Our latest tech revolution is delivering productivity gains to support and operations teams across the globe. While this benefits employees by reducing cognitive load and streamlining operations, it’s also at risk of adding friction to the holistic customer experience asking users to engage wth a chat for just about everything.
LinkedIn is awash with posts promoting prompt engineering, prompt-as-UI productisation, and our need for chat-based workflows. Within our organisations, this makes sense. We want our teams who already operate through chat to do it better, faster, and with less effort. For our customers though, does it make the same amount of sense?
AI has become unavoidable
While we’ve swiftly slapped the somewhat incorrect AI label on our latest brand of technological advancement, it is undeniably the start of a new paradigm. Like the advent of the new age of technology in the early 2000s, we’re riding the elevator straight to the top. Gone are the days of needing to take meeting notes, or do those pesky human tasks, checking emails, replying to customers, or even writing code - there’s a prompt for that now.
The achilles heel of the current ecosystem, however, is that it requires a small mountain of written context and a well trained model. Under the hood these models operate by assigning numeric values to groups of words (if you’ve had the pleasure of watching Apple’s Severance, yes, the scary numbers are real), mapping them relationally, and finally, constructing a statistically predictable reply. It’s a bit more complex, but you get the gist.
This critical mass of algorithmic brilliance has brought us to a junction, however, where context is king, data is paramount, and everything requires a prompt. But does this deliver what customers now and in the future actually need, let alone want?
We’re pushed towards conversation
There is nothing wrong with conversational support. As an end-user it’s my go-to preference when I have the time to deal with an issue in the moment. When I’m running between fires though, sometimes I just want to hit send on an email and know there’s a human picking it up on the other side. Customers are a bit the same, whether they’re wanting a face-to-face call or an email, conversational channels shouldn’t be our only engagement surface.
If you’ve ever watched a sci-fi movie and wondered why we don’t have holographic and motion-capture interfaces in the workplace, the reality is simply that they’re too high-effort for people to use. Who really wants to type on a keyboard that has no tactile feedback and your fingers can phase through? Who wants to be waving their arms around for 7h a day to operate machinery? It’s simply not an effective experience, and we love easy-to-use experiences.
AI in its current form has much the same feel. While it can boost productivity and efficiency, having a big “do the thing” button in any case would be just as effective as a conversational prompt. So why are we allowing our experiences to be pushed to a chat-first reality?
Refocusing the customer experience
While conversational support will continue to have a time and a place and AI can cut costs and free our teams up to work on more meaningful work, we shouldn’t lose sight of the overall customer experience.
Before we get to productising AI, we should also remember there’s an easier avenue where we embrace it as co-pilot instead of replacement. Implementing it into internal workflows to suggest responses or surface context is a quick first-step to build team trust and sanity test reliability before you’re unleashing it on customers.
Once we’re ready to put it in the pilot seat, we can also apply it to proactive approaches before making it directly customer-facing. Using it as connective tissue in workflows using behavioural data and system state as context. What better solution than pre-empting customer needs before they’re even reaching for a messenger or email form?
Ultimately, in low-complexity environments AI will win-out when it comes to the scalability of supporting and serving our customers, but it’s critical that we don’t let this current evolution allow us to forget the lessons we’ve learnt over the past two decades. Our customers often need us to meet them where they are, not where we want them to be. So let’s keep doing that. And use AI to help us do it in the best way.