cognibly

AI Career Services Platform

Cognibly—Job aggregation tool powered by AI including performance metrics.

OVERVIEW

My Role

Lead Product Designer

Team

Reid Thomas (Founder)

Niko Leban (Backend Developer)

Jerry (Frontend Develeoper)

Timeline

December 2024 - Present

Breakdown

As the Lead Product Designer at Cognibly, I designed the website and platform from the ground up. The platform aims to provide users an automated process of getting jobs that fit their background by using AI and mathematical vector embedding.

I conducted research through surveys and usability testing to improve the experience users would have on the new alpha release. I collaborated with my team to resolve any issues that were design and tech related through an agile framework.

HIGHLIGHTS

$25,000 Azure credits

With the designs, our founder was able to pitch the idea and designs to Microsoft and received funding to continue the development of Cognibly

Credibility in AI

Through collaboration I was able to provide users with more context and clarity to our match making algorithm

Optimizing tasks

I redesigned how users interact with different features resulting in an increase in usability and optimization

CONTEXT

It’s not user-friendly.

The current design, created by our Founder, is not very user-friendly. With the large volume of listings, navigating it can be challenging. While the design provides thousands of jobs daily and organizes them by best fit, it resembles a spreadsheet, which is not visually appealing and adds to the existing problem.

THE PROBLEM

My findings.

I conducted a short survey to get an idea of how people felt about job searching and its relation wtih AI.

People concerned about sheer volume of job listings

On the current platform it’s hard for users to navigate through a datasheet of jobs and all the information

People feel uncertain about AI matching

Telling the user a score isn’t enough to prove why they are a good fit

Mapping out a user's journey.

From the research and my own understanding, our users experienced frustrations with AI matching.

DESIGN PROCESS

More features in a simple way.

We wanted the platform to have the functionality to support multiple career trajectories and visualizations of job data. All this needed to be included while enhancing readability.

Wireframes.

Given the old designs on AirTable, I came up with some initial conversions of those features into a user-friendly layout.

Conducting some testing of different designs.

A/B Testing.

I had a total of 4 participants who tried out the 2 different versions of the design mockups.

Reviewing the feedback.

Version 1

This version had a simple add icon and a smaller input overlay field

Digestible information on the home screen

Information-heavy when creating a profile

Adding a custom profile was a bit confusing

Version 2

This version had a text with the icon and more space for the input fields

Users completed tasks faster

Filter features were easy to adjust

Unused space on the right column

Making some changes.

Progress update

Instead of all the information on one page, it can be split in different sections and a bar can show the progress of creating a profile

Clarity

Ensure that the feature for creating a profile is clear and the user should know that they can have multiple profiles

Applying the feedback.

After integrating the feedback into the design, I created the mid-fidelity prototype.

Working with developers.

Front-end issues.

As more of the designs got approved by the founder, the front-end developer started to work on the designs but quickly noticed an issue with the readability of the home page.

The solution.

I suggested that instead of sticking with a table design and having to deal with different widths, we can reuse components that we already have on a separate page to keep it more consistent.

Addressing use of AI and operational costs.

I began to explore ways to provide users with more credibility of our matching process. The first method was too costly as it was require extra use of AI. I dug deeper and found more information as to how we actually match users and used that to provide visual context.

COLLABORATION!

We were able to figure solutions before further development was made by resolving technical and resource issues with design.

FINAL DESIGNS

Building trust with the users.

Since most of the survey respondents felt uncertain about how AI matches people with jobs, we knew this would be a problem down the line. To solve this, we use AI to tell users why they are a fit. Once that trust is built users will feel confident in AI.

Main dashboard.

Glimpse of information

At a quick glance, you can see what the platform offers and each component has an entry point to see more

Filters

Users can change the time filter to view different instances of data

Multiple profiles

We want to support users with multiple career trajectories to maximize results

Job listings.

Why they're a fit

Users are given a reason why they are given a fit, to help build trust about AI matching them with jobs

Users can search for specific job titles and companies within their job database

Tailored Documents

To help users save time, we tailor resumes and cover letters to fit each job

Performance metrics.

Valuable data

A way to help users track their progress of a job search is essential to stay organized and improve their workflow

CONCLUSION

Testing uncovers hidden issues.

The goal of the redesign was to enhance the user experience of the platform from a spreadsheet to an intuitive interface.

But, with research and usability testing, I was able to find underlying issues with how people felt about integrating AI into their job search. In the end, I was able to design a trustworthy product that not only met the needs of the users but also the business.