2023 was all about chatbots. ChatGPT stole most of the headlines, but as companies across the planet worked hard to get their bots chatting, many questions started to bubble to the surface.
With their narrow conversation flows and questions that often don’t lead to an answer – or an agent – chatbots don’t always seem all that revolutionary. You wouldn’t be the first customer service professional to ask themselves if a chatbot is really any different than an IVR.
Which brings us to another uncomfortable question. If your chatbot seems like an IVR, are you doing something wrong?
In this post we’ll look at chatbots, their functionality and how they differ from the tried-and-true IVRs they sometimes seem to mimic. We’ll also talk about how to make your chatbot better, and we’ll consider whether you might be expecting too much from your ‘bot.
Of course, this conversation is evolving quickly as technology changes. But that doesn’t mean we can’t ask critical questions to improve our outcomes using the knowledge and tools we have today.
So, let’s dive in!
All Chatbots are Not the Same
It’s probably helpful to start by pointing out that not all chatbots are alike. Understanding a few basic differences can help clarify your sneaking suspicion that your chatbot might really just be an IVR in disguise. (Props to Jeremy Watkin, Director of Customer Experience and Support at NumberBarn, who first posed the question on Twitter and got us thinking about the topic.)
There are Two Main Types of Chatbots
Chatbots can be used for a number of different applications – but behind the curtain, you’ll typically find two main types of chatbots in popular use today.
1. Simple, Rules-Based Chatbots
These task-oriented chatbots are widely-used. Often lurking in a website, they’ll typically spring to life a few seconds after you land on a web page. They usually offer basic customer support, using simple pre-determined rules and conversation flows to respond to customer’s questions. These are the bots that most of us are deploying and the ones we’re mainly concerned with in this blog.
These bots are also described as task-oriented, declarative, deterministic or decision-tree bots and they work like a simple flow chart: If someone asks A, respond with B.
2. Smart, AI Driven Chatbots
Smart, AI or data-driven chatbots are much more sophisticated, interactive, and personalized than simple, task-oriented chatbots.
For starters, they are language wizards. They use natural language processing and machine learning to learn and grow as they chat with you. They should get better with every conversation and of course, many of them actually speak – though they can use a typed-interface, too.
These clever chatbots also seem to be able to read your mind, employing predictive intelligence and analytics that can make your chats feel very personal. They remember your likes and dislikes, make awesome recommendations, and can guess what you need before you say it!
Amazon’s Alexa and Google’s Siri are examples of smart, data-driven chatbots that learn as they talk to you, using artificial intelligence, natural language processing and machine learning skills.
What is an IVR?
An IVR, or interactive voice response, is an automated phone system that also typically has call-routing capabilities. If you’ve ever been asked to “Press 1” to speak with an agent, or “Press 2” to go back to the previous menu, then you’ve interacted with an IVR.
Like chatbots, IVRs are used for basic self-service. They can also route calls and collect pre-call information – such as when you are asked to enter your account number before being connected to an agent. Functionally speaking, they can be similar to a chatbot – they can offer a limited response to a simple question and can easily help with many basic self-service options. Like chatbots, they’ve got their strengths and weaknesses.
DID YOU KNOW:
Using chatbots can cut your customer service costs by as much as 30%. Source: IBM
What are Web Call-Backs?
At a glance, Fonolo’s Web Call-Backs (formerly Visual IVR) can look a lot like a chatbot, too. Visually, Web Call-Backs uses a simple web-widget with menu options to allow your customers to select a call-back from a live agent. If desired, they can also be programmed to collect pre-call information like account numbers, order numbers and the reason for calling. These powerful tools are excellent in tandem with a digital chatbot because of their ability to connect your digital customers to a voice-call if they need further assistance.
What Are the Differences Between Chatbots and IVRs?
Chatbots and IVRs have a lot in common as both are simple, self-service customer interfaces. But we’ll look at some of the things that makes each technology distinct.
- IVRs are usually deployed on voice calls and chatbots are usually found on websites.
- IVRs typically screen and route calls to an appropriate agent, while chatbots tend to offer more self-service options.
- IVRs employ voice recognition technology, and so do AI-driven chatbots. Simple chatbots do not have voice recognition software.
- IVRs follow a linear script, making them slower to use than simple chatbots because they can only react to one small piece of information at a time.
Where Can I Get a Chatbot?
If you’re just getting started with chatbots, choosing from the automated platforms and drag-and-drop chatbot builders is a common place to start. You’ll build your bot’s conversational flow based on a pre-determined set of rules. Many are on the market, including Microsoft’s Power Virtual Agent and HubSpot’s smart chatbot builder, for example.
If My Chatbot Functions Like a Bad IVR, Am I Doing Something Wrong?
Many bots are built using pre-built, turn-key deployments, and that’s OK. They are designed to permit simple functionality and can do that adequately. The problems begin with rushed logic and design that isn’t well thought-out.
You’ll improve your outcomes if you begin by mapping out very clearly what your chatbot is for and what it needs to do. Proceed with your end-user in mind and don’t scrimp on development time.
There are two things that typically go wrong:
The bot is not designed to be conversational, nor does it consider the needs of the end user.
The data we’re talking about is human language and conversation flow. If the designers don’t spend enough time collecting and analyzing language to ensure the bot says the right thing at the right time, the chatbot will underperform.
How to Improve Your Chatbot
1. Take your time when deciding the scope of your bot and when answering the question: “What should your bot do?”
To answer that question properly, you need to think about these related questions:
- What does your user want to do?
- What are their pain points?
- What value can a bot offer, or what need could it satisfy as well as, or better than, a human? (i.e.: What will encourage your user to engage with your bot.)
2. Design whole flows and complete conversations to be as simple and complete as possible.
3. Let your customers know what the chatbot can do. For example, the bot can say: “Hi, I’m Bonnie and I can help you with purchases, returns, and frequently asked questions.”
Think About What Your Bot Should NOT Do
Your bot can’t do it all. And you shouldn’t expect it to. There will always be some conversations that need to be forwarded to a human. And remember that your bot can be used to route these callers in a smart way to the correct customer service agent, by asking some pre-call questions.
A product like Fonolo’s Web Call-Backs is key to this transaction because you still need a way to link your digital customer to the voice channel. With call-backs, your customers can choose a call-back from the next available agent, or schedule a call-back for a time of their choice: any time up to 15 days in the future.
It’s an important piece of the chatbot puzzle, and key to integrating chatbots into your customer service workflow.
Do I Need to Choose Between a Bot and an IVR?
Probably not. As we’ve discussed, IVRs and chatbots each have unique qualities that make them suitable in different situations. Many businesses will need both to adequately deal with customer’s questions. And most will want a product like Fonolo’s Web Call-Backs to help bridge the gap between different self-service technologies and integrate chat flows with live agents when needed.