How to build your best bot: the Bot Stack Compendium

Micaela Neus
ddouble
Published in
4 min readSep 7, 2016

You might have noticed the craze for all things conversational these days.

Great tools for developing chatbots have appeared at a dizzying rate, and that’s a good thing. They take some of the complexity out of designing and developing bots as well as analyze their performance in the wild.

Choosing which tools to use to build your dream bot becomes a bit daunting in the face of so many amazing options. Do you know how to make a strong selection? What other choices could you have made, if you knew more about the options?

You need to arm yourself with information to answer these critical questions. We made the Bot Stack Compendium for that purpose.

→ UPDATE: You can submit new tools to be considered for the Compendium HERE.

The Backstory

Sandi and I started playing with various tools and soon got curious about how other people chose what to use build their bots. We threw together a quick “bot stack” survey and got some solid insights.

Insights, yes. Answers? Not quite.

Both the survey and our own experience convince us that there is no “right way” to build a bot, and best practices are still far from being solidified. The standard bot stack doesn’t exist. Your use case, target platform, team strengths, timeline and a thousand other concerns will inform how you structure your product. You don’t need us to tell you what the best way to build your bot is.

What you need are enough options to make an informed choice.

Enter the Bot Stack Compendium.

The Compendium tracks tools and services to build bots. Each one is tagged with a category and linked to its source and documentation. We also identify the level of engineering skill needed to make use of the tool, its pricing schema, its licensing and other handy information.

Pro Tip: Open in a new tab for a better look. Make a copy in order to sort/filter the results.

A thing of beauty, best viewed in its own tab. Open the Compendium.

The Compendium will give you a clearer grasp of the landscape of frameworks, services and platforms related to bot building. It’s restricted to tools for text-based agents so you won’t see things like Alexa Skills or APIs for identifying photos listed here.

What you will see is many resources for incorporating artificial intelligence into your bot despite the fact that most bots don’t use them (yet).

Our categories (e.g. AI/ML, KB, NLP) might seem rough or even wrong to you. That’s okay. Our intention was to orient people to options and sources of inspiration, not overload them with jargon. That said, we welcome thoughtful feedback in comments on this post.

Many products listed in the Compendium appeared in our bot stack survey. A few products aren’t yet publicly available and we only list them because we got early access. Thanks to everyone for the spirit of openness.

Note that tools marked Beta mostly require an invitation to join the beta testing group.

But wait, there’s more and more and more and…

Companies continue to put out new tools for building bots, adding to a long list of mature tools that pros have used for years. To keep up, the Compendium needs to a means to evolve alongside the industry.

We created a form to accept new contributions from the community for exactly this reason. Use it to share tools you love, whether or not you make them yourself.

Most days.

Did we leave something blank that you can fill in? Let us know through the same form by creating another entry for the same product. We’ll contact you if we have questions reconciling the two entries.

The world of conversational agents is vast and the terrain shifts daily. This overview should give you a better sense of the landscape, an idea of major trends, and the urge to go exploring.

Please feel free to make a copy of the Compendium for your own use. When you make or find a tool you want to share, use the form to add it to the list. Send us a tweet @DDouble to let us know what you think.

Want more? Follow our Publication! Good things are in the works.

We want to thank everyone working to develop and maintain these incredibly feats of engineering and creativity that we’ve called simply “tools.” We know how much passion and intellectual elbow grease goes into them. It’s heroic. Your work opens up unknown numbers of opportunities. Thank you.

Go, team, go!

--

--