THE CLIENT
Healthy Babies Bright Futures

An AI-powered, multilingual chatbot that gives personalized recommendations for reducing lead contamination at home.

=

Healthy Babies Bright Futures (HBBF) is dedicated to measurably reducing the largest sources of babies’ exposures to toxic chemicals that harm brain development. Whether it’s clean drinking water, lead-free homes, or pesticides in produce, HBBF targets chemicals with the strongest body of evidence supporting developmental harm. Their work is changing lives and policy across the country, and making homes safer for millions of children, with a focus on protecting low-income, BIPOC communities that are most likely to not have access to clean water and air.

The Problem

Lead contamination poses a toxic threat to families. In 2017, Exygy partnered with Healthy Babies Bright Futures to help families locate potential lead in the home, test for lead, and create a plan to mitigate contamination. Exygy was challenged with creating a unified system that could seamlessly connect a high volume of users’ unique information to HBBF’s science-backed recommendations and samples from a testing laboratory.

Additionally, HBBF had a larger goal of scaling their reach — both increasing the number of families accessing their mitigation recommendations, and as well as enabling their team to identify families that were at high risk for lead exposure.

The Approach

Increasing User Engagement with a Chatbot

In the beginning, Healthy Babies Bright Futures wanted to create a lengthy survey on home-based lead. However, we know that long-form surveys generate low engagement. We needed to make it as easy and intuitive as possible for users to exchange data with HBBF.

We recommended that HBFF pivot from creating a long-form survey to an interactive, web-based chatbot. The ability to create a customized user experience would help HBBF meet their goal of increasing the reach of their work, ensuring more users were motivated to complete activities to make their home lead-safe.

Designing a User-Centered Chatbot Experience

Our design team was excited about turning a long-form survey into a conversational user experience that could adapt based on the user’s input. For example, if a family’s home was built before lead-based paint was banned in 1978, the chatbot needed to ask additional questions about their circumstances. If their home was newer, it could move forward the user to the next section. 

A user flow that maps the sequence of the lead testing process.

The chatbot’s design was informed by:

  • A Design Jam that allowed us to co-create a solution with our clients through sketching initial ideas together, as well as convening on a final solution by discussing and voting on the most relevant ideas.

  • Testing with users as a part of the initial product development to understand how they are using the chatbot and how to improve the experience.

  • Mapping the chatbot’s logic along with the client. This map organized the chatbot’s branched logic, which was based on HBBF’s verified research, to create a streamlined user experience that prioritized the most important lead testing scenarios.

  • A “human” voice and tone enabled us to communicate with users in a way that was conversational, supportive, and friendly. 

Step-by-step instructions break down the testing process into a digestible format.

Building Logic: From Spreadsheet to Automated System

When we began digging into the science behind HBBF’s lead mitigation plans, we were given an Excel spreadsheet with algorithms that HBBF used to calculate chemical risk. This posed an exciting challenge for our engineers, who needed to create automated software out of an existing manual process. 

We worked closely with the HBBF team to make the spreadsheet compatible with our automated server, in addition to building a custom upload system for the testing lab to incorporate their results to user profiles. 

Our custom system moved the chatbot’s data through a series of exchanges:

  1. A custom server asked users a series of questions, and collected data was sent back to our server. 

  1. Our server processed their answers through a spreadsheet algorithm developed by HBBF, which contained insights from their expertise in chemical analysis and exposure mitigation. 

  1. The testing lab processed any samples they received from the user, and incorporated those learnings into the user’s profile through a custom upload.

  1. We then codified the HBBF algorithm and shared the results and personalized recommendations (below) back to the user.

Results from lead testing are compiled for the user.

Sample recommendations for users with a significant level of lead in their water.
The Impact
quote mark

Vida: The Lead in Homes Chatbot

Together with Healthy Babies Bright Futures, we designed and built Vida: a chatbot that empowered parents to find and rid their homes of lead sources before children are exposed. The chatbot engine applied data on hundreds of lead-exposure risk factors and everyday items that parents could test for lead. It gave customized actions to help families get rid of lead sources, drawn from scientific literature and public health agencies. 

Due to Vida’s mobile-reactive capabilities, it had the potential to not only be used at home, but also in doctor’s offices and other healthcare settings.

A key component to scaling Vida’s impact was that it was accessible in both English and Spanish, ensuring that the households of 40+ million Spanish speakers in the US were able to get support in creating a healthy, lead-free home environment.

Although Vida is no longer live, it is a testament to the power that chatbots have in transforming dense questionnaires, engaging users, and harnessing data to help people make tangible changes for a healthier life.

quote mark

Want to work together?

We are always looking to get in touch with partners to help build healthy and resilient communities together
contact us