Saving Jane -  How conversational A.I. can help you avoid feeling sick as a parrot

I was sitting in reception, waiting for a meeting at a well-known football club

 Their digital team were running late, so I had some time to kill.

 To pass the time, I got chatting to the receptionist *Jane*. (name changed to protect the innocent)

When I say chatting to Jane, I mean talking briefly, then stopping, as the phone was ringing off the hook. She was probably answering a call every, say 45 seconds that lasted about 10 seconds. I couldn’t help but overhear the slightly exasperated response Jane had to trot out on around 99% of the calls, which went something like this;

“Hello, reception this is Jane speaking, how can I help? Um, no, this isn’t the number for match day tickets, ring them on this number. Have you got a pen? No?

Ok I’ll wait……(this is taking a while)

..No problem, thanks for the call, bye now.

I asked if that was pretty common (it was) and if they had any kind of basic automation in place, maybe a Facebook assistant.  Website chatbot? Simple voice automation? Basic FAQ? Voice skill?

ah. ok.

‘Nope’ came the slightly jaded reply. It got me thinking about the wasted opportunities, time and cost (and Jane’s sanity obvs)

Oh Jane, it really doesn’t need to be like this.

Chatbots are your saviour

Some farmer’s maths;

  • Calls per hour = 50
  • Calls per day = 350
  • Calls with the same response from Jane = 95% of 50 = 45 x 7 = 320 ish
  • Missed opportunity to upsell from a live agent — say 10% of calls could lead to a sale = 30 calls a day. Tickets are average of £35. Possibly £1000 in revenue per day not being closed. Blimey.
  • And one pretty miffed receptionist

Now, while the numbers are a guess, and it was only poor Jane in reception if you extrapolate that out to any business that deals with customers (i.e most of them) it’s not a massive leap to see how basic automation –  a digital assistant chat or voice integration could help — even if it was just for that one receptionist not having to answer the same questions over and over again.

Absolutely bonkers. Here’s what they could and should be doing for starters;

  • Integrate a simple Q&A chat/voice module to deflect the vast majority of requests. This is a pretty straightforward thing to undertake.
  • Talk to Jane, she’ll know exactly what questions are being asked (in this case one big question)
  • Automate the top 20 queries being asked through social media, Facebook, your website, and the call centre.
  • Check social listening tools — such as Sprout, and Buzzsumo and use as an additional corpus of data to train conversational intents, answer questions better and keep customers happy.

Then put together a pilot to test and learn from with a small cohort of internal users/regular customers and a small corpus of test data — extracted from actual conversations to get some useful baseline data from which to take the next actions. This could be a simple FAQ assistant to gather that data and get a feedback loop going.

This stuff works, just be sensible, answer core questions, amend and update regularly to test where you are at, what the impact is on your KPIs – the number of calls, enquiries answered, lead generation through these new channels – as well as customer NPS/ happiness scores. That will give you a steer as to what you should be doing next

So what tools are available where to start?

The basics are pretty straightforward, you can use effective tools like drift, intercom, landing.io, Chatfuel, and Dialogflow or further up the chain integrate with fully loaded services like Salesforce,  Amazon,  Google Assistant, and Oracle Digital Assistant to tie in with Live chat, integrate with your CRM and CX platforms and activities.


So, if you’re business has any sort of customer service function that is purely people based –  and you’d like to find out how automated conversational AI can get help save you hard cash, drop us a line at hello@thesynthetic.co

If not for yourself DO IT FOR JANE 🙂 You’ll be over the moon (Enough with the stale football puns. Please. – ed)