I recently moderated a panel discussion titled Supercharge Customer Service and Agent Productivity with AI.
It was fantastic! Unfortunately, the webcast ended before our industry experts could answer all the attendees’ questions.
Fortunately, one of our panelists—Pete Weaklend, Managing Partner at PeakCX Partners—agreed to respond to the questions in writing! We decided to share his answers (I chimed in just a little) here so the entire 7.ai™ community could benefit from his insights.
Two suggestions before we get to the main event ….
OK, without further ado ….
Speech analytics. When you use AI to ingest and analyze 100 percent of contacts, speech analytics is a great way to ensure customer interactions are always handled professionally and without bias. Compared with traditional quality assurance, which typically reviews a random sampling of 1-2 percent of calls, speech analytics can fully monitor all calls to ensure quality, professionalism, accuracy, and compliance across the board. Sentiment analysis can quickly determine problems that can be coached and corrected. Coaching sessions can be recorded and analyzed in the same manner.
Despite the plethora of vendors with seemingly similar offerings—there are differences. The best vendor and products for your organization, of course, are fully dependent on your unique business characteristics and requirements. And that means your decision must follow the data. First you need to assess the contact drivers, process/support tool inefficiencies, and areas of high associate and/or customer effort or dissatisfaction. These assessments typically produce several areas of “low hanging fruit” that will carry the strongest business case for return on your investment; start there.
Of course, in your initial discussion with the vendor you’ll have to cut through the “presentation layer” to get to real proof of capability, which can be a challenge. Companies with truly effective offerings will encourage you to talk with their customers so you can validate their claims and solidify your business case.
There are many mature AI technologies. Again, what’s right for your company depends on its specific business needs, your support channel structure, the size of your support organization, and your goals.
No surprise, but speech analytics (and associated processes) is likely to have the largest and earliest impact. When implemented properly, this technology quickly (and continually) improves overall customer support and satisfaction, boosts ROI, and lowers operating costs.
While AI capabilities are always evolving and improving, other currently mature offerings include chatbots, conversational AI, AI-driven noise cancellation, and robotic process automation (RPA) that automates basic workflow tasks.
The maturation of agent assist capabilities may supply the greatest opportunity to improve customer service and boost agent productivity in the years ahead. While this technology exists today—AI can already assess customer interactions in real time and proactively provide agents with most relevant next workflow steps and knowledgebase information—the capabilities of these tools to empower agents, ensure accuracy, and boost efficiencies will continue to rapidly improve. Of course, all of this translates to a better customer experience.
Integrating properly designed AI support capabilities (e.g., chatbots that seamlessly escalate to a live chat agent) within your website improves e-commerce two ways: It guides customers to the right product(s), thus enhancing their understanding of product benefits and increasing sales conversions, and it helps customers complete their purchases. Bots designed to trigger a conversation based on certain business rules (such as web page flow, time on a page, etc.) make it likelier customers will engage.
Properly designed and integrated AI solutions aligned to specific, data-based business use cases drive increased customer satisfaction. Poorly designed solutions reduce satisfaction. It’s as simple as that!
Stop us if you’ve heard this one before: You’ll achieve excellent customer support only by designing and implementing custom solutions based on your specific business data and customer needs. Questions with singular answers or answers that can be clearly derived from a data set, as well as transactional and process-driven workflows, are successfully handled by bots. More complex customer needs should be driven to a live agent. The most successful and efficient models leverage integrated solutions: The bot handles upfront data capture through to resolution where possible, but quickly and cleanly hands off to a live agent (along with relevant data) once the inquiry exceeds its capabilities.
Call trends during and after the pandemic are highly specific to the business/industry. For some, overall business and support volumes have increased during the pandemic (e.g., online retail), while others have decreased due to business decline (e.g., the travel industry).
The primary disadvantages involve selecting the wrong tool for the job, or in poorly designed, implemented, and refined use cases not well aligned to your specific customer support needs. Always integrate AI within your overall support strategy—and never prioritize efficiency gains (or cost reductions) over improving the customer experience. The two can and should always go hand in hand.