--case-landbotWeb + Product · Integrations · 2021 – 2023
Landbot.
Integrations without a single webhook.
Landbot helps companies build chatbots without code. I joined to lead its website migration from WordPress to Webflow, protecting the blog that drove its organic traffic. Then I moved to the product team, in the squad that connects Landbot with tools like HubSpot, Salesforce, Google Drive, Calendly and WhatsApp.
- SEO kept, then improved
- Mobile
- Marketing autonomy
- No-code integrations
- Role
- Web Designer → Product Designer
- Timeline
- 2021 – 2023
- Scope
- Webflow migration, blog and SEO, brand design system, integrations squad
- Users
- Marketing and sales teams, mostly non-technical
Fig. 1 · Web + Product · Integrations
My role
Hired to lead the website migration. Once it shipped, product designer in the integrations squad.
- 01
Leading the migration from WordPress to Webflow, with the blog, the main source of organic traffic, as the critical path.
- 02
The audit and the five rules of the brand design system, so marketing could build new pages on its own.
- 03
A key mobile iteration of the site and the blog.
- 04
In product: integrations with third-party tools, designed so non-technical users don't have to set up webhooks or APIs.
TeamWeb: with the marketing team, no dedicated engineering for the site. Product: the integrations squad, with PM and engineering.
Context
Landbot helps companies build chatbots without code to capture and qualify leads. Its website was the main acquisition channel, and its blog ranked better than competitors for its target audience.
Landbot doesn't replace a CRM or a sales tool. It connects to them, so the whole funnel works as one. That makes integrations part of the product's core value, not an add-on.
The company bet hard on no-code as a culture, not just a feature: non-technical people doing technical things. Looking back, it was a precursor of what vibe coding is today.
- Live site
A live site that couldn't stop converting during the migration.
- SEO
A blog whose SEO ranking couldn't drop.
- No-code users
Users who build chatbots precisely because they don't write code.
Problem
Chapter one, the web: the site had grown for years on WordPress, one landing page at a time, each added by a different hand. It wasn't one ugly page. It was the lack of a system: CTAs that gave “Sign up” and “Preview” the same weight, two typefaces mixed between menu and submenu, no memory of what already existed. And any mistake in the migration could cost the organic traffic the blog had earned.
Chapter two, the product: connecting a chatbot to a CRM, a calendar or a spreadsheet often meant setting up webhooks and APIs. For a non-technical marketer, that's where the funnel broke.
Landbot doesn't replace your CRM. It connects to it, so the integration is the product.
Key decisions
- 01
Protect the blog first.
The blog was the main source of organic traffic and SEO, so the migration was planned around it. The goal wasn't to keep its ranking. It was to come out ahead: better performance and fixes for issues future indexing updates would have penalized.
trade-off: The blog set the pace of the whole migration.
- 02
Audit before redesign, then five rules.
Every proposal traced back to the exact screen that motivated it. Then five rules instead of a hundred fixes: one color per function, one typeface per role, unified containers, motion only where it helps, retina only where it pays off.
trade-off: Weeks before the first visible change.
- 03
Integrations without webhooks.
Instead of asking users to wire APIs, each integration became a guided step inside the chatbot builder. The technical part stays behind the scenes; the user just picks the tool and maps the data.
trade-off: Every integration is designed one by one, instead of one generic webhook for all.
Craft
Chapter one in pictures, the audit and the published homepage. Then chapter two: integrations inside the builder.
Outcome
SEO kept, then improved
The blog came out of the migration with better performance and fixes for future indexing penalties.
Mobile
A key mobile iteration of the site and the blog.
Marketing autonomy
New pages built from the system's components, without a design ticket for each one.
No-code integrations
Tools connected to the chatbot without webhooks, measured by customer service cost and the share of incident tickets.
Reflection
No-code was an early version of vibe coding: non-technical people building technical things. Designing for them taught me to hide the webhook, not the power. Behind every “it just connects” there's a webhook someone decided the user shouldn't have to see.