Managing Simplii’s Chatbot Experience

Managing Simplii’s Chatbot Experience

At a Glance

Early 2023, Simplii Financial launched its Chatbot on to transform its customer support experience. This aimed to introduce a new way of how customer inquiries and support interactions were handled.

Here are some of the milestones of success:

👍  25% increase in Automated Resolution Rates

🤖 100+ intents to broaden chatbot competence

💸 $15,000 cost savings by reduced human intervention

Deep Dive

(1/4) Problem

In an increasingly digital landscape, Simplii Financial faced the challenge of high customer service handoff volumes, leading to longer wait times and decreased customer satisfaction.

The existing chatbot was failing to efficiently route customer inquiries, causing frustration and a negative impact on the user experience.

As the financial industry moves towards round-the-clock customer service, Simplii recognized the need to leverage technology to improve accessibility and response times.

The chatbot (powered by Ada) was the first point of contact for many customers, but its limited capabilities meant that too many interactions required human intervention.

(2/4) Insights

Through a data-driven approach, here are some key insights we found to optimize the chatbot’s interaction with clients:

  1. 🧠 Deepening Client Engagement through Chatbot Intelligence
    • Clients desire comprehensive engagement that go beyond interacting with the chatbot for transactional support.
    • Integrating regular feedback loops into the chatbot’s development cycle ensure its capabilities evolve in alignment with client expectations.
  2. 💬 Navigating Customer Interactions
    • We identified high engagement for basic product inquiries and a preference for clickable options.
    • There’s a seasonal rhythm to our interactions, marked by a surge in queries from international students and heightened activity during GIC renewals and promotional periods.
    • These patterns not only informs us about when and how our client seek information, but highlights the opportunities for tailored interactions / communications expected from our chatbot.
  3. 🔎 Bridging the Query Resolution Gap
    • A comparative analysis of the chatbot’s query resolution success versus escalations to human agents uncovers significant discrepancies.
    • This gap reveals specific areas where our chatbot falls short, offering a clear directive for targeted improvements in its understanding and response mechanisms
  4. 🤖 Elevating Understanding with Advance Natural Language Processing (NLP)
    • Delving into the effectiveness of our chatbot’s NLP capabilities, we’ve pinpointed a a crucial need for more sophisticated NLP techniques and its limitations.
  5. 📚 Filling the Knowledge Void
    • A systematic review of unresolved queries has laid gaps in our chatbot’s knowledge base, directing us to continuously refine and iterate on chatbot content. This ensure the chatbot caters to a broader spectrum of inquiries with precision and depth.

(3/4) What I Did Next

Given the insights we gathered, we explored a series of targeted initiatives to elevate our chatbot’s performance and relevance to our users’ needs.

  1. 📈 Integrating Feedback with Analytics for Continuous Improvement
    • Established a system to capture and analyze real-time user feedback, integrating it with existing chatbot analytics.
    • Utilized data analysis tools to sift through large volumes of user interactions, identifying common queries and unresolved issues.
  2. Streamlining Intent Development for Agility
    • Created a workflow for quick development and integration of new intents. This involved setting up a streamlined process for scripting, testing, and deploying chatbot responses.
    • Collaborated with technical teams to ensure that new intents were technically feasible and aligned with the chatbot’s existing architecture.
  3. 🔎 Employed a Prioritization Matrix for Strategic Focus
    • Developed a prioritization matrix to rank new intents based on factors like frequency of the query, potential impact on user satisfaction, and feasibility of implementation.
    • Regularly reviewed and updated the prioritization matrix to reflect changing user needs and business objectives.
  4. 🤩 Optimizing User Journeys with a Top-Down Intent Strategy
    • Developed and implemented a top-down intent management strategy, prioritizing queries based on their frequency and impact.
    • Reconfigured the chatbot's decision trees and response triggers to create a more streamlined and intuitive user journey.
    • Conducted iterative testing to refine these pathways and ensure ease of navigation

(4/4) Outcomes

👍  25% increase in Automated Resolution Rates

Our chatbot's ability to autonomously resolve inquiries saw a remarkable 25% improvement.

This leap signifies not just enhanced efficiency but a direct impact on user satisfaction, reducing the need for human intervention.

🤖 100+ intents to broaden chatbot competence

With the integration of over 100 new intents, our chatbot has grown significantly more adept at addressing a wider array of user queries.

This expansion reflects our chatbot’s enhanced relevance and adaptability, ensuring it remains a vital resource for users seeking support.

💸 $15,000 cost savings by reduced customer support volumes

Given these chatbot enhancements, this translated into tangible financial benefits of $15,000.

This underscores the dual achievement of elevating user experience while optimizing operational efficiency.