Improving Cerego

Work      About

Understanding how instructors feel about creating study material with the help of AI




Instructors have PDFs, presentation slides, and sometimes textbooks or manuals. What they often don't have is time to manually convert existing material into Cerego.

Cerego has AI-powered tools that automate the grunt work of the task of converting materials, however the reactions from our beta users have been extremely mixed.

I wanted to dig into this disparity and understand how instructors and instructional designers felt about using AI to assist them in converting existing material into micro-learning lessons on Cerego.









UX Research

Challenges & Constraints 

This was my first time leading a UX Research project at Cerego. I enjoy research but don't feel too confident about my skills. Luckily, my teammates are happy to collaborate and assist.

Cerego recently began venturing into the SMB onboarding training space, which means accounting for instructor and learner needs that are quite different from our existing Higher Education users.

What kinds of research?

Foundational Research

To understand who's using our content authoring product? What are their frustrations, daily schedule, needs, and wants?

Evaluative Research

How is our current content authoring tool and AI-powered features doing to solve problems for our instructors/instructional designers?

Discovery Research

What should we do next to improve our content authoring product and instructional design services?

Why do some users embrace automating quiz creation while others are skeptical?

Mostly automated

Some instructors and instructional designers quickly embraced our automation features. I wanted to understand why where their trust for our system stems from and what their process is for quality assurance and curation.

Mostly manual

I interviewed users who only manually create lessons from scratch. I wanted to understand whether it was lack of awareness or distrust in the idea of automating lesson creation.

"I don't like your default so I re-write them."

– Instructor

What tools have our instructors embraced that we can draw inspiration from?

When asked about the tools that helps save their time, our instructors mentioned a couple products that use AI, even if they may not be aware of it. I wanted to understand why some AI-powered products were embraced and why others aren't.


They may not realize it but there's a lot of machine learning that goes into SPAM filtering, categorization, and smart compose.

Credit: The Keyword (Google Blog)
Credit: The Keyword (Google Blog)


Our instructors like the suggestions that Grammarly provides them without assuming what's best for them.

Credit: CCSD
Credit: CCSD

Literature Review

There is surprisingly not a lot of high quality resources available to the design community for designers who want to learn more about taking a human-centered approach to designing with machine learning and AI.

I wish more companies building AI-powered products can share their experiences, failures, and learnings with the design community. Here were some useful readings I found:


People + AI Research, Google

How to design for AI-enabled UI,

AI UX: 7 Principles of Designing Good AI Products, UX Studio Team

AI and the Future of UX, UX Design Collective



I liked how easy it is to go along with their suggestions for improving your writing but also how causal it is to dismiss anything that you don't want.

Learning from other teams, as a team.

I attended several talks with our data scientist and engineering lead to see how other companies' designers approach designing AI-powered product.

(Side note: I always invite non-designers to come with me to design events because designing great experiences goes beyond the responsibilities of designers.)

Designing Products with AI

Designers talked about their experiences designing for AI-powered products and what they learned when people didn't want to use them.

Credit: AIGA

GHC Product Inclusivity Panel

Designers shared their experiences designing AI-powered products that went wrong because of bias in the model and how designers can work with data scientists to be more aware of these issues.

Credit: Google

Our data scientist sharing a summary of our learnings


Guiding Principles

This research project allowed me to understand our instructors and instructional designers a lot better. I created a set of guiding principles for how we should approach AI in our content authoring product.

Product Recommendations

As a result of the research, I had a lot of ideas for how we might improve our content authoring product to balance convenience and control for instructors. Some of these ideas are currently in progress and will be shared once we launch to the public.

Ongoing research

The BBC Bitesize team has created some of the most content in Cerego. I was lucky to visit them in the UK with our engineering lead, instructional designer, customer success manager to understand their needs better.

Back to Top