We’ve written about how AI use is expanding across the enterprise and early-adopters are leaving the competition behind.
We’ve shared how AI is a superpower for customer support execs facing the daunting task of delivering scaled, personal and cost-effective support.
Today, we debunk the AI myths that trip up these well-meaning leaders.
Trying to learn about and adopt AI against the backdrop of hype and over-promotion is hard. While solid technology is emerging and breakthroughs are happening, the noise from media, technologists and non-AI solutions scrambling to stay competitive is deafening. While many customer support functions are on the front lines of use, others are hesitant. Half-truths and sensationalized stories about AI keep some leaders sidelined and missing out on business transformation. But the right technologies do exist for enterprises’ most challenging issues. Don’t let fear, uncertainty or misinformation bring disappointment and limit your opportunity to give customers and employees an improved experience.
Below are six of the most popular myths, or misconceptions, about AI use in customer support:
AI Replaces My Team
AI augments human talent and brings new levels of efficiency, not obscurity to human involvement. Interactions will always be dynamic as customers bring ever-changing needs, wants, context and perspective to each exchange. From routing automation or routine answer delivery, to pattern and trend identification for next-step recommendations, AI simply improves rep interactions. Instead of replacing customer support, AI can help teams evolve from focusing on closing out tickets to delivering high-quality support – the tools simply reinforce process and workflow efficiencies.
AI is All The Same
Over-marketed technology confuses support leaders into believing that any tool sold as “AI” is actually AI. This oversimplification makes it challenging to understand what is being bought and its potential to deliver improvement.
True AI requires a level of system “intelligence”; that is, it can change its programming or responses over time based on real-world experiences. This is markedly different than basic data science techniques plumbed into available software that get lumped into the AI conversation. Within a customer support context, leaders will run across chatbots, virtual assistants or agents, predictive analytics and other integrated platform systems promising new ways to identify customer issues, learn from gathered data, offer real-time support, curate email, handle issue resolution, define patterns and behavior and drive personalization. As leaders explore options for an intended use case, it is essential to know what the AI inside the tool actually is.
AI is Too Expensive
With new AI platforms from Amazon, Microsoft, Google and others, plus the growth in open-source computing, AI development is quickly becoming democratized. This, and the maturing of AI technology sectors like machine learning, have made affordable enterprise AI a reality. In a customer support context, AI tools may be no more expensive, light or deployable than popular SaaS solutions. Many will actually improve human touch, experiences and revenue generation or cost savings at scale. As such, support leaders may find their colleagues in IT, finance and business systems behind AI initiatives.
AI is Set-It-And-Forget-It
Like the newest customer service rep, an AI tool also needs training for maximum effectiveness. The difference is that the AI never eats, sleeps or takes vacations. It learns, won’t forget and can intake or process data thousands of times faster than its human counterpart. But one cannot expect canned AI applications to be unleashed off the shelf and immediately overhaul customer support. Their effectiveness grows over months and years as they handle more cases or see more data. Some human backup is important, but this investment of time and availability pales in comparison to the benefits achieved. AI is smart, but not a panacea. Some learning is required.
AI Always Breaks
This myth is fueled by poor interactions with brittle “AI” bots (Note, the term AI is applied loosely here!). While AI tools will mature through training, relevant data feeds and basic error correction, they rarely require the vigilance some worry about. In most cases, concerns about AI maintenance come from bad experiences with scripted or “dumb” tools built with pre-defined conversational flows, huge decision trees or that are fed by keyword scripts based on assumed customer queries. Good, natural language processing technology doesn’t break or stall when asked something off script. A robust chat or text based AI tool will be built from word vector technology and should parse language to understand the intent of a question.
AI Disrupts Existing Workflows
Companies are wise to find AI tools that work alongside existing workflows and technology stacks. Fortunately, modern AI isn’t an “all or nothing” proposition – nor will it generally require massive re-architecture of systems or workflows. In most cases, AI tools are designed to meet users where they are. In the case of support, many AIs mesh with common rep activities to reduce issues of non-adoption. As well, the best AI have great user experiences so users don’t feel like they are interacting with a machine. Using an AI-powered knowledge base and bot as an example, the well-designed tool can be deployed within Slack or MS Teams and leverage content in live CRM or ticketing applications like Salesforce or Zendesk. The tool could learn from historical data in popular corporate messaging tools and reference data in other third-party knowledge bases like Confluence or ServiceNow.
AI myths hold too many support leaders back. Don’t let science fiction storylines and outdated industry examples slow the AI agenda. Talk to peers and learn from trustworthy vendors. These conversations will challenge common assumptions and thoughtless AI worries. In doing so, you can distill fact from everything else– and finally apply AI.