July Success Spotlight: Breaking Down Answer Bot
By Ashley Mo
“Why use answer bot?”
If Helpshift Customer Success Manager Ashley Mo was an answer bot, this would be one of her most frequently asked questions. The short answer is that answer bot is a new customer service tool that refines the self-service process, makes it far more user-friendly — all while deflecting tickets. For this month’s customer success spotlight, though, Ashley has provided in-depth responses to the related questions below.
1. What actually is answer bot, how does it work, and what are the main benefits?
You are probably thinking “what can answer bot do for me” and “how can it improve my business?” Well, there are a lot reasons why you should start using answer bot, so first let me provide additional context about how answer bot works.
Answer bot is an AI-powered bot for both your web and in-app messaging user experiences that automatically and instantly responds to the end user’s initial message. Answer bot responds with up to three suggested relevant knowledge articles that help the user self-serve. The goal here is to reduce ticket volume so agents have less backlog.
By using a combination of Natural Language Processing and Machine learning, answer bot both detects the language and finds the articles from the knowledge base that are most relevant.
When you manage a support team, it’s also important to consider answer bot from the perspective of your two primary stakeholders: your end users and your agents. You want your end users to be able to receive a response and have their issues resolved as quickly as possible, and you want your agents to likewise be more efficient. Answer bot accomplishes both of these goals.
That’s because your end users will receive an instant response 24/7 (regardless of working hours), and your agents will be able to focus on higher-level support inquiries. In terms of ROI, brands that use answer bot will see improved ticket deflection of up to 20 percent.
One additional benefit is that answer bot is multilingual and available in 19 languages. This allows our global customers to support their users from all over the world.
I often encourage my customers to think ‘big picture’ in terms of scalability for their support team. Answer bot allows teams to scale in a cost-effective way because teams can maximize productivity of their existing agents as opposed to increasing headcount.
2. How do I prepare for answer bot and use it most effectively?
Answer bot suggests FAQ articles from your knowledge base. For that reason, it is important to ensure that your knowledge base is in good health. In order to properly evaluate the efficacy of your existing knowledge base, follow these three tips:
- Revise your FAQs by looking at the way your users frame their questions. Looking at the historical data, you can see how your users frame their questions. Rework and supplement your existing FAQs based on this data, and add keywords to match the common queries (even those that include misspellings). You ultimately want to make sure that the knowledge articles are up-to-date and cover a robust array of topics.
- Improve/assess your FAQs by evaluating key metrics. At Helpshift, we provide our customers with analytics such as “FAQ analytics” to help provide them with insights into the performance of their FAQs. There is an FAQ Deflection report that provides metrics on successful and failed deflections, including a section on issues that were created without prior FAQ views. This visual guide will provide insights about the percentage of issues created when the answer bot has been displayed, but the user has not clicked on it. The “FAQs List Report” additionally provides a comprehensive list of all FAQs in your support portal, so you can focus on FAQs that have a high percentage of dislikes. “FAQ Search Terms” is another report that allows you to look for the most popular search terms that have returned with no results. Incorporate these keywords into relevant FAQs, or create new ones to meet this demand.
- Frame your “greeting message” to help users better explain their questions. Include phrases like “can you please describe your problem?” and “don’t use hi,” etc. to help your users best help themselves.
3. How can answer bot deflect more tickets and how do I measure its performance? Monitor these reports every week and keep in mind that full optimization may take regular iterations. Helpshift also provides an advanced analytics report called Power BI that allows teams to track overall effectiveness of answer bot. This information can be used to view:
- The number of times answer bot was used over a given period of time & how many issues were deflected by answer bot directly.
- The percentage of FAQs that has been suggested and read over time, as well as the efficacy of FAQ titles in terms of compelling users to open them, and the FAQs that have the highest deflection failures (if you’re interested in details pertaining to answer bot analytics, read more here).
For more information about how to get started with answer bot, please reach out to your customer success manager or request a demo.