engage, formulate a question, and trust they'll get a helpful answer.


For those just browsing, it's too much friction.

Starting a conversation takes effort

Shoppers must notice the widget, decide to engage, formulate a question, and trust they'll get a helpful answer.


For those just browsing, it's too much friction.

NARROWED SCOPE

We started ambitiously to tackle both onboarding and content creation in a single because they're so intertwined for adoption.

We recognized scope bloat and narrowed the scope into phased projects so we could ship a focused ‘launch‑ready’ experience faster, lower delivery risk.

Phase 1: Essential setup

Get merchants almost live with minimal effort. The goal is to be ready to flip the switch, not to be perfect.

Later: Content expansion

Mature the knowledge base over time—add core FAQs, address top contact drivers, organize categories, and iterate continuously to make self-service truly effective.

THE SOLUTION: PHASE 1

A 3-step onboarding wizard that provides fast time-to-value

We transformed a perceived time-sink into clear, guided steps that show launch-ready setup can takes minutes, not days.

Core configuration

Basic setup essentials like subdomain and store connection, prefilled whenever possible to reduce friction.

Brand alignment

Logo, color scheme, and typography setup ensures the Help Center reflects merchant brand identity, which is critical for building customer trust.

Quick content creation

Editable templates for the most-searched topics let merchants populate their Help Center fast (later replaced with AI articles).

Unlock paid features

Merchants on paid plans get an additional step to turn on automation features, ensuring they leverage what they're paying for while building product awareness.

Nudge to publish

After setup, a contextual modal prompts merchants to either add more articles (create or import) or go live, depending on their current state.

RESULTS

This feature was released to all customers in February 2024.

15% increase

in Help Center adoption

5% increase

in merchants with >1 article

PROCESS IMPACT

Scaling research velocity through async testing

I used a new tool for async testing (Maze) to accelerate research on a tight timeline, cutting a typical 2–3 week interview cycle down to 1–2 weeks. Its success led to broader adoption across the product team.

After collecting interest from other designers, I wrote a proposal for senior leadership to secure a budget for a team subscription, and built documentation for the team to easily onboard to the tool.

LEARNINGS

Building AI features demand a different design mindset

The biggest uncertainties came from the scraping model, not the interface. Evaluating and shaping AI behavior became as central to the design work as designing the interface itself.

Chat complexity requires deep foundational understanding

Chat has countless configurations and edge cases. I invested in building on my foundational knowledge and worked closely with my PM and engineers to ensure we were covering all of the different use cases.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16