5 Steps to a Successful Launch
Chatbots help employees and customers find the answers they need by offering instant self-service support. However, poor implementation often causes negative user experiences, which could affect your entire support structure.
The more you use and develop it, the more Capacity will take repetitive tasks off of your team’s plate. Conversational AI is a powerful tool that can make work better and more productive for everyone—if used correctly.
Read below for 5 strategic principles to ensure your success and avoid common mistakes:
1. Set Goals and Create a Dedicated Chatbot Team
Manage your chatbot with a dedicated team and define a strategy. We see that implementations succeed most often when they have a project manager, use subject matter experts, and collaborate with Capacity’s Customer Success team. Use this stage to build a scalable foundation, then plan to expand.
Define what you hope to achieve with your chatbot. Shorter response times? A certain deflection rate? Set reasonable goals from the outset and design your strategy from there.
Assemble an internal implementation team to manage the chatbot. Pull in other teams for collaboration and feedback. Take advantage of your team’s strategy and development skills, as well as subject matter expertise from agents. Your Customer Success Manager will also help guide you through this stage.
Ask for user feedback early and often with our built-in surveys and reactions. We recommend gathering feedback within two weeks of launching Capacity. Once you make improvements, announce new changes to your users with our Articles feature. Rinse and repeat.
2. Prioritize User Experience, Always
At the end of the day, your chatbot needs to be useful—or it won’t be used. There are a few things to keep in mind when designing your chatbot and helpdesk integration.
Show off what your chatbot can do. Suggest questions that users can ask and use visuals to inspire interaction. For example, in your teaser and welcome message, include sample questions such as, “What’s your pricing?” or a small image of a lock to indicate password reset capabilities.
Keep humans in the loop. There will always be questions the chatbot can’t answer. Offer escalation options so that users can reach your team when they really need to, via helpdesk tickets or livechat.
Avoid page pushing. Users want to get their answers in one place. Limit links and other external resources by building actionable answers and processes within the chatbot itself.
3. Start Small, Then Expand
When using new tech, it’s important to start small. Be patient with your implementation and stick to built-in features at first. After you’ve learned which features work well for your organization and users (and which don’t), start to customize as a next step.
Create a good first impression with your chatbot. Use out of the box capabilities and avoid the extra complication of optimizing at first. A robust knowledge base, with accurate and up-to-date information, is key to building user trust and encouraging engagement. Your chatbot should first and foremost be a useful point of contact for users.
Support your workforce’s existing skills. Focus on the tools your team knows how to use, and integrate with those first. Your Customer Success Manager is always available for help and advice, so make use of their expertise!
Use the data and experience you already have by including your agents in the process. Since they work with your users every day, they know best which questions are most asked and why. Ensure agents are trained to update and maintain the knowledge base; agents can simultaneously resolve tickets and update knowledge within the ticketing system. Review nuanced issues as well to see if AI could handle them.
4. Plan a Practical Automation Rollout
Capacity’s chatbot does much more than answer questions—but so many features can be intimidating. When implementing end-to-end automation, plan several phases that center user experience.
Start off with your top 10 questions and requests. Things like “When are you open?” and “What is your mission?” are good places to start. Ensure there is a path to escalation for nuanced issues, like a form fill or livechat routing.
Reduce repetitive work for your agents. Build human-handled requests, like lead qualification and customer information collection, into the chatbot. Knowledge articles can be very useful here as well.
Adapt and overcome! Keep track of your analytics to improve existing features and experiences. Prioritize making a few small improvements on features with high success rates. Once you are fully optimized, try out one or a few new features at a time. Customer Success can help you choose which features to try next and support you as you implement them.
5. Look at the Whole Picture of User Success
We love a high deflection rate at Capacity, but chatbot success involves so much more. To get a bigger picture of how your chatbot reduces support costs and increases ROI, look at these key metrics:
Consider balanced interaction success. Compare metrics across interactions and performance, such as intent recognition, positive or negative user feedback, number of interactions per use, and resolution time.
Keep collecting feedback. Does the chatbot actually improve the end user’s experience? Is it failing to meet user expectations in any way? Continue to optimize your no-match and feedback tickets; update your knowledge base and be sure to respond to chat users. Build confidence in your bot, keep your users engaged, and promote overall success.
Don’t discourage automation with unfair KPIs. Use the chatbot to streamline the agent’s workload, not increase it. Balance automation completion metrics and human-handled completion metrics equitably and with context.
AI self-service can supercharge your support capabilities and reduce unnecessary work. To get the most out of your investment, carefully plan and implement your strategy using the tips above.
Don’t forget, our Customer Success team stands ready to help whenever you need it.
Now go forth and maximize your Capacity!