Motion AI Blog
The latest features, platform and industry news.
By very popular demand, we are stoked to announce the release of our Get Name module! This module will allow you to parse out a user’s name from their response to the module! We’re looking forward to seeing the awesome bots you create using this new module!
Today we are incredibly excited to unveil Motion AI’s new “Node.js Module” which gives anyone the ability to easily code complex chatbots directly within their browser, with no external servers required.
This integration was made possible thanks to AWS Lambda, with a spotlight announcement of our integration at AWS re:Invent today. Check out our documentation and sign up for free to get started — read on to learn more!
Our mission at Motion AI has always been simple — we want to make the process of training, structuring, and deploying chatbots as easy and as painless as possible.
The Motion AI platform is designed around the modular creation of bots, assembling various NLP building blocks to enable users to create any conversation flow desired. As we say, “if you can draw a flowchart, you can build a chatbot!”
Up until today, users looking to manipulate their Motion AI bots outside of our platform configured their own web servers to listen to and respond to webhooks.
With our new Node.js Module, this is no longer necessary. Whether you want to interact with an external database, connect with a third-party API, or do virtually anything else — there is no need to leave our platform.
Each Node.js function created through Motion AI is passed a payload object that contains metadata based on an end-user’s response to the bot. This data can be acted upon within the Node.js module to craft a bot response to be returned at the end of the function.
As a demonstration, let’s make a dynamic, API-backed trivia bot. You may test the end result on Facebook Messenger here.
To begin with, we structured a simple conversation flow that asks for the user’s name and greets them. Next, we ask them what difficulty they’d like to play at (Easy, Medium, Hard)
Because that is pretty basic stuff, dedicated Node.js Modules would be overkill. But in order to ask the user trivia questions, we want to pull from a large database of thousands of questions and answers (through a REST API), and that is where our Node.js Module comes in.
Once the user has selected a difficulty, we then pass them to one of three dedicated Node.js Modules: “Fetch Trivia — Easy”, “Fetch Trivia — Medium” or “Fetch Trivia — Hard”.
Each of these three modules are very similar in nature — they pull from separate databases of questions depending on the difficulty level.
Below you can see the code, with comments, used to retrieve easy trivia questions.
In this example, because our callback to Motion AI contains the key-value pair “continue = false”, Motion AI will continue to ping this same module until “continue” is no longer false. That is the desired functionality in this situation, where the trivia game can go on as long as the user desires.
Motion AI exposes a “customPayload” key that you may use to store and retrieve working background data. In this particular demonstration, we use this to store the last correct answer for any trivia question. This allows us to compare it against the user’s response (their answer) and determine whether or not they were correct.
Try the finished bot for yourself on Facebook Messenger here.
And just like that, we’ve created an API-backed trivia bot that will dynamically ask thousands of trivia questions. This particular bot demonstration took approximately 15 minutes to build.
Prior to our Node.js Module solution, users would be required to create a convoluted flow that either used webhooks to an intermediary server, or, they would have to manually input the thousands of trivia questions into our interface (pro tip: that would be extremely inefficient!)
Looking beyond the demonstration
The reality is that this demonstration only scratches the surface of what you can do. This opens the doors to quickly creating bots that can do everything from image recognition to machine learning, without ever leaving your web browser or setting up a server.
Bots are just one of a multitude of spaces where we anticipate serverless computing will play an increasingly important role. It goes without saying that we are very enthusiastic about the future of AWS Lambda, of serverless computing, and of the endless opportunities brought on by this amazing technological paradigm. We hope that you are just as excited as we are!
A look at Motion AI, available in public beta today.
We debuted a prototype on Product Hunt last November, where we reached #1 for the day and earned a spot among the top 70 “hunts” of all-time. Our vision? A platform allowing anyone to visually build, train & deploy chatbots to do just about anything.
“If you can draw a flowchart, you can create a chatbot.”
In the time since, we’ve been busy readying our platform and have watched with excitement as interest surrounding chatbots has continued to rise.
Last week, we began onboarding users of our waitlist to private beta, and today, we’re pleased to announce the general availability of our public beta.
Bots created with Motion AI are designed like a flowchart, which can be as simple or complex as you please — with as few or many branches as your project requires.
At launch, we support bot deployment to SMS, Email, Webchat, Facebook Messenger and Slack. In addition, we offer an API to give developers the flexibility to integrate with other services or roll their own mediums.
A world of possibilities
It’s been a blast seeing how the community uses the platform over the past 7 days—we’re seeing everything from a public opinion polling bot, to a movie recommendation engine and patient triaging bot for a hospital, to name a few.
Combined with webhooks from IFTTT and Zapier, users can build advanced bots that trigger actions on external services — for example: a bot that scans a knowledgebase to handle basic customer service questions, and helps the user submit a support ticket (to ZenDesk, Groove etc) if assistance is still needed. Check out a video of this particular example here.
We are preparing to launch our Bot Store, which will surface curated, pre-built bot templates from our community.
Whether you are looking to build a bot to automate your home, supercharge your customer service process, conduct surveys or do something entirely different, the Bot Store will enable one-click creation of bots for a wide range of usecases.
Launching our platform only scratches the surface of what we have in store… we are so thankful for the outpouring of grassroots support that this project has had from the entire bot and maker community.
Ready to build a bot?
Thoughts and lessons at the intersection of Product Hunt and funding
Just over a month ago, Motion AI made its debut on Product Hunt, reaching #1 for the day and earning a spot among the most upvoted products of all-time. Fast forward to this week, and we’ve just announced the closure of $700,000 in seed.
The community that Product Hunt has built is very targeted, with highly influential and engaged readers. Your Product Hunt launch may not bring down your servers or result in a phone call from your web host, but it can surface your product/concept to a very important demographic central to your future fundraising, recruiting and customer acquisition efforts.
Below the jump are some tips that helped improve our Product Hunt experience!
Engage the community
Product Hunt has a very technically adept community of readers and commenters. Not only that, they’re polite (yes, polite Internet commenters!) and love to engage with makers. Return the favor!
This is not a new piece of advice — most all Maker Success stories offer this suggestion. But here are a few specific things we did that helped:
- Aside from engaging people with a Product Hunt-specific landing page (using IntroBar or similar), consider making additional quirky or fun PH-inspired alterations to your site. In our case, this meant renaming our demo “Hoover’s Pizza”, and incorporating the glasshole kitty as an avatar in said demonstration.
- We did our best to engage with each commenter on Twitter to learn more about their interest. This actually resulted in several great customer acquisition leads and contacts, and even an angel investment in one case.
Create viral loops
You’re not allowed to explicitly ask people to upvote your product — and this is a great rule to have in place, ensuring the playing field remains fair. But that doesn’t mean there aren’t things you can do to promote the potential for upvotes by directing users to your Product Hunt page.
For example, every user who signed up for our waitlist was given the option to tweet out their position on the waitlist, with a chance to move up in queue everytime someone clicked on their link (check out Waitlisted for a great SaaS solution for this)
When their friends clicked the link, we didn’t send them to our homepage… we sent them to our Product Hunt page. This brought more eyeballs to PH and in turn, surely assisted with upvotes.
I just ran a report in Google Analytics and found that nearly 3,000 additional users flowed through these referral links. If only 5% of those users opted to upvote the product, that’s 150 additional votes right there.
Raising money? Bring Product Hunt into your pitch
Product Hunt is an almost universally lauded barometer of tech interest… Everyday, VC firms and angels check Product Hunt to see what’s hot. If you perform well on Product Hunt and seek funding, be sure to mention that! We devoted an entire slide of our Pitch Deck to this.
Particularly in early-stage investments, social proof is among the most definitive signs of demand… and when it comes to social proof, Product Hunt is a well-trusted indicator of this.
Make your Product Hunt connections last
Think of your PH journey as a networking opportunity, not a one-off engagement.
Consider adding a waitlist (if pre-launch) or Product Hunt specific mailing list prompt. Product Hunters want you to succeed and love to hear about your journey forward.
Over 5K individuals and companies joined our waitlist, and it’s been a great way to keep in touch with each and every one of them since then.
There’s no silver bullet to performing well on Product Hunt, but there are a number of steps you can take to improve your chances and optimize your experience.
Study the top hunted products each and every day, check out the comment threads and take note of what people like (and don’t like)
Ultimately, if you don’t think your product would pique the interest of the community in the first place, these tips won’t be very beneficial.
Going back to looking at PH as a barometer of tech interest, the reality is that understanding what the Product Hunt community enjoys has implications that run far deeper than just your tenure on PH.
If there’s anything I didn’t cover, feel free to ask!
Happy Product Hunting 🙂
The round was led by Crush Ventures with participation from Hyde Park Angels and a host of prominent angels— including quantum computing company 1QBit CEO Andrew Fursman, Guy Gamzu and Fiverr CEO Micha Kaufman.
To say I’ve had a wild (and exciting) past 35 days would be a huge understatement!
After announcing Motion AI’s funding round about 2 weeks ago, we surpassed the initial ask of $350K in just 6 days. The round quickly became oversubscribed, with $700K of commitments streaming in over the days that followed.
I’m told the round is one of the fastest to come together in Greater Chicago history, which is an exciting validation of what we’re building.
It was just over a month ago that Motion AI first debuted on Product Hunt. This was a tremendous way to introduce the concept to the public, and the success of this launch has greatly helped move ahead many of the company’s goals.
For those not familiar, we are working to build the “UI for AI” — an end-to-end platform that simplifies and streamlines the many aspects of deploying Artificial Intelligence.
This funding will allow us to make key hires, build out our engineering team — and most importantly, launch our product to general availability in the near future.
I’m so thankful for the outpouring of grassroots support I’ve received since the beginning of this journey.
One of my favorite parts of this entire process has been engaging with prospective users and partners of our product. If you have any questions or comments, please feel free to email me… I’d love to hear from you!
Make sure you’ve signed up for our waitlist to stay up to date on the very latest. We can’t wait to get our product into the hands of the public.
Why we’re building the UI for AI — and a status update on where we’re at, where we’re going, and how we’ll get there.
Reception and feedback was incredible. We reached #1 for the day, and with 1,100+ upvotes, Motion AI became a Top 70 hunt of all time.
It was fantastic to see our thesis validated, and hear specific use cases from thousands of excited individuals.
Today, I’m excited to announce the opening of our angel round. We are seeking $350K and accredited investors may access additional details on Angel.co. More information is available at the bottom of this post.
For those who aren’t familiar with Motion AI, we’re building the tools required to simplify the normally complex nature of deploying Artificial Intelligence to business applications. Our product allows companies to easily train and deploy robots to handle any range of tasks from customer service to food ordering and healthcare tasks. In a nutshell: we’re the UI, for AI. Motion AI is IFTTT meets Omnigraffle, for Machine Learning.
“If you can draw a flowchart, you can create Artificial Intelligence.”
Over 5,000 companies and individuals have joined our waitlist, and the pizzeria demo on our website has served up almost 10,000 virtual pizzas.
To date, Motion AI has been self-funded, thanks to the success of prior ventures. And while we’re capable of bringing Motion AI to general availability without funding, this round allows us to expedite that process and bring on key hires sooner than we otherwise would be able to.
So far, I’ve done much of the engineering myself and have relied on contractors as-needed. Thanks to our strong debut, we already have many job candidates who are passionate about AI and our mission.
We will be fully launching our product around six months after the closing of our round, with initial onboarding of those on the waitlist beginning around month three.
Just since our Product Hunt debut, we’ve made great strides on the development of our core product. At present, the back-end is 60–70% completed, and our front-end is coming along very fast. Thankfully, our pizza demo is far more than a hardcoded prototype — it runs entirely on our platform.
Our infrastructure is predominantly based on Ruby and Python, making use of MongoDB and Node.js. Stanford CoreNLP was our starting point for Natural Language Processing, while Theano has guided our Deep Learning efforts. We’re currently hiring, so be sure to check out our job listings here!
I’m incredibly excited about the future of Motion AI, and truly thankful for the great support and feedback I’ve received from the community so far. I believe that together, we will create the best, most robust Artificial Intelligence platform around.
If you’re an accredited investor and would like to participate in our angel round (minimum investment is $2,500 — but priority is placed on larger investments), please email me: [email protected]
Finally, be sure to join our waitlist and follow us on AngelList for the very latest!