Meet Hannah

Improving employee productivity by providing easy access to HR information.

Project overview

Hannah is the chatbot solution designed to help employees find HR content faster and reduce the amount of calls and emails to HR support representatives.

Role

I worked with a small team consisting of HR support specialists and subject matter experts. I was responsible for the research, visual and interaction design of the chatbot.

Close project
Top

Understanding the user

The audience

Target users for this venture were employees and managers at The Hanover. To better understand the users and their experiences using the intranet, I sat down with several managers and employees to get a sense of their daily intranet use.

Examining patterns

We explored common behavioral patterns like navigation, learning preferences and user experience; but also expanded our scope to include sporadic patterns like time constraints.

Building empathy

Personas developed to get a better understanding of the employee’s background, pain points, motivations and goals.

Employee persona

Manager persona

Discovering pain points

User journey map developed to uncover the employee's pain points as well as explore opportunities to make improvements.

Defining the problem

The latest iteration of the company's intranet equipped employees with a lot more HR content. But it was challenging to dissect through the phase-by-phase navigation style to get to the desired content. 
As a result, employees spend greater time searching for HR information than attending to their work responsibilities, making them less productive.

Solution

Enabling quick access to HR content

Hannah is the chatbot solution designed to revolutionize access to content on the Intranet. Chatbots are expanding rapidly and according to IBM:

  • 53% of customers would happily message than call a customer service agent 
  • 85% of all customer interactions will occur through chatbots by 2020

The purpose of the chatbot at each stage is to help the user achieve their goal by reading keyword input from the user and serving up direct links to HR content on the intranet. By doing that, employees would spend less time looking for content or calling HR support and can spend more time carrying out their work responsibilities.

Design

Creating a meaningful experience

Designing a great conversational experience starts with thinking less about what we wanted the chatbot to say and more about what the user wants to do. In our case,  I focused on the employee’s need to quickly access HR information from the intranet.

Establishing conversational chat interface

Choice-based or open-ended?

At this point I was  stuck between designing a choice-based or open-ended conversational flow. Initial web research showed that choice-based flows made it easier to control conversations and guide the user to the desired content but testing this option with users, I discovered a lot of the employees felt a bit overwhelmed with the amount of choices presented which quickly led to another pain point.

A good number of the employees wanted the freedom to communicate their intent to the chatbot but the choice-based flow presented an interaction constraint.

On the other hand, by engaging the user in open-ended conversational interaction patterns, we were able to encourage the user to communicate naturally as well as collect information that provided a better understanding of the user’s active intent, needs and desires.

By designing an open-ended chatbot, we were able to deliver personalized content with a greater potential to satisfy the employee’s needs.

Keeping it simple

I made a conscious effort to simplify the open-ended conversational flow as much as possible and after some iterations, I arrived at a simple flow.

  • Chatbot opening
  • User motive
  • Navigation
  • Ending

Building conversational flow

User flow established to inform an open-ended conversational flow.

Crafting personality

Conversational interfaces exist for better interactions between humans and computers. So it was crucial to design a chatbot that was relatable, relevant and capable of being understood by the employees.

Chatbot opening message

User intent

Chatbot navigation

Feedback from user

Chatbot opening prototype

Testing early

During one of our early user test sessions, a manager asked “Can the bot detect if its interacting with an employee or manager?” We realized the chatbot was unable to do that but the question presented a new challenge to think of ways to design an inclusive chatbot that serves the employee and manager population at the same time.

A few more iterations later, I decided to combine the chatbot’s response to employee and manager content. This presented users with the option to either view manager or employee related content.

Manager and employee content prototype

Iteration

Results

After launch we saw a 45% increase in employee engagement. We also noted a 23% decrease in calls and 18% decrease in email inquiries to HR support. Post-launch tests with employees revealed an increase in efficiency as a result of spending less time searching for HR content on the intranet.

Lessons learned

One mistake I made designing the company’s intranet was leaving out a site tour. I assumed the website would be intuitive enough for employees to immediately understand how to navigate through the site. Learning from this mistake, I made a conscious effort to design a chatbot that clearly explains how it operates and how users can engage with it.

Reflection

This was one of those design projects that I had the opportunity to talk to A LOT of users. Naturally, I wasn't sure what to expect but I actually (thoroughly) enjoyed the process and had several 💡 moments from engaging with employees.

I’ll also note that a lot of the solutions developed initially came from ideas I got from speaking with several employees.

If I would have done anything differently, it would be exploring other avenues to communicate with the chatbot like voice activation or an image-to-text option. Then again these could be dream list features for future iterations of Hannah.

Thanks for reading!

I'd like to hear about your project

Next project

Boston University student portal