AI for impact.com
Designed for Improving Adoption and Predicting Trends
How did we start?
In 2023, I co-led the first AI chatbot project, driving innovation through design, research, and strategy while collaborating with teams to uncover fresh user insights and possibilities.
This initiative placed us at the forefront of AI integration across the organization, enabling us to anticipate user needs by addressing their most frequent asked questions.
This project was also designed to ease the strain during busy seasons and uncover valuable user trends, empowering teams to work more efficiently and encouraging stronger adoption of key platform features.
Summer seasons were peak times for Impact, making it essential to alleviate the workload of small customer success teams by efficiently handling repetitive "how-to" and "where" inquiries, enabling faster resolutions during these high-demand periods.
Integrating the chatbot into the tools and platforms our customers regularly used was essential for addressing the needs of Account and Agency Managers. Through our research, I discovered that Slack was the most commonly used platform.
15%
Increase in User
Acquistion
How did we get there?
The first steps in creating the chatbot was laying out core functionality the chatbot would have such as:
Information Retrieval: Quick access to account details, status reports, and performance analytics.
Insights and Recommendations: Offering data-driven suggestions for improving account performance and identifying new opportunities.
Feature Requirements
Next, with the business objectives in mind, I made sure to lay out high level feature requirements needed for user validation. Some requirements I focused on for V1 were:
Chat MUST abide by user management settings. e.g. users can only get responses about brands they have permission to access
Add a thumbs up/thumbs downs for feedback and validation 🔑
Feed our sitemap and a brief description of each page to the chatbot so it is able to direct users to relevant features/pages
Alfie, the AI Assistant
After extensive collaboration, the team chose the name "Alfie" for our chatbot. The goal was to create a more personal and approachable experience, helping users feel like they weren’t just interacting with a machine delivering generic answers. Instead, we wanted Alfie to feel like a thoughtful assistant who genuinely cares about their questions and strives to provide the best possible answers.
Perfecting
Performance 🚀
Key Insights I found through testing:
Limitations:
``V1 only renders tables; cannot analyze or respond to data directly.
Bugs
Occasional echoing of user requests.
Discrepancies:
Interchangeable terms cause data input errors.
What did I learn?
Collaboration Drives Progress
I worked closely with our developers and customer success team to refine and structure the language models. Company-wide testing revealed key insights, emphasizing the power of teamwork in refining and improving the chatbot.
Feedback from clients emphasized the importance of integrating tools into existing workflows, like Slack, to drive adoption and engagement.
Meet Users Where
They're At
Speak Their Language
Creating friendly and approachable chatbot responses wasn’t just about improving the user experience—it directly supported key business objectives like boosting platform adoption and reducing the workload on customer success teams.
Implementing a feedback system with thumbs-up and thumbs-down options allowed us to continuously refine the chatbot based on real user input.
Feedback at the Forefront
