Dialogflow: Google's chatbot platform doesn't disappoint

I was quite excited to try out Dialogflow for several reasons. First, I knew it had recently been bought be Google, and I know they wouldn’t buy anything flaky. Secondly, I had just discovered that the company behind it has a long history (well long in chatbot terms!) - it was founded as Speaktoit back in 2010. It first produced a virtual assistant for mobile performed similar tasks to Siri and Google Now. In 2014 they opened up the technology behind their app as API.ai, enabling third parties to make use of their speech recognition, text-to-speech and natural language understanding tools. Finally, after Google bought them in 2016 the product was renamed Dialogflow, in October 2017. I’m not quite sure why they settled on that name, perhaps to go well with Google’s flagship machine learning product, TensorFlow? Personally I thought API.ai was quite a good name.

Anyhow, as you’d expect from Google, the product is sophisticated. This comes with the downside of being not quite as easy to use as some of its competitors, in particular Chatfuel. However, once you’ve got your head round the concepts, the benefits of the additional power will definitely pay off.

Dialogflow is based around “intents”. An intent is a representation of what the user has said in a form that the computer understands. So if you say “I’d like a train to Norwich”, this may be mapped to an intent that looks something like:

• Intent name: “train”
• Destination: “Norwich”
• Departure location: unknown
• Desired arrival time: unknown

The nice thing about this is that once the system knows what you are looking for, it also knows what questions to ask you to help you achieve your goal. So it may say, “Great, what time would you like to arrive?” to help determine your desired arrival time.

But what’s really powerful about Dialogflow is its built-in ability to recognise entities. These are natural language expressions that are recognised as belonging to a certain category, for example place names or times. For example, for the query “I’d like a train from Luton to Norwich arriving at 3pm”, it is able to detect that Norwich and Luton are “cities” (technically Luton is a town) and that 3pm is a time, which is returned in ISO-8601 format (for example “15:00:00”).

Added to this is another really powerful feature: “contexts”. This allows your app to remember things that have happened and reuse that information at a later point. For example, after booking the train to Norwich, I might say “What time does it leave again?” and the bot would be able to know that I’m talking about the train booking.

What’s really nice about Dialogflow is the testing capabilities, which provide a simulation of your bot as it would work in Google Home. This means you get to actually talk to your bot! What is a little weird is that you first have to say “Talk to my test app”. After that, you’re actually conversing with your bot.

How much does it cost? There are two options:

• Dialogflow standard edition is free, and allows a maximum of 1,000 requests a day and 15,000 per month. It is also limited to 3 queries per second (averaged over a minute). Support is from the community or by email.
• The enterprise edition costs $0.002 per request for text queries and $0.0065 per request for voice interactions, and supports unlimited requests, and a maximum of 10 queries per second. It also possible to get Google Premium Support for this edition (at an additional cost, unless you are already paying for Google Cloud support).

It’s hard to beat this platform: it has excellent features at a reasonable cost. Although the learning curve may be steeper than other platforms, the payoff is worth it.

What next?

We think Dialogflow is one of the best tools for building chatbots. That’s why we’ve created a special tutorial on how to create a support chatbot just for Dialogflow.

However it’s not for everyone. There are plenty of other tools out there. Check out our top chatbot building tools for non-technical users.

All the reviews, one PDF

Choosing the right chatbot tool can be difficult. There are a lot of tools, and their quality varies a lot. The purpose of ChatbotTech is to make choosing a chatbot tool as painless as possible. We've reviewed and rated the top chatbot tools so you can quickly find the one that is right for you.

Sign up below and we'll send you a free PDF containing all the reviews, and ChatbotTech's guide to choosing a chatbot. We'll also send you occasional news and updates (no more than once a week).