Many leaders in business today think about #AI in the form of either Bots, #GenAI, or a silver bullet to reduce costs. My recent experience has shown me that, despite how much people may know about AI, they often pigeon-hole their thinking, applying "blinkers", and not seeing outside the box of their own understanding.
On a recent project programme there were several use cases for AI in different forms, achieving vastly different objectives. Let's unpack what that meant and what the output looked like.
Understanding the customer
When it comes to #customerservice, being able to be proactive based on current call topics, cases logged, and call volume, has not been a particularly tangible goal.
It is often just out of reach, requiring a lot of time and financial investment to get right. The technology is available, but it is challenging to find a) budget, and b) capable engineers specialising in that area, who can harness the potential.
Call volumes to the call centre were high. We knew that we needed to increase call deflection by empowering customers to solve problems themselves, and to push requisite information to them; rather than them looking for it, thereby reducing the need to call.
Customers need to feel that a brand understands their needs, and in these instances, where volumes are high, we needed to create content where it didn't exist, or repurpose content where it was available so that customers could use it effectively.
We knew that we needed to produce content faster than in-house teams allowed, especially given the influx of calls. The siloed nature, and ownership of content and processes, made this very difficult.
The Challenge
So here is the cause. In almost every single organisation, there are large tranches of duplicate content, resulting in inaccurate content or answers as there will often be discrepancies.
Enter AI, stage left.
Imagine trying to ground your AI on that content and use generative AI to provide answers to questions. Doesn't seem like a great idea, does it?
Now imagine people attempting to find content, not knowing what to trust or where to find the content. Searching and returning noise, or using an AI bot or similar and receiving an inaccurate answer, is less than optimal and hugely frustrating.
What do business users inevitably land up doing? Creating their own content, saving a version to their desktops, creating their own maintained knowledge base. All of these things contribute to the ongoing degradation of content quality, accuracy, discoverability, relevance, availability and accessibility.
So where to from here, how is it even possible to overcome this if AI is not the silver bullet?
Well, that's a journey.
Part Two we investigate how we needed to approach the issues at hand.
Comentários