Lessons learned in building sustainable and accessible chatbots

In August of 2021, CARE constructed and launched its first chatbot in a report 4 days to reply to and assist present and former CARE Afghanistan workers who’re struggling within the aftermath of the US leaving Afghanistan. Since then, CARE continues to experiment with utilizing chat for influence.

CARE, in partnership with Meta, envisioned a undertaking that might leverage CARE chatbots on WhatsApp as a method to assist optimistic public well being outcomes in two nation places of work, CARE Bangladesh and CARE Nigeria. The purpose was to measure the impact of supplementing pre-existing customized applications in nation places of work that promote early childhood immunization by chatbot interventions.

All through 2022, CARE Bangladesh and CARE Nigeria have been growing and testing content material and studying tips on how to use it Turn.ioA chatbot creation and administration platform, and developed Fb advert campaigns to drive chatbot consciousness and adoption.

TL; DR?

• In 2022, CARE determined to develop chatbots on WhatsApp to assist CARE Bangladesh and CARE Nigeria applications to advertise immunization in opposition to communicable illnesses, chatbots have been launched and promoted from December 2022 to February 2023.
• CARE has realized many necessary classes about chatbot logistics, content material improvement, constructing chatbots, in addition to launching and selling the service.
• Key info included growing a sensible schedule, budgeting generously to plan the chatbot’s content material and communication move, and constructing a sustainable organizational construction round managing the chatbot.

This undertaking took quite a bit longer than we anticipated. It appeared as if every thing that might go flawed had gone flawed. The cool half about that’s that it taught the CARE crew so many classes. Listed below are the steps CARE will take subsequent time. We hope it’ll function helpful ideas for different INGOs within the sector who’re experimenting with WhatsApp chatbots for program influence.

abstract