The First of its Kind

🌕 Oct 2017 - Oct 2022

For over 5 years I had been working on Odus.AI — an AI-based conversational chatbot platform. This platform became the first visual chatbot builder on the market, the first of its kind among AI bot platforms. I was one of its founders and was responsible for UX/UI design, user research and team management. I managed a team of 7 people and 3 outsourcers.

Research.
Analyze.
Create.

Start Point

Starting this project, we knew almost nothing about the industry, but the team that we were had a lot of experience in related fields. Therefore, first of all, I began to study the market, competitors, technologies and tools available at that time.
‍

Divide & Conquer

Research was a two pronged approach:
Сompetitive Analysis: Identify what we must have, and how we can differentiate Odus by researching the top 3 competitors currently in the market.
Customer Development: Ensure we solve the user's problem for both sides - developers and customers by having conversations with 20 individuals (devs, managers, users). Through both research methods, two threads emerged: accessibility and efficiency. This opportunity guided our design decisions.
‍

First Steps

Having studied the results of the research, we came to the conclusion that it is necessary to create the simplest possible interface. At the same time, the developer should be able to create a powerful chatbot that will really benefit its users. Our team developed a powerful software solution that I had to wrap in a simple and understandable interface.

My design process:
‍

High-detailed prototypes of the main pages and core interactions.

Dashboard with your chatbots and stats.

Introduction. The onboarding process for new users.

The main area where chatbots are created.

After all layouts are ready, we need to add comments for developers.

Odus Design System: I developed a design system because several people worked on the project and we outsourced part of the front-end work.

Download full size image

Wrap Up!

During the first year after the launch, the platform attracted 2500+ businesses and 10000+ chatbots were created. To be completely honest, I have to state that the UI has been constantly refined and improved over the two years. As a result of this small venture, the startup was sold to a large technology company.
‍

Lots of Practice

🌕 Apr 2013 - Sep 2017

Before founding Odus, I worked for two agencies within 4 years where I worked on many different projects as a product manager and designer.

For example, as a product manager I contributed 10+ big projects of chatbots for large Russian companies such as Synergy, Alfa-Bank, Russian School of Management, and others that attracted 50.000+ users in total. I conducted workshops on development and marketing. Also, as a product designer I was responsible for UIs for a big site for job searches GorJob.ru, and improved user engagement – from 12 up to 18 pages per visit.

But first of all I want to pay attention to an unusual and complex project. I was lucky to work on adapting the site of Wikimart for a touch trading terminal. Wikimart was a huge online retailer in Russia like Amazon in the US. Unfortunately, due to a bad business environment, this project went bankrupt and closed. But let's not talk about sad things, let's talk about the challenge that fate has prepared for me...

Research.
Analyze.
Create.

Background

Starting the project, our team already had a partially updated design of the company's main website. I was faced with the task of redesigning the terminal based on an already existing design system.
‍

Finger instead of Сursor

To understand how the user behaves using the terminal on the trading floor, we invited a focus group of 10 people of different ages and physiques to create persons. The result — the main problem in working with the UI was the size of buttons and text links.
‍

Ten Iterations

In order to find the best version of the interface, I had to create 10 different versions and conduct A/B testing. The main focus was on the homepage and category navigation, since terminals are mostly used by new customers and first impressions are very important.
‍

Detailed prototypes and design system.

Final homepage with categories and popular goods.

Product page: tabs with information about the item.

Download full size image

Ready to use!

The new interface was launched on all terminals in the company's shopping centers. This allowed us to increase the number of new customers per month by 10.52% and increased the average check by 18.67%. The main conclusion: The closer a touch interface resembles objects from the real world, the easier it is to use. In particular, this applies to controls. The fewer small elements, the better. Especially text links — a very unfortunate solution.

New Frontiers

🌕 Jul 2012 - Feb 2013

Before I started working with various agencies and external clients, I worked for a TOP 3 tech company in Russia called Mail.Ru Group, now better known as VK after they acquired the company of the same name. I only worked there for a year, but it was my first unique experience in a company of this size.

About the project: Our team consisted of 6 people, 3 developers and 3 designers. I completely redesigned the UI of the games section and the news feed for the My World (Moi Mir) social network in about 6 months. Although this project was completed many years ago, they still use the design framework I created. Unfortunately, I have not preserved the more materials, and besides, I am forbidden to upload them. So let's move on to the next big project. It's also my first big startup.

It Was Before Uber

🌕 Sep 2009 - May 2012

I started this project after I got tired of being a just freelancer. Yes, this may sound strange now. But Taxofly was created before Uber was born in its current form. In this project, I was involved in everything related to the product - UI, analysis, customer research, and designing the product structure from scratch.

Research.
Analyze.
Create.

From Scratch

I started this project with my friends when I was still at university. We wanted to disrupt the market for ordering a taxi. Our team had talented developers and a poor understanding of the industry. But this didn't stop us and we began to study business processes.
‍

Amazing and Incredible

At the time, such a concept seemed unbelievable. There is no competition - it is impossible to peep how they do it. Therefore, our analysis consisted in studying the principles of taxi operation and a broad analysis of audience preferences.
‍
Customer Development: We called 15 taxi operators and 50 potential customers. It is necessary to determine the preferences of both parties in order for the platform to work correctly. We've found that the speed and ease of use of our apps are critical for customers and taxi operators.
‍

Step-by-step

This means that the UI should contain minimum elements, and the ordering process should contain as few steps as possible. I developed prototypes and UI for the site, mobile applications for drivers and customers, as well as a web interface for dispatchers. At that time, early Android smartphones and classic phones with Java applications were popular. Therefore, we were very limited in technical terms.

I created and redesigned the UI many times because I was constantly in touch with customers and drivers. They told us what they liked and didn't like, and I worked with the developers on the feature roadmap and interface redesign. This endless process lasted until the sale of the company in 2012.

US Patent US20120084225A1
‍

Detailed prototypes and user flow.

The process from registration to the trip took 3 screens: Registration, filling in the address, and choosing a car. The prototypes above reflect all states, including intermediate ones, and show the user flow for all sections of the client application.
‍

Android app: Steps to order a taxi.

The company website with info for clients, drivers and partners.

The first and only application for ordering a taxi online on Java phones in the world.

Flying high

After many iterations, we managed to win the hearts of clients throughout Russia and neighboring countries. In total, the service was present in 28 cities. The average time for ordering a taxi is 37 seconds from the moment of registration in the application. Shortly after I left the company, it was sold to an American investor and then acquired by a competitor. A good ending to such an exciting journey, I suppose.

Alex Chernyshev

The First of its Kind

🌕 Oct 2017 - Oct 2022

Shared knowledge

Lots of Practice

🌕 Apr 2013 - Sep 2017

New Frontiers

🌕 Jul 2012 - Feb 2013

It Was Before Uber

🌕 Sep 2009 - May 2012

In conclusion, I'd like to say...

Some other projects...

That funny life story

Want to learn more?