As a tool for improving phone conversations, artificial intelligence was sitting in a distant corner of my priority list, until Krystal Hobbs introduced me to Jim Brown who interviewed me on his podcast Sales Tuners which was heard by Richard Smith of Refract.ai who called me and booked a conversation.
The result of that conversation?
This summer I’ve joined up with a client’s sales team, making outbound calls to prospects. It is always valuable for me to get on the phone, talk to strangers, uncover changes in how people are using the phone to communicate. And there’s an added bonus because this team is working with me to test the Refract.ai platform. We dial our calls using the platform so the content of the call can be analyzed using artificial intelligence. It feels like we’ve got a bit of magic on our side.
We haven’t created a ton of data yet and, as with all new things, we’ve run into a few snags, but a brief test indicates we have the opportunity to capture some important insights that will help improve communication and increase sales. Here’s what I mean, using only information for one of my conversations:
- My average speaking time on the call was 55%. On the surface, this sounds good. It shows that I’m inspiring conversation and that the prospect is speaking and sharing information in almost equal measure. When we have more data, we’ll know the truth. Perhaps, for a product that requires some education, I should be talking 70% of the time? Or perhaps I should be talking less? With more data, we’ll be able to “craft” our conversations differently to get optimum results.
- The conversation switches from me to the prospect, from me to the prospects, etc, 45 times. A switch indicates that the prospect is engaged. This is a fascinating aspect to analyze but again, what does it indicate about my ability to sell? More data will reveal an ideal number of switches for an ideal and engaged prospect.
- The average time I spoke uninterrupted was 11 seconds and the longest segment was 50 seconds. While this could indicate that I’m great at delivering small bits of information and giving the prospect time to think, it could also indicate that I’m too generous with the prospect, giving them more time to speak while missing out on opportunities to educate and inform.
- My speaking speed is 139.2 words a minute. Experts indicate that clear communication requires between 110 and 150 wpm, so I don’t think I need to focus on making changes here.
- I asked 14 questions during one conversation. This is an average of one a minute, but the analysis also tells me that I asked the majority of the questions in the second quarter of the call. (It divides each call into 4 equal sections.) Is this good? I’m not sure yet. Does it mean I have a lot of engagement at the front of the call but not at the end … where I should be moving the prospect through my sales process?
- And finally, the platform tells me what “filler words” I use. We all have them and I’ll admit that I haven’t paid attention to mine for quite awhile. My #1 filler word? “So”. I used it once a minute. Yikes! I said “Great” and “Right” six times each, and “Fantastic” and “Really” three times each. Hmmm. While I know I can’t change all of these at once, I’ve already started paying attention to “so”.
While this limited data doesn’t yet provide any valuable analysis, it does confirm that using AI in this way has advantages. Oh, and one I didn’t mention. For years I’ve listened to audio recordings of calls, scrolling backwards, clicking pause, writing down words or phrases or comments to discuss with clients. It is very time-consuming. Refract.ai includes a transcript of the call, as well as the audio. Alleluia!
I’ll keep you informed of my relationship with AI in the coming months. If you’ve got a specific question or comment, do include it in the section below.
Fascinating Mary Jane,
can the AI also measure the time it takes to get to the point of the conversation? I wonder sometimes about people that call me about my business and then do small talk for almost 2 minutes before they come to the point. I am talking about people that don’t know me. (smile)
I always try to be succinct and aware of the other’s time constraint but wonder if I am also doing the long “hi, how are you and isn’t this weather beautiful” dance before starting the conversation for real. (smile)
Yes, Jacqueline, AI can measure how people might be wasting time on the phone. Using tags or key words, you can search the call and identify exactly where conversation is taking place that is, for lack of a better word, inappropriate. So you could search for the phrase “How are you?” or you could choose other words such as “vacation” or “hobby” (or anything) to see if the call moved into a very personal conversation. It is quite fascinating. Thanks for your interest.
What a great opportunity to be able to try this new program. Do you know if this could be used in a call centre? I may know someone who would be interested.
Yes, Roberta, the system can certainly be used in a call centre. It does have the ability to monitor numerous individuals at the same time. Have your friend connect with Richard. I know he’d be delighted to provide a demo and discuss all the possibilities. He’s a very nice person, easy to speak with and very respectful.