Trustworthy AI is spoken about often. That’s because no matter how advanced and clever your conversational AI is, if it’s not trusted it won’t be used. And if it’s unused there’s a sizeable business impact.
Why trust erodes
Trust erosion happens over time. Poor AI experiences that build up over time eventually force customers to no longer want to use it and so they seek out alternatives. The customer relationship erodes.
This is a problem because it undermines the ROI of your AI strategy and has consequences not just financially but also operationally and reputationally.
When a customer interacts with a conversational AI chatbot, voicebot or virtual assistant they are expecting convenience, a simple, fast, and helpful exchange. When they get friction the trust erosion begins. The bad experiences are remembered – much more so than good experiences that meet expectations.
It’s the little things, that happen over and over again that erode trust:
- The incorrect answers given with certainty
- Repeatedly told “I didn’t understand that”, and unable to progress
- Being asked to repeat information already provided
- Stopped or blocked from accessing a human agent
- Tone mismatch, e.g. serious situation is met with a cheerful tone
Consequence of trust erosion
The consequences of customer dissatisfaction arising from their AI experiences, where they end up not using it, undermines the ROI of any AI strategy. The negative business impact when trust goes is caused by:
- Escalations to human agents increase – resulting in a rising cost to serve
- Ongoing AI investments no matter how large cannot fix the declining containment rates
- Brand perception damaged over time
- Customer disengagement leading to loss of customer lifetime value
Customer behaviour signalling trust breakdown
Customers who no longer have trust in AI, change behaviour and there are several metrics that will highlight that along with the obvious usability issues. Behaviours that point to trust erosion include customers:
- Going straight to requesting a human agent
- Changing the communication channel e.g. chat to phone
- Abandoning the conversation
- Being more abrupt or emotional in their language
- No longer using self-service options
Building trust credit
Trustworthy AI should be a priority for every organisation and top of the conversations you have with your AI provider.
Customer trust has the greatest impact on your business and with it you can shape your customers’ future behaviour; influencing lifetime customer value, and immediate financial opportunities.
Organisations must start looking at ‘trust scores’. Often you find conversational AI success measurement is focused on containment, deflection, and resolution time. However, they don’t tell you how much confidence customers have in your AI.
Important metrics, not individually, but taken together will tell you how much confidence customers have in your AI:
- Time to requesting a human
- Conversation drop-off
- Repeat contact after bot interaction
- Self-service adoption vs self-service avoidance
These can act as directional signals to understand if trust is eroding and enable appropriate action before irreversible damage is done. If trust is lost, to recover it is extremely hard. It is far easier to maintain trust – and that requires a proactive approach.
Trust erosion can be an expensive hidden cost of conversational AI if it is not considered from your initial conversation with your AI provider, embedded into the solution design principles enabling proactive tracking of customer behaviours.
Start tracking conversational AI trust and enjoy increased customer satisfaction, revenue, and lifetime customer value.
