Customer Service & Chatbots

How to become a psychotherapist

Full, non-edited transcript by descript.com

Chatbots, eCommerce, and how to become a psychotherapist. it’s Re:iterate, the digital marketing podcast. Hi, it’s Clint Griffin, and welcome to reiterate this week we are talking customer service and chatbots. Seeing how we can make that process simple for you, literally understand it and perhaps go ahead and make your own one and see how it’ll work in your customer experience department.

Any discussion around customer service and websites lately. Has been around AI, machine learning, and in a very big chatbot’s come up as well. How do they work? What do they do and how can you use them to make your user experience heck of a lot better? Well, the thing about chatbots is they’ve been around quite a long time.

The first chat bot was called Eliza, and it was developed in the 1960s mid 1960s by a guy called Joseph Weizenbaum. He developed a chat bot that was a psychotherapist. It didn’t have any particular intelligence, but it could answer questions and to make people feel as though they were engaging with a psychotherapist.

In fact, there’s a story that one point that his secretary Austin to leave the room and close the door when she was chatting to Eliza. I’m not sure what that is about psychotherapy. You can make your own decisions. The thing is that chatbots are really just a combination of words in a, in a tree.

Did she use to answer questions that up in a place in front of you and how this works for a customer service is that the can take a whole lot of the pain away from that process. So usually a personable furniture, customer service agent, or they will type a message on a computer to service agent.

And, or send an email. And what happens is the content doesn’t make any sense. They’ve lived off their contact details or they haven’t described what particular item they’ve got a problem with. Um, a response might be, I have a problem with my machine, and then the conversation goes, what kind of machine is it?
So, ma’am, um, what model is it? Where did you buy it? Can you describe the problems? Can you send us photos? This takes a long time. It makes the process laborious for both the customer and the port agents on the other end of the line. Um, and it makes it costly. So chatbots are quite a simple way to three months.

You could sit a check water in Excel, you have an opening statement. Hi, how are you and what would you like to talk about? And then you can give the customer three or four categories. You could say to them, would you like to talk about our specials? Would you like to talk about our products? Would you like some information about us?

Would you like to know the location of our nearest store? And by typing in a few of these details, you very quickly get to the bottom of what the customer is looking for. You might use geotagging to figure out where the nearest store is, all through a series of questions. You might find that they need a replacement bell for a hairdryer, and they can get that from the store or get it online.

Here. A lot of that takes away the human interaction, um, and makes the process very simple. What it doesn’t do is onset every question all of the time. The problem with checkbox is they all limited, and I think it’s important when you are sitting one up to understand the limitations and to realize that chatbots aren’t AI.

They simply a series of questions and answers designed to give responses that make sense to customers and help them understand or work out what they need to know. So things to be wary over the chatbots, they can get complicated really quickly. Try and keep them simple. Um, you might want four or five different ways for you to be able to help the customer.

As I said, um, who near a store, um, the cost of an item. Do you have this in stock? And maybe you’ve got something about warranties, which is quite a simple thing to fill out with just a link through to a warranty page. Because if you’re talking about products, for example, he might have 200 products.

So the query around, um, how much is, this can become quite deep and detail. How much are you a vacuum cleaners? Will, which vacuum cleaner. Um, I want a handheld vacuum cleaner. Well, there are five types of handle making penis. I would like to know about this one. Oh, that’s one is this much. Oh, well that’s too expensive.

Have you got one that’s cheaper? And you can see soon that these things add up to quite a complex array of, of questions and answers. The thing about them is that each onset and a check bot needs to be scripted. This isn’t machine learning is no particular science behind them other than answering queries with a prompt.

So it’d be very careful, um, with the partner that you choose or look online and see what veil available platforms there are that are free, that you can build your own chat bot and integrated into software. The RFU around that I’ve seen that I’ve used, they are quite simple. The is a cost to some of them after you scaling it to a certain degree, but by and large, you get a very simple understanding of what you want to do in our checkbox work really quickly.

I found the easiest way to make my first one was to open up Excel, go to the middle of a page, and in the first cell type in, hi, my name is, how can I help you? And then give three or four answers under that. Well, three or four categories perhaps. Then underneath that, you put those three or four categories in different rows, and then you see from the how you build it, those categories to answer questions that go deeper and deeper and deeper.

The ultimate goal of the check bot is to either answer the person’s query quite quickly or to gather enough information to allow that person to then chat or get information back. From a customer service agent or a real person without them having to go backwards and forwards and it’s, and it should save time.

It should say resources, and it should allow two things, customer satisfaction, and also allow your customer service agent to get on with really more interesting things and things that are much more profitable for business. Like upselling. Cross-selling, recommending things, taking the time to phone a customer and help them through certain complaints.

Because if you can do that, rather than telling them, telling them where the nearest store is, you’re adding a much more value to the customer’s life. You’re adding a lot more value to them that the way they see your business, and ultimately you provide yourself a lifelong customer. So if we think about chatbots, they give, are trying to keep them really simple.

Think about trying to design your own one in Excel and see how that tree works and how the knowledge unfolds and see how you can then translate that back to a customer service agent. That makes their lives easier. That lets them answer queries to customers and ultimately frees them up to be able to do deeper and more meaningful engagements with customers on other levels.
And of course. Lastly, don’t forget about the importance of data. What is the point of having a bot? If you know it’s collecting the data? Customer service teams can give you back enormous amounts of information and the reporting, but sometimes things are hidden because maybe the data tells them things you don’t want to know about.

Then, or it’s things that are so repetitive that I don’t even think it’s important. With a bot, you have it in a format that you can look at, you can analyze, you can see what the biggest queries are, where the smallest queries are, where there are suggestions and where we are falling down. Whether it’s logistics, our service, our products, their price point of a product, a person in store.

The cool thing about the bot is the information is there. All you need to do is look at it. It’s Rachel business and make it stronger pain. Don’t forget, every now and then, um, your team, I keep praise and getting praise from a customer is the most valued thing that you can get. So don’t be afraid to hand out the praise when it’s there and reward people as well.

That’s about it for this week. Please don’t forget to subscribe and we’ll send us a mail at podcast@reiterate.org. Have a great week and we’ll chat again soon.


3 Replies to “Customer Service & Chatbots”

  1. Solomon

    Now that you know how and where people will chat with your bot, you must decide how smart you’ll make it. Basically, this means whether you’ll use true artificial intelligence or machine learning to power it. The alternative is to provide a simple, menu- or command-based conversational UI. Sticking to simple commands is easy to set up in house by designing conversation trees, but if you want machine learning, you’ll likely need an outside source to power it. It’s worth mentioning that machine learning might make a chatbot seem more sophisticated, but keeping the conversation too open can also be overwhelming for the user; some bots benefit from a simpler approach. Whichever you choose, opt to provide contextual buttons throughout the conversation so users always know how they can respond.

Leave a Reply

Your email address will not be published. Required fields are marked *