Design Strategy
Save
the bot
Save
the bot
8% loaded

[Top screenshot]

“Im liking the tone of the bot, "let's see what is in there" makes me feel like Im exploring the product rather than carrying out an exercise.”

Save
the bot

[7 minutes read]

Company Profile

Name: Genesys

IR, Galway, SaaS, Enterprise, B2B

“Every year, Genesys delivers more than 70 billion remarkable customer experiences for organizations in over 100 countries. Through the power of the cloud and AI, our technology connects every customer moment across marketing, sales & service on any channel, while also improving employee experiences.” [genesys.com in 2021]

Business objective

Bots have been the biggest innovation to reduce costs for Genesys customers in the call center service industry. Until now, the company had to rely on developing chatbots with third-party applications and connecting them to its own software suite via interfaces.

Of course, this meant dependencies and additional costs in a growth area that was important to Genesys' business customer base. After I started in March 2019, Genesys wanted to sell a fully-featured bot authoring software by the end of the year.

"Build a working bot authoring app by the end of 2019 that will enable Genesys to offer customers their own solution, reduce costs and reduce third-party dependencies."

In this case study, I tell the story of how a group of courageous product people and I persevered and convinced the US sales team to move forward.

Topics

Project stats

The team

Team composition

Cross-functional:
2 Product Managers, 1 backend developer AI, 1 frontend developer, 1 UX UI Designer

Supported by:
Design Lead, Shared Junior User Researcher

Areas of improvement

Complete new bot authoring application for Genesys

Our processes

Scrum

Application Type

Cloud app for tablet & desktop

My stats

My project duration

Mar 2019 - Mar 2020 · 1 yr

My role

Senior UX Designer AI, Onsite

My responsibilities

All UR, UX, UI, Design tasks

My tools

Pen & Paper, Workshop Office Supplies, Adobe XD, Miro, Atlassian Confluence, Atlassian Jira

[Top image]

The design shown above was the result of a series of meetings and workshops with representatives from a wide range of disciplines. We focused on putting together the necessary features for a bot author to accomplish the tasks.

Problem

Problem

Only for pro devs

The language, data, AI and information scientists were spread across three different continents. Our first focus was on defining all the tasks that need to be completed so that a conversation bot can be created by our customers. Therefore, this version was particularly suitable for specialists who already had the experience building chatbots.

A competitor copy

Everyone involved probably already had solutions from competitors in mind. Because the focus was on user flow and UX, we used a standard UI kit for these initial UI designs. In this combination, what we had created up to this point also looked like established products on the market.

Screenshots of IBM's Watson or AWS Lex shared by senior stakeholders look pretty similar. Our Genesys unique selling point was missing.

High costs for setup & maintenance

In interviews members of the senior sales team in the USA told me that getting started was the biggest problem for customers. As a rule, the salespeople had to acquire all the skills themselves in order to then set up the chatbots for Genesys's customers using alternatives from AWS or IBM. This usually got really expensive very quickly. The prospect of needing an expensive expert to set it up in the future also put off many customers.

USP Onboarding?

Could a good idea of how we can get future users on board easily and safely make all the difference?

[Top video]

The video above shows an early prototype I created for feedback sessions with senior stakeholders. To address limitations due to the many time zones, I introduced the “UX Helper” for colleagues who couldn’t participate but wanted to give feedback.

Full prototype

It might be that the full prototype is still available under the following link:

Next topic

[Top image]

The screenshot above shows the moment of the bot's conversation with the user explaining the Ghostbusters scenario.

Users & mission

Users & mission

Kevin

What if even simple office workers who handle calls could create a bot? Maybe you saw Ghostbusters 2016? Have you perhaps noticed Kevin Beckman, played by Chris Hemsworth, is the new secretary of the Ghostbusters? The brilliant product manager Jonathan McKenzie had the great idea for a very special persona mission and user story.

Who you gonna call?

In our persona mission, Kevin from the Ghostbusters has the brilliant idea to make his life a little easier. Our Genesys Click and Go bot is designed to help him with this. In the best case scenario, it's so easy that he can also provide a bot chat service for the new Ghostbusters branches in Los Angeles and San Francisco.

[Top image]

In the image above, I have cut out the part of the Confluence page that shows the ideas that were most voted by the workshop participants.

Process
& iterations

Process & iterations

Meeting minutes on steroids

Simple meeting minutes would never have been able to consistently and securely communicate complex content to the highly fragmented global team. So that participants who were unable to attend a meeting were not forced to watch the entire video recording, I took live notes and sketches on the screen during the sessions.

I was then able to publish these as clues with a screenshot as a basis for the summary in Confluence. This meant everyone was able to access the information they needed efficiently and transparently. All of this helped to enable the functionality of the bot app in the first step.

First hybrid sketching session

We agreed in mid-April to make the topic of 'onboarding' a priority. I had already conducted sketching workshops with the team at Rentalcars several times. They really open the minds of a team and generate lots of awesome ideas.

This type of workshop was completely new to the cross-functional Genesys team that I invited. For me, however, it was completely new to organise this meeting with a group that was both in the Galway office and remotely in Belfast. A little more effort had to be made to ensure that colleagues could participate remotely equipped with the appropriate tools.

Design Thinking

The entire process followed the concept of Design Thinking which meant that the workshop did not stand alone. Similar to my meeting minutes on steroids I organised and documented the workshop using the software Confluence from Atlasian.

The session begins under the heading 'Empathize' and here a product manager with a PHD in AI talked about her experiences. I gave a summary of the interviews I conducted with sales on the topic, presented the persona idea and insights I got from a competitor analysis.

We then started to 'Ideate' using a modified version of the popular sketching method ‘Crazy Eights’. Participants had 10 minutes to outline around five suggestions for how we can onboard users like our persona Kevin into our bot authoring software.

Vote, prototype & Test

“Everyone has 3 min. to show their max. three best ideas they had to the group. All remote participants can show their mockup into their webcam. After that please send me a photo of your mockup to alexander.kruck-kneip@genesys.com. At this stage, we will give no feedback.” [From the original meeting minutes of '2019_04_12 - PW - Mockup Session - Bot Author Onboarding']

Because of the hybrid situation, I have decided that all participants will have until 12pm the next day to cast their vote. The winner of the session with 7 stars was the idea of the AI developer 'Guided conversation setup'. This idea was the basis for the next two steps in the design thinking process, prototype and test.

[Top video]

The video shown above shows the section of the prototype with which the user test was carried out.

Final design
solution

Final design
solution

De:Bot

Based on the best idea from our sketching workshop, we wanted to present our users with a character who would be fun to have a conversation with. Starting from the design elements of the Genesys logo brand, I finally designed the face for our bot authoring software Dialog Engine, De:Bot. In many extra shifts, I created a complete icon set and components that fit the design language used.

Click & go bot

In our previous concepts, we have not yet managed to create something for Kevin that would allow him to simply click together a bot from simple components. As I was playing around with basic shapes to find a solution, the idea of the bot component library came about.

The Library

The special idea was that a new dimension was suddenly possible for our application. In addition to the possibility of structuring elements of the bot authoring software simply and clearly, the option of a 'Marketplace for bot components' was also on the table.

With this idea, Genesys or third parties could create paid modules that users could purchase to make it easier to complete the task in the Ghostbusters chat.

Conversational onboarding

Yes, the ideas just flew back and forth between the team members. But back to our first idea of creating a bot that could also safely and easily get non-developers on board. Together with product management, we designed a bot-led conversation that allowed all important elements of the application to be presented. Of course we used the Ghostbusters story as a theme and as a special treat, De:Bot turned into a ghost during the conversation.

Full prototype

It might be that the full prototype is still available under the following link:

[Top video]

The video above shows another section of the prototype that was later used in feedback sessions with senior stakeholders.

Next topic

[Top video]

Feel free to turn on the audio for the video above to hear our wonderful PM putting the final, important touches on it.

Results &
key takeaways

Results & key takeaways

User testing

In early July, we finally tested the prototype with seven internal users and members of the senior US sales team. In summary, points 'Good overall design and Information Architecture' and 'Good story behind the onboarding' stand out particularly positively. However, the entire onboarding process was a bit too long. To solve this, there were suggestions to have an option to ask the bot for a break during the conversation.

User quotes design

“Text hierarchy and icons are obvious.”

“It seemed to be obvious to the user to use the select all box to select the Library items.”

“Very clean and modern.”

User quotes IA

“Im liking the tone of the bot, "let's see what is in there" makes me feel like Im exploring the product rather than carrying out an exercise.”

“I loved it :-) it does feel like it took me a long time to get to this point. If I was a real user I maybe would have clicked on through some stuff but I was analyzing everything to give you feedback.”

“It feels like someone is helping me.”

Green light

Nobody would have thought that our bot creation application could have such a wonderful conversation with its users. We not only completely convinced and, above all, inspired the senior sales team in the USA but also ourselves. We got the green light and the full support of all senior management.

Now that we were at the point that sales would support us to test with real users we didn't know how to proceed immediately. No design team has ever been at this point before at Genesys. My answer was organising a couple of ideation sessions with senior stakeholders with the goal to get lined up ten B2B users for testing within the next two months.

Let's line up real users

The product team regularly sent a presentation with an overload of information to customers to get in touch. The response rate was quite low. But I was able to get everyone on board with the idea of a little product promotion video. The copy and animation for the promotion video was ready, but we were missing sound.

To convince our PM I did the first voiceover. My German accent was probably not so well received and the product manager showed how it can be done so much better.

Within two weeks we were able to double our expectations and lined up more than twenty customers who wanted to support us doing regular user testing on Dialog Engine.