Crafting a winning CX strategy for a customer service voicebot

The client sought our assistance to understand what constitutes an excellent customer experience (CX) for a customer service voicebot in their industry. They aimed to move beyond technological focus and explore the user-centric aspects of voicebot usage as they were conscious of the competitor telco company's solution, considered a widespread failure by customers.

• Conducted desk research focusing on best practices as well as psychological and user experience literature.

• Performed usability tests and in-depth interviews with 20 customers to assess competitor voicebot solutions.

• Developed an insight-led CX strategy for designing the new customer service voicebot.

• Literature Review: Explored voicebot and human-machine interaction literature.

• Usability Tests: Conducted scenario-based tests to evaluate the user experience of the competitor solution and understand detrimental effects on customer experience.

• Coding Sheet: Tracked key metrics such as customer-bot interactions and misunderstandings to quantify qualitative data.

• Storytelling: Used extensive storytelling in the strategy document to help empathise with voicebot users and get buy-in from stakeholders.

• Video Summary: Compiled insights from usability tests for impactful presentation.

• Our analysis showed that during user testing the competitor telco company's voicebot solution successfully recognised the customer intention at first attempt in only 42% of the cases (14 out of 33). Interestingly, however, our participants’ assessments of whether they felt they achieved the goals our scenarios gave them were not dependent on their assessment of their interaction with the voice bot.

• We coined the term 'voicebot-paradox' to describe users' perception of high responsibility but low control while using voicebots. We pointed out that this internal conflict in the users was rooted in not just technological limitations (i.e. limited 'intent recognition') but also psychological factors: user anxiety and interaction uncertainty.

• We recommended strategies for building a voicebot with better CX, including mental model development and conscious design of conversation flow, supported by a comprehensive compendium of insights.

• The playback of our insights and the CX principles we developed, supported by scientific evidence, best practices and short video snippets from the interviews, achieved astounding success. Our client managed to use this to convince internal stakeholders to give more weight to user-centric considerations during the development phase; subsequently recognising our project as the 'Best spent money this year'.

1) Always pilot new approaches (and often old ones too). Quantifying elements of qualitative interviews is tricky, as it requires even more attention to the details and adherence to the interview plan from the researchers. My key learning is that whenever using new research methods at interviews, always do a test-run with all the researchers involved to clarify any questions about the method and prepare for any potential curveballs.

2) The more junior the team, the more time needs to be allocated to alignment. It was my first time managing a junior colleague who had just started his service design career, which required more time and supervision than anticipated. Mentoring and teaching colleagues with less experience is crucial for the organisation’s resilience, but it shouldn’t take away time from doing high-quality work. Next time, when preparing to work on a project with inexperienced colleagues, I would advocate for allocating more time to mentoring them.