Why software support AI chatbots should supplement, not supplant, human experts [Q&A]

Many enterprises have begun to rely more heavily on chatbots to provide software support, and this often means customers find it hard to get in contact with an experienced, human engineer when they encounter an issue that they need help resolving.

While this might save costs in the short term, it can seriously damage the company’s brand in the long term. We talked to Craig Mackereth, EVP, global service delivery at Rimini Street to find out about the overuse of AI chatbots for enterprise software support and ways that vendors could use generative AI to actually improve the customer experience.

BN: You argue that enterprise software vendors rely too heavily on AI chatbots for support. Why?

CM: We've all seen the headlines and heard anecdotes about the power of generative AI. It's clearly a powerful new technology that provides business value for all sorts of applications. It can be used to help improve customer service, but it’s not a quick-fix panacea, nor should it replace humans entirely for many experiences. It seems some enterprise software vendors have been relying too heavily on AI chatbots for support, and that can be a major misstep. IT teams with mission-critical issues will want to know they can work directly with a human expert to solve their complex problems, and some vendors are making that increasingly difficult in the name of cost efficiency.

By adopting AI chatbots for support, some vendors aim not to give their customers a better experience, but to cut costs by minimizing the time they have to interact with their users. They refer to it as 'call avoidance' or 'call deflection' in the industry. Today, software customers are finding it harder and harder to contact an experienced human engineer when they encounter an issue that they need help resolving.

In replacing support experts with bots to cut costs, we see a race to the bottom that disregards customer concerns instead of addressing the root causes of problems. It's a myopic strategy that can ultimately lead to both customer frustration in the short term and damaged loyalty in the long term. And that could have a significant impact on these vendors’ businesses.

Of course, vendors should not ignore AI chatbots altogether -- that would be just as short-sighted. Instead, deploying them in the right place and time, and remaining focused on improving the customer experience by supplementing -- not supplanting -- expert human support, can help ensure the customer experience remains paramount.

BN: Are there other concerns about AI chatbots?

CM: First, the reality is that enterprise software support is rarely easy. But it's not going to get any easier by preventing customers from engaging with human support experts. In particular, the demonstrated tendency of AI chatbots to hallucinate represents a potential landmine. Would we want to rely on a chatbot that makes things up in response to customer inquiries?

Second, generative AI technology is still in its infancy. Choosing winners and losers at this early stage is virtually impossible. The industry is evolving rapidly, best practices are gradually taking shape, and a wave of consolidation will surely be coming (perhaps at an unprecedented level). Would you want to gamble your business future on a chatbot support option that could evaporate or be rendered obsolete by the next disruptive innovation?

And third, there's also the data privacy issue. Customers need to be assured that the cloud services that make AI chatbots commercially viable are not being used to mine their data to feed the insatiable training needs of Large Language Models (LLMs). So far, we’ve seen some of the more prominent chatbots offering responses that have been exposed as regurgitating info from competitive services. Just imagine the reputational damage if your AI chatbot provided Bank A with proprietary information from Bank B!

BN: What's the best path forward?

CM: AI is here to stay, and adoption will inevitably continue to accelerate. New, innovative AI-powered processes and services are being developed and will shake up the status quo whether you’re on board or not. For organizations that engage with enterprise software vendors leveraging generative AI for support, here are three critical questions to take into consideration:

• Where's my expert? IT teams with mission-critical issues will want to know they can work directly with a human expert to solve their problems. Ensure you have access to such experts, and not just stochastic bots. Make sure that the commitment to provide a backup human expert is guaranteed with a response time based Service Level Agreement (SLA)
• Is my IP safe? Understand how the vendor is gathering your data to feed its LLMs. Make sure there are solid agreements and guardrails on what happens with your data. You don't want to expose your -- or your customer’s -- intellectual property.
• What's my recourse? For IT teams encountering generative AI pitfalls -- fabricated and/or biased information, question misinterpretation and inconsistent answers, and a lack of empathy -- what's the path away from peril? Look for recourse for chatbot frustrations and check social media to see how others have found the experience.

Image credit: phonlamai/depositphotos.com