Should You Turn Supporter Care Over to AI?
May 1, 2023
Kevin Schulman, Founder, DonorVoice and DVCanvass
No is the short answer. The longer one, you should be using AI to significantly boost productivity and increase the number of successful outcomes.
It’s worth noting supporter care is more than donor care. There are plenty of charities (e.g., health, education) whose donors also make use of the program/services side of the house. This volume of inbound inquiry often dwarfs the donor related ones.
But just donor volume can be significant on an absolute or per agent basis. Heifer International, for instance, gets tens of thousands of inbound inquiries each year, backloaded at year-end and their agents often handle multiple inquiries simultaneously across multiple modes – phone, email, online chat, etc. It’s a lot.
Newer staff everywhere are often less productive and require more training and turnover tends to be higher. Supervisors can often spend 20 hours per week coaching lower performing agents.
Using some form of automation built on machine learning is standard fare in the commercial sector.
Some use GPT but if so, they fine-tune the model with their proprietary datasets specific to their call center domain – e.g., tech support or retail customer service. Why do this?
Same reason we train GPT for the copy we write, the generalized model is just that, general. No amount of prompting will work as well as giving the AI your samples/examples.
Doing this on steroids means taking all the call center recordings and email or chat exchanges and feeding them into the beast. Importantly, each call will also include key outcome data,
- how long the call took and
- whether it was successfully resolved
This means the AI is learning how the best agents operate and providing that guidance to the low performing ones in the form of,
- suggested prompts to more quickly diagnose the issue
- suggestions to maintain a good call dynamic
- helping find the best answer or resolution faster.
How does it perform? Agents supported by AI,
- take less time per conversation
- handle more conversations per shift
- resolve more calls and
- resolve more calls to the satisfaction of the customer
That’s called a win, win, win, win.