DeepL
Onboarding for Language AI
This is a story about enabling growth for a multi-product platform with 100+ million of users.
DeepL is a machine translation platform that helps hundreds of millions of users globally to translate, write, and refine text, documents and speech using AI.
I led product design in the Acquisition & Onboarding team, collaborating closely with a PM, Engineers, UX Researchers, UX Writers, and stakeholders from Product, GTM, Sales, and Customer Support.
The Challenge
We discovered a recurring issue within enterprise accounts: users were unaware of the product’s more advanced features such as e.g. team glossaries, tone of voice settings or document translation. Despite the product’s sophistication, most engagement stayed within the basic text translation flow.
This limited feature adoption not only reduced perceived value but also led to lower activation and higher churn among enterprise seats. The challenge was to design an onboarding experience that helps users discover and understand DeepL’s advanced capabilities, without adding friction to their core workflow.
The Approach
We began by conducting user interviews and shadowing sales calls to understand how large enterprise teams were onboarded and where friction occurred. Additionally we tapped into research of teams that owned pro user features. These sessions, combined with analytics data, gave us a solid insight of user behavior and team dynamics. Our research focused on two key user types: enterprise end-users and admins across European markets.
Key insights
Through interviews and data analysis, several consistent patterns emerged:
Users were unaware of advanced features and often didn’t recognize their names.
Many struggled to understand what certain features did or how they fit into their daily workflow.
There was no clear location where users could learn about features after an initial onboarding.
The information architecture made it difficult to navigate between DeepL’s growing ecosystem of products.
Enterprise admins lacked visibility and tools to enable their teams efficiently.
Problem statement
New and returning B2B users struggle to comprehend the breadth of DeepL’s capabilities beyond translation. They underutilize their Pro plans, leading to suboptimal adoption and higher churn, a risk that increases as we introduce more features and products.
Goal
Guide PRO users to experience DeepL’s value and become autonomous in unlocking its full potential through onboarding and continuous learning experiences tailored to their context and needs.
Our strategic initiatives centered on:
Driving awareness and reducing time-to-value
Adapting onboarding to different user segments
Establishing continuous onboarding across the product ecosystem
The vision tied directly into DeepL’s broader transformation into a multi-product, B2B-ready SaaS platform.
The Solution
As our team and the whole organization was suffering from growth pains like frequent restructuring of teams, lack of technical resources and strategic shifts, we decided to follow a simple principle to be able to ship: use proven tactics to get tangible results quickly, then iterate and evolve. Rather than choosing one direction, we implemented a combination of complementary solutions, e.g. creating a holistic onboarding system that scales across products and user segments.
Given the complexity of DeepL’s enterprise landscape and the team’s limited technical bandwidth, we focused on solutions that could deliver immediate value and scale later.
Flow overview
The redesigned journey guided enterprise end-users through a structured but lightweight flow:
Signup within a multi-seat account
Intent questionnaire to capture user goals and needs
Getting Started sidebar as a persistent learning companion on the home dashboard
Contextual onboarding tips and prompts triggered by user behavior or plan type
Intent questionaire
Asking for high level intent to customize the onboarding content to user needs
Contextual onboarding modals and tips
We implemented onboarding modals and it´s dynamic logic. Brief, contextual prompts helping users discover new functionality in a non-intrusive way.
Prototyping and Testing
We built low-fidelity wireframes and Figma prototypes to quickly validate our concepts with users. Usability studies and interviews with both end-users and admins revealed key friction points early.
Each major initiative was then launched as a two-week A/B test. After each experiment cycle, we analyzed data in Metabase to evaluate hypothesis validation and identify opportunities for refinement. This lean, test-and-learn approach allowed us to deliver progress continuously while collecting measurable insights on adoption and engagement.
The Impact
The launch of the new onboarding and home experience delivered strong signals of success across key business and product metrics. Within the first months, we saw:
clear increase in feature adoption rates
a rise in active seats
a drop in churn among enterprise accounts.
Early cohorts showed higher activation after 30 days, validating that the improved guidance and contextual discovery helped users realize more value from DeepL Pro faster.
Reflection
A completely new team. All team members new to the company. No backend support for the first 6 months. Frequent organizational reshuffling. And yet, the onboarding initiative is just one of three major initiatives that we have successfully delivered as a team during my time at DeepL. I am beyond proud and humbled having been part of this team and the work we have delivered.