AI-assisted bots are reshaping customer digital engagements – and chatbots are leading the way. But customers today expect more than just a new interactive FAQ. They want the bot to know them, engage them and make decisions based on their specific customer history.
My organization works with company’s contact centers to help them design and build effective digital engagements using the most acclaimed customer engagement application on the planet – Salesforce. But the concepts we use to support customers in their digital journeys are relevant – no matter what application is chosen.
In our recent Town Hall session we shared four ways that today’s bots must support customers.
-
Understand – Bots should have the ability to gather data in new ways, to understand the customer better – often before the live interaction even begins.
-
Interact – Chatbots and virtual assistants should be able to deflect inbound customer interactions and automate processes and repetitive tasks.
-
Answer – Bots have the ability to unlock knowledge, to recommend answers to complex customer requests via self-service options.
-
Augment – A successful bot is designed to support human interactions, not just replace them
-
Guide – Bots should be able to guide both the customer and the employee, via predictive and machine learning models, to present the next best action to the customer.
In a recent report, Forrester reported that “while 90% of customer service leaders agree personalization is core to the future of automation, 63% of customers will leave a company after just one poor experience, and almost two-thirds will no longer wait more than 2 minutes for assistance.” It seems companies must do more than just deploy a bot – they must have a plan to build a foundation that allows the bot to learn and improve with every interaction. We believe there are six key elements to any well-designed bot:
-
Assess – Review your current service use cases. Assemble a team. Establish clear KPIs.
-
Build – Leverage your existing data and build clear workflows.
-
Test – Test your conversation flows; understand how they model the expected performance.
-
Deploy – Setup a launch plan and ensure your bot will perform as expected. Train your live agents on what to expect and what the bot is capable of doing (and not doing).
-
Monitor & Measure – Track your usage metrics, create usage performance models; analyze the data to improve processes, conversations, and expected usage models.
-
Iterate – Re-train your bot based on usage data, successful and unsuccessful flows/answers. Adjust your conversation flows accordingly.
-
Repeat
In working with contact centers throughout the world, we’ve learned that the success of any digital strategy depends more on a clear strategy, a strong implementation plan, and designing a bot that can augment human interactions.
Based in Memphis TN, contact center industry veteran Bob Furniss serves as a CCNG Academy member, working as a leader in Slalom’s Global Salesforce Practice, focused on all-things related to contact centers.
