ENHANCEMENT

Helping users help themselves:
Support hub in the purchase journey

Digital only Digital-first

To broaden Sonnet's omnichannel strategy, the team held an all-day hackathon to collaborate, brainstorm and problem-solve. The team dived into the end-to-end journey, and one of the ideas that came out of the session was to improve and expand both inbound and outbound channels for Sonnet. We asked ourselves:

• How can we leverage existing data to deliver personalized quote flow experiences that increase policy binds?

• How do we maintain our commitment to a digital-first experience while providing enough support for customers throughout the purchasing process

As we deep dived into the idea, it became evident that customers navigating the quote flows often get stuck. Many find themselves confused about what is being asked, why they need to provide certain information, or how to get help when they encounter issues. This confusion leads to drop-offs. To address this, we need to rethink how we guide users through the process., how we should offer better assistance and make the overall experience feel more intuitive.

What type of support do the customers need?

I began the work by going through our chatbot platform to identify the questions our customers ask, what they need help most with, and determining pain points. Here's what I discovered:

• People seek guidance. Regardless of where they are in the quote flow, they are looking for guidance in selecting the best coverage for their needs that is both comprehensive and cost-effective

• Our customers are looking to talk to Insurance Advisors

UX strategy & proposal

To address our customers' need for guidance throughout the quote flow, I proposed a more intelligent support panel experience that includes:

• Data-driven insights (powered by AI):

  Leverage behavioural and vehicle data to surface helpful tips.
  Example: Honda Civic drivers like you often choose Sonnet Shift and select these coverages.

• Quote flow guidance:

  Provide clear directions that explain what’s coming next and what customers can expect.

  Example: What you can expect on the next page: For the next step you will need information about additional drivers. You will need their date of births and driver license details.

• Relevant FAQs:

  Include frequently asked questions tailored to the specific step the customer is on, minimizing the need to search elsewhere.

• Clear contact options:

  Provide easy access to real-time support via phone, live chat, or email.

Build → Ship → Learn → Repeat

• Build: The work started with collaboration between design, content, and development. We worked together to outline the user experience, set goals, and confirm technical feasibility. I worked closely with our lead developer to ensure we could create what we designed within the timeline. As with any product launch, trade-offs were necessary. While we originally planned to include AI powered support suggestions, we learned that integrating AI would significantly delay launch. The business decided to prioritize launch to market, agreeing that some support for the customers was better than none. This allowed us to move forward with a functional MVP and planning for future enhancements. Key obstacles we faced:

• AI data integration would have delayed the launch

• Technical complexity in integrating our chat vendor into the support hub

• Challenges pulling and customizing content from our CMS into the panel

• Ship:
We launched an MVP product that delivered relevant in-flow support during the quote and purchase journey. This allowed us to gather feedback, monitor engagement, and validate that we were solving the right problems. The initial release included:

  • A collapsible support side panel

  • Tailored FAQs based on page context

  • Step-by-step quote guidance

  • Easy access to contact us options

• Learn: The initial phase began with reviewing multiple Fullstory sessions to understand why customers were abandoning their quotes. We noticed a consistent pattern: users often dropped off when asked to enter driver’s license details, either their own or someone else's, or when asked about detailed property information. In many cases, it was clear they simply didn’t have that information readily available. To address this, I began by conducting competitor research, analyzing user journeys, and outlining key technical considerations. I also developed a lo-fi design proposal to explore how a “Save & Continue Later” experience could work within our flow.

  • Saw clear opportunities to included dashboard-specific FAQs in the dashboard to surface relevant content (e.g. moving, policy changes)

  • Inconsistent provincial targeting. For example, a New Brunswick customer received Ontario focused FAQs

  • A few UI issues were identified that had slipped through QA and needed refinement

  Key wins:

  • Users frequently clicked on the Need Help? tab and often left the panel open throughout their session

  • Users engaged with FAQs, indicating relevance and value

  • Many clicked on the Eligible Offer, showing interest in available promotions


  • These insights gave us a strong foundation for prioritizing improvements and validating that our MVP was delivering meaningful value.

• Repeat:
With these next steps we aim to improve the user experience:

  • Add dashboard-specific FAQs to reduce unnecessary chat volume and support common tasks like moving or policy changes.

  • Localize FAQs by province to ensure users receive content that matches with their location (e.g. New Brunswick vs. Ontario).

  • Fix minor UI bugs identified during Fullstory review and QA gaps.

  • Explore AI-driven content recommendations for future iterations.

  • Continue to monitor user behaviour to collect data for future optimizations and measure the impact of the changes.

Key learnings & business impacts

By launching an MVP produt, we validated assumptions, uncovered usability issues, and identified enhancement opportunities. Shipping the MVP led to reduction in support volume as users began leveraging the contextual FAQs during their quote process. Additionally, the MVP laid the groundwork for future enhancements like personalized content. Some key takeaways:

• Starting small allowed us to learn quickly and iterate with confidence

• Close collaboration across design, content, and development helped balance user needs with technical constraints

• Real user behavior was essential in shaping the roadmap and refining the experience

Results: Best SQ2Q Yet 🥳

Post launch, Sonnet has seen an increase of almost 3% for auto and a slight better than 3% improvement in Property for the SQ2Q (start quote to quote). If we can sustain these numbers it’s the best numbers of SQ2Q the business has ever seen.