Home / Software & Service News / 5 steps to training a chatbot

5 steps to training a chatbot

Training a bot in five steps.


So you’re thinking of implementing a chatbot like every other company on the planet. Now, the issue is about training the thing.

There has been an endless stream of coverage about all of the wonderful things chatbots are going to do for business by automating conversations with customers. The problem is, bots are only as good as the training that goes into them and training isn’t something you just knock off in a few days.

Many bot startups seem to want to treat bots like an IVR, choose the top 20 use cases and training your bot around those needs. Considering IVR is probably the most hated piece of technology invented in the last 50 years, replicating that process is probably a horrible idea.

Very few, if any, bot vendors have actually operated these things out in the wild with millions of customer interactions over a number of years so it makes sense that we’re not yet hearing about the best practices when it comes to training a bot. The reality of training a bot is a little more complex than might be obvious at first glance.

At the highest level you need to consider all of the various channels that customers are using to contact you (such as Facebook Messenger, Twitter, Kik, WeChat, web chat, mobile apps, Amazon Alexa, and many others), the unique UI capabilities of each channel and multiple languages to be supported. You definitely don’t want to train in multiple systems.

This is a little more complex than just plugging in the top 20 use cases, so let’s look at the best way to train a bot to deliver real value to your enterprise and customers.

Step 1: Pre-training your bot

This is what many bot companies are pitching as the entire training program, when it is really only the first step of a real training plan. This is where you go to your customer care or tech support department and see what they think are the top reasons your customers are contacting your company. By sampling the chat logs, call recordings and digging through the typical support material that the reps use you will get closer to understanding the customer intent. This is a decent place to get started on training your bot.

Step 2: Analysis of all customer interactions

Every day your customers are telling you what they want to know and what they really care about. Many bot companies assume that you will know what your customers want and if you don’t that it is your problem to figure it out. A tool that analyzes customer questions in real time is a key piece of any bot solution. The analysis will tell you what your customers actually want to do with your bot and will let you focus your training efforts where it pays the biggest dividends. This is the step where you determine the bot priorities relative to customer needs. This is the most important part of any bot training framework.

Step 3: Train the bot

With all of the rich information coming from the analytics, you now know how to train the bot. A good bot solution will now let you match your bot to the best platform and take into consideration each platform’s capabilities and limitations. What you present in Facebook Messenger won’t work in a pure text messaging bot, and what you can present in Twitter may not work in your web chat UI.

A well architected system will let you train for both conversational and Q&A types of platforms in one system. Multiple languages should be automatically generated or updated if you have created a new scenario or just updated an existing one. Training multiple bots to solve this problem is insanity, as described in this previous Venturebeat article.

Step 4: Measure the effectiveness

Use every question from every customer from every channel every day to determine how effective your bot is. There are many ways to measure the level of satisfaction of your customers during their interaction with the bot. Over time you will determine normal thresholds for each parameter and then can use anything outside the norm as an indicator of failure. Your false positive rate should be monitored closely here. Delivering the wrong answer is far worse than not delivering an answer at all so this is critical to measure. If you are delivering the wrong answers more than 3% of the time your system should be taken back to the drawing board.

Step 5: Continuously improve

The process of training your bot never ends. Even if you have a simple pizza ordering bot you’re going to have to continuously learn new languages from your customers about how they want to order, add new platform support and new products. If you have a more complex business and are using your bot for customer service then you should plan to invest a considerable effort in ongoing training.

Don’t get caught up in the bot hype and deploy something that is going to create more work in the longer term. A fully thought out customer strategy will provide large customer satisfaction and economic benefits.

Get more stories like this on TwitterFacebook

Click Here For Original Source Of The Article

About Ms. A. C. Kennedy

Ms. A. C. Kennedy
My name is Ms A C Kennedy and I am a Health practitioner and Consultant by day and a serial blogger by night. I luv family, life and learning new things. I especially luv learning how to improve my business. I also luv helping and sharing my information with others. Don't forget to ask me anything!

Check Also

What we’re watching: ‘Raw’ and ‘Feast of Fiction’

Welcome back to Video IRL, where several of our editors talk about what they've been watching in their spare time. This month we're kicking things off with some seasonally-appropriate horror fare, that you can catch right away on Netflix or Amazon Prime. Then it's time for a Gundam throwback before Kris Naudus points out a couple of YouTube food channels perfect for binge eating or binge watching.

Them / Raw


Timothy J. Seppala

Timothy J. Seppala
Associate Editor

To get into the Halloween spirit, I've been watching at least one horror movie a day since the end of September -- the lower the budget, the better. Problem is, so many of the American low-budget or indie horror offerings on Amazon and Netflix are crappy Paranormal Activity clones, cheap-thrill gore-fests or uninspired found-footage "documentaries." Whether it's by design or coincidence, I've found that French horror movies have held my attention the most lately. Specifically, 2016's Raw, as well as Them, from ten years prior. They're more psychological thrillers than straight-up horror, but that didn't stop me from being more on edge while watching them one afternoon than I was during A Haunting in Saginaw, Michigan, late at night. Both start with a car crash, but they couldn't finish any more differently.

Raw, recently added to Netflix, tells the tale of a vegetarian girl in her first week at a prestigious veterinary school. During a hazing ritual, she's forced to eat a raw rabbit kidney. She immediately gets sick, throws up and wakes herself up that night scratching a full-body rash to near bleeding. This bout with food poisoning is just the beginning, though, and soon protagonist Justine finds out she has a taste for forbidden fruit. As the remaining 70-ish minutes unfolded, I lost track of how many times I clasped my hands over my mouth, agape in shock, to stifle my shouts of "OHMYGODWHATTHEFUCKISEVENHAPPENING?!"

But French director Julia Ducournau balances every body-horror scene either with something pedestrian twisted into being unsettling (like a horse on a treadmill) or with something that makes you ask how far Justine can go before someone confronts her about her new diet. And those questions keep coming right until the credits roll. I can't say I enjoyed watching Raw, but it was a hell of a ride.

The same goes for Them, currently streaming on Amazon Prime. Its focus is narrow, centering on a young couple living in a cavernous farmhouse, terrorized over the course of a night by unseen horrors. The camera never quite gives away who (or what) the perpetrators are, and revealing the twist would be a sin. As with Raw, its atmosphere and overall creepiness won me over straightaway. The scariest part? Realizing that I've probably driven past a shot like the final scene countless times and not thought twice about it. If you're willing to read subtitles, both of these should make you shiver and scream more than The Conjuring 2 on HBO Go could ever hope to.

Mobile Suit Gundam The 08th MS Team


David Lumb

David Lumb
Contributing Editor

I'd heard that a lot of anime had left Hulu, but I scanned their selection anyway looking for classic shows I'd missed, like the original Mobile Suit Gundam. They don't have that -- but they did have a series I didn't finish the first time it aired on Toonami, the 1996 classic Gundam side story The 08th MS Team. Unlike the franchise's other show released the year before, the massively successful Gundam Wing, 08th ditches the brand's typical pretty-boys-in-unbeatable-robots for a grounded and sobering story about the people who get caught up in wars -- desperate soldiers, civilians and guerrillas alike. It's dirty, honest, utterly humane and gorgeously animated.

It's also a little preachy and melodramatic, and it shows its age with odd sexist moments. While it's still the Thin Red Line of the Gundam universe, I remember it far more fondly from when my 14-year-old self grazed the series on its first American airing. There's something sad in seeing an old favorite for the flawed media it always was. Much like Waypoint's Rob Zacny, I've grown up and seen a lot since I first caught the show as a starry-eyed teen. I still think The 08th MS Team is a wonderful little 12-episode miniseries with a big heart, but I won't revere it so highly -- and will think a little harder about who I recommend it to.

Feast of Fiction / Binging with Babish


Kris Naudus

Kris Naudus
Senior Editor, Database

Back in March, I came home from a trip only to discover that my oven didn't work. The cooking gas in my building had been shut off due to a leak. My building management seemed to be on it, so I made do with a combination of microwavables, toaster oven and Seamless. Unfortunately, weeks and months went by, calls to the city were made and permits were issued, but, even as I write this in October, gas still has not been restored to my building. My landlords eventually threw their collective hands in the air and began installing electric ranges in every apartment, so a few weeks ago I was finally able to cook for myself again.

I am so jazzed to be able to make food. Hot food! Scrambled eggs! Steak! Cookies! I started reading food blogs and cookbooks, and shopping to refill my pantry. I'm halfway through Kenji Alt-Lopez's The Food Lab, a huge 900-page hardcover that talks about the science of how food cooks. On the lighter side, I've also been reading food-themed comics like Delicious in Dungeon and Food Wars. And the latter title (which is also an anime) ended up sucking me into a YouTube hole of food videos that I've been obsessed with ever since.

You see, the very first chapter of Food Wars features the "Gotcha" Pork Roast, a bacon-wrapped potato loaf that hero Soma Yukihira makes to save his family restaurant. It looks pretty tasty, so I searched for recipes and pics online and stumbled onto Jimmy Wong and Ashley Adams' Feast of Fiction, a series that demonstrates how to make various foods seen in cartoons, video games and comics. If you ever wanted to taste Steven Universe's beloved Cookie Cat ice cream sandwiches or Kirby's super-spicy curry, there's an episode for you. One thing I really enjoy is how they also incorporate crafts into it, showing how to make paper wrappers for your Reptar chocolate bars or genuine-looking Ecto Cooler Hi-C boxes.

I've been marathoning through the episodes, which the YouTube algorithms have definitely picked up on at this point, throwing food show after food show into my suggestions. One that caught my eye was Binging with Babish. Where Feast of Fiction mostly sticks to the realm of kids' cartoons, anime and video games, Binging with Babish is a little more mainstream, covering foods from popular media like Mad Men, Seinfeld and House of Cards. Still, there's a bit of overlap -- both Babish and Feast have done their own takes on the Ultimeatum from Regular Show and Krabby Patties from SpongeBob SquarePants. But the recipes are different, and I watch the shows for the personalities. Feast of Fiction is pretty silly (and there's a cute dog), while Binging with Babish is a little more subdued. Not that Babish can't be ridiculous as well -- the Moist Maker is one of the most ridiculously complicated sandwiches I have ever seen, basically asking you to cook an entire Thanksgiving dinner.

Sadly, I still haven't done a lot of actual cooking since getting my stove back. I'm having too much fun watching other people do it instead, with the added bonus that I don't have to clean up the mess.

"IRL" is a recurring column in which the Engadget staff run down what they're buying, using, playing and streaming.

css.php