Get rewarded for investing more into SapientX:
As A.I. technologists, we were initially excited when Siri and then Alexa first came out, but ultimately felt disappointed by their limitations. We realized that we could offer an alternative AI product that allows you to talk to your technology products as if you were talking with your best friend.
Our products understand up to 99% of requests, speak 40 languages, work both on and offline, and they protect user data. Companies we work with range from auto and motorcycle companies to companies that offer digital signage, teleconferencing systems, vending machines, and more.
Mia is a prototype assistant developed for Mitsubishi. It is not currently available on the market.
The Problem
Current voice systems have surprisingly low accuracy, requiring users to learn specific commands to communicate. Most only speak a few languages, limiting the users they can reach.
Most current voice systems also need high-powered servers and an internet connection to run. In addition, they profile their users and sell sensitive data.
The Solution
Nara assistant prototype for NEC. It is not currently available on the market.
We started SapientX to make using your voice assistant as easy as talking to your best friend. Our conversational interface needs no user training, functions with or without an internet connection, and never sells user data.
SapientX speaks 40 languages, making technology accessible to more users than ever.
Zia languages demonstrator.
The Market
The voice AI market is growing by leaps and bounds. The voice AI market is expected to reach $31.82 billion by 2025, with a CAGR of 17.2%.
This is a highly active M&A environment. Our peers have been acquired by companies such as Apple, Microsoft and Samsung.
Our Traction
Our first big customer, Mitsubishi Motors, challenged us to build the Mia voice assistant prototype when they were exploring adding a voice assistant to their vehicles for their next generation of Outlander SUVs. We also created an automotive prototype for LG, a supplier for the dash screens used by Tesla.
Previous versions of our technology have been used by our co-founder Bruce Wilcox to win the Loebner Prize four times.
In 2018, we closed our first seed investment round and began work with companies such as Samsung, KTM, Zero electric motorcycles and Magneti Marelli, formerly a division of Fiat Chrysler. In 2019, we prototyped with Volvo, Yamaha, NEC and Indian Motorcycles.
While Covid-19 offered some challenges, we also found new opportunities. Our new Telecaster Teleconferencing system lets us bring our assistants to Zoom, WebEx, and other video conferencing systems. We are also working with vending machine companies to add a touch-free voice and avatar interface for their customers.
In November, we demonstrated our integration with Unreal’s new Meta-Human game system. This allowed us to display conversational characters at life-size in high resolution. In September, our RayGun Studio authoring tool reached beta user testing. This summer, we released the Zia assistant for Beta testing in English, Chinese, Spanish, and Japanese.
We hope to begin shipping our software in our customers’ products beginning in 2022.
What We Do
Our AI voice system solution is designed for rapid prototyping and flexible integration. We run on or offline, in text or voice, and with or without an avatar. We use best-of-breed third-party technologies such as Spotify, TomTom, Wikipedia, MapBox, and Google Maps.
Our target applications include cars, motorcycles, and a wide range of other consumer products. We deliver white-label solutions tailored to meet our customer’s needs and tools designed to have them maintain and extend the voice system over time.
*Images used in this graphic are of the prototype software. Software is not currently available on the market.
The Business Model
We adapt our revenue model to the norms of the industries that we work in.
For example, with cars, this means paid or unpaid prototyping and user testing leading to per unit software sales along with yearly, recurring maintenance, and support contracts.
Sample Pricing:
Cars & Trucks - $20/vehicle + $5 yearly maint. + support.
Motorcycles - $10/motorcycle + $5 yearly maint. + support
How We Are Different
We offer higher accuracy, user data security, 40 languages and we can run on almost any platform.
When dozens of companies told us that they would use our software in their products if it spoke Japanese or German or Chinese, we knew that we had to take up this challenge in order to grow.
We decided to do what seemed impossible and develop support for not just a few languages but a total of 40 global languages and dialects allowing us to reach 5.5 billion potential users.
The Vision
SapientX is one of the few products built from the ground up to support natural human interaction, replacing older command-based systems. We believe many new products will use a conversational interface, changing the standard for voice systems.
This is already a large market and growing rapidly. A number of our peers have been acquired by larger players. We believe that the best way to reach the most users may be by joining a larger company with an established sales pipeline.
Field testing our system in a Chrysler Pacifica van for Magneti Marelli.
OUR LEADERSHIP
SapientX was founded by longtime friends, David Colleen, Bruce Wilcox, and Maclen Marvit, who wanted to go beyond the keyword-based chatbots used by voicemail systems to make a voice technology that could interact with users naturally like they were talking to their best friend.
From there, we assembled a team of passionate AI scientists, software developers, and world-class animators. Together, we are well-equipped to bring our vision of a better voice system to life.
Why Invest
*Images used in this graphic are of the prototype software. Software is not currently available on the market.
With strong established relationships with Fortune 500 companies’ in a fast-growing, multi-billion dollar industry, and a product that has higher accuracy, security and accessibility, we believe we are the future of voice assistants.
We have everyone we need except for one last person: you. Join our team as an investor and share with us this thrilling ride as we climb taller peaks and change the future of AI voice systems.
Thanks for choosing us!
*Images used in this graphic are of the prototype software. Software is not currently available on the market.
There may be an app and (pretty soon) a bot for just about everything, but this might be a first. SapientX, a company that has been in stealth mode for the last year, has created chatbots for Donald Trump and Hillary Clinton that provide words directly from the candidate’s mouth on topics ranging ..
A number of chatbot offerings were demoed at a special AI-focused session of the Health 2.0 annual conference in Santa Clara, California on Monday.
SapientX™ < www.sapientx.com>, a leading supplier of conversational AI software, today announced it will be showcasing the next generation of conversational AI (complete with a 3D avatar) at CES January 9-12 in Las Vegas (Sands Convention Center, Hall G, Booth 51516).
往人工智慧的概念或議題，往往只存在科幻電影中，然而前些時日機器人 Sophia 在沙烏地阿拉伯的演講驚豔世人，讓大眾驚覺原來 AI 的觸角已悄悄延伸至我們生活中。在今年 CES 展中亮相的 SapientX 會話式聊天機器人便是 AI 的應用之一，該產品更被外媒評為 2018 CES 十大最愛產品。
Bots are officially everywhere — they’re available to assist with almost every part of our lives, from shopping and gift-giving to news-tracking and political decision-making. They can even tell us what to make for dinner.
“If you’re going to cast a vote, make sure you’re casting a vote from a position of intelligence and knowledge, and don’t let the propaganda from either side sway you,” SapientX board of advisors member Jonathan Hirshon told VentureBeat in a phone interview earlier this year. “Make your own decision
Company | : | SapientX Inc. |
Corporate Address | : | 4600 Smith Grade Road, Santa Cruz, CA 95060 |
Offering Minimum | : | $9,999.60 |
Offering Maximum | : | $4,177,800.90 |
Minimum Investment Amount(per investor) | : | $100.75 |
Offering Type | : | Equity |
Security Name | : | Series Seed-3 Preferred Stock |
Minimum Number of Shares Offered | : | 15,384 |
Maximum Number of Shares Offered | : | 6,427,386 |
Price per Share | : | $0.65 |
Pre-Money Valuation | : | $7,921,386.00 |
Investment Incentives*
$500+ | Tier 1
Beta Access - Join our test team for early access to our Sage multi-languages assistant and get a copy of our team t-shirt.
$1,000+ | Tier 2
Previous Tier + a custom SapientX leather key fob from Etsy made for us in Canada.
$2,500+ | Tier 3
Previous Tier + a quarterly update call with the founder David Colleen, CEO.
$5,000+ | Tier 4
Previous Tier + an invitation to a picnic lunch at FlipJack Ranch (www.flipjackranch.com) in the Santa Cruz Mountains with David Colleen. David will autograph your t-shirt if you're not careful.
$10,000+ | Tier 5
Previous Tier + a full day at Flipjack Ranch with founder David Colleen to discuss product ideas and the design of voice systems for Fortune 500 customers. The day will wind up with a farm-to-table dinner at the camp ring.
*All perks occur when the offering is completed.
A crowdfunding investment involves risk. You should not invest any funds in this offering unless you can afford to lose your entire investment. In making an investment decision, investors must rely on their own examination of the issuer and the terms of the offering, including the merits and risks involved. These securities have not been recommended or approved by any federal or state securities commission or regulatory authority. Furthermore, these authorities have not passed upon the accuracy or adequacy of this document. The U.S. Securities and Exchange Commission does not pass upon the merits of any securities offered or the terms of the offering, nor does it pass upon the accuracy or completeness of any offering document or literature. These securities are offered under an exemption from registration; however, the U.S. Securities and Exchange Commission has not made an independent determination that these securities are exempt from registration.
[The following is an automated notice from the StartEngine team].
Hello! Recently, a change was made to the SapientX offering. Here's an excerpt describing the specifics of the change:
Issuer is extending campaign end date.
When live offerings undergo changes like these on StartEngine, the SEC requires that certain investments be reconfirmed. If your investment requires reconfirmation, you will be contacted by StartEngine via email with further instructions.
We are happy to announce that we have two new customers for our trade show assistant software; Modev for their Voice 22 show in Washington DC, October 10-12 and Questex for their Digital Signage Experience show in Las Vegas November 17-19. Please join us at those shows. Our team is busy designing new characters to act as show guides to inform and entertain show visitors. We think it's a big win to place our characters in the lobbies of trade shows because it exposes our product to many potential customers.
In addition, we are busy working on Anna, a prototype guide for the Edmonton Airport. Anna can guide you to your gate, tell you where to pick up your luggage and even tell you which restaurant has the best poutine. Because this is Canada, she speaks both English and French and she can even tell you the latest hockey scores. Go Oilers! We are in discussions for another airport in Canada and one in Germany.
Xander at the AWE show last May
Most of our work, here at SapientX, has been in building the best conversational system on the market. Occasionally, we work in other areas to add important functionality. Earlier this year, we tasked Conrad with developing Sage's ability to see using a webcam and machine learning software approaches. I'm happy to report that this is now operational.
Our first use of this technology is for our assistants that live in kiosks at airports and trade shows. We want our assistants to know when people approach so that we can greet them. When a user begins to speak, Sage can hear them and now see their mouth move. We can then triangulate this position and have the avatar turn to face the user as they speak. It's a subtle but powerful effect. All of a sudden, the character seems more engaged and alive. If two or more people approach, Sage can turn to address each of them as they speak. Love it!
BTW... the beta testing is going great and we have 1,000 people helping us to improve Sage and her five personas. We will be inviting more of you $500+ investors in this week. Thanks again, we are learning a lot from your interactions with Sage.
A bit more... some of you who joined us as investors have not yet completed your paperwork with StartEngine. I know... it's a beautiful summer day out there, but this is super important to all of us here at SapientX. Thank you to the best investors in the world! David Colleen
SapientX has worked in 40 languages for several years now. It may surprise you to know that Alexa only speaks 15, Google 18 and Siri 21. To continue our leadership in language support, I'm happy to report that as of last week, we are now able to automatically detect, within 3 seconds, what language a user is speaking in and offer to respond in that language for them. Cool hunh.
Our first use of this is for an airport assistant in Canada. Their top three languages are English, Canadian French (Quebecois) and Chinese (Mandarin). We plan to begin live public testing of this system this fall.
BTW... last weeks launch of the beta site is going super well. We will be inviting more of you in this week. Thanks for your patience as we ramp this up.
Another change of topic... some of you who joined us as investors have not yet completed your paperwork with StartEngine. I know... it's a pain, but it's super important to all of us here at SapientX. Thank you to the best investors in the world! David Colleen
I'm really proud of my co-founder Bruce Wilcox. He's been hard at work updating our ability to handle off topic questions with a new machine learning system that's answering almost anything that we throw at it. Combined with our highly accurate symbolic reasoning system, we are getting conversational results way beyond anything that I have ever experienced before. I can hardly contain my enthusiasm!
And today, we will begin to share it with you.. our investors. At first, invites will go out to just a few... then more and more of you. Sage, our general speech system, is not perfect but with your help as test pilots, I believe we will be the best. You will get to try Sage via five different characters:
Enjoy!
Kendall, at StartEngine, tells me that some of you that invested in SapientX have yet to hit the okay button one last time to complete your investment. If you need help, please just ping the friendly folks at SE and I'm sure they will be happy to help.
Once you do that, please get off of that computer and go enjoy some summer weather!
David Colleen
Last year, I wrote an essay on how machine learning was being used… or misused in our conversational AI world. Since then, systems like GPT3 have begun to challenge our views on the use of ML. Unlike the older ML systems, used by Amazon and Google, GPT3 seems to achieve over 90% accuracy in giving a good conversational response. The largest issues are that GPT3 is not yet open source, it only runs on a $200k computer and it needs a lot of electricity to run.
That said, as good engineers, we are taking notice and we are getting promising results in our initial tests. Our goal is to take the best of what SapientX does and have GPT3 handle what we are weakest at. I’ll share the results when we are further along.
We are also using ML to automatically determine what language people are speaking and to determine who is speaking so that our avatars can address multiple people in an audience.
These are exciting times for us and I’m happy that you are co-pilots with us on this journey!
David Colleen, CEO
2021 Essay
We know the old fairy tale. The clothes-less king has all of his subjects convinced that he is clothed in luxurious appointments… but one small boy has the nerve to point out that the king is actually naked.
Such is the case with artificial intelligence (AI) and machine learning (ML). Broadly, artificial intelligence is any computer process that emulates how humans do things. Over the last 65 years, the AI tree has grown many useful branches of exploration. Each branch became like a unique tool in the toolbox of an AI engineer. Like any good engineer, it’s important to know which tool to use for each task. The newest tool at their disposal is machine learning.
ML has become one of the most hyped technologies in recent years. This hype has become so prevalent that I had an AI professor in Sweden tell me that if it’s not ML, it’s not AI. This is of course silly but it’s an example of how many people are drinking the hype-filled, ML Kool-aid. This is a dangerous error because ML is now being used for tasks that it’s not good at.
You may have seen the press reports about 40% of startups that claim to use AI are falsely claiming it. Well, this meme is only partially correct. To be sure, some of the companies lied about their use of AI, and some even secretly used humans in a ruse to deceive investors and customers. That said, many of the persecuted companies do use AI… but not ML. Sadly, this faulty reporting delivers a memorable message to the point that many now believe ML to be synonymous with AI. This is bad for our industry.
Machine learning is difficult to talk about because no one knows how it actually works. Machine learning is a bit like pattern recognition in that as you feed massive amounts of data into the system, it begins to recognize common patterns and their relationship to a response. Because ML is statistical, it never actually “knows” or “understands” anything… it generates a “best guess”. Consequently, ML systems generate false responses approximately 1 time in 4.
As I wrote in a previous post, conversational systems began as rule-based chatbots in the ’60s. In fact, IBM Watson, Apple’s Siri, Google Assistant, and Amazon Alexa all began as rule-based systems. All four companies now tout their systems as being ML-based. The truth is that the companies only began to add machine learning to handle user requests that the rules did not address and today, their reliance on ML varies. In benchmarks, conducted by ZDnet and Ark Invest, these voice assistants generated a significant error rate with Google at 12%, Siri at 25%, Alexa at 27%, and Watson at 29%! In our own testing, Alexa has actually declined in accuracy as they have increased the use of ML in their system over time. These error rates are consistent with typical ML performance and it’s a major limiting factor for the use of these systems.
When Bruce Wilcox began designing what became SapientX, he ignored prevalent trends, and in a very practical approach, he focused on which AI tools worked best. He was inspired by symbolic reasoning research going on at Carnegie Mellon at the time. Symbolic reasoning is a form of rule-based AI-augmented by part of speech tagging, pattern recognition, and concept matching. Since our system analyzes inbound text like your old English teacher, we argue that, unlike ML, our system actually can understand what you say. The proof is in the results. We manually train our systems until they achieve 99% accuracy. In the wild, actual results vary. Because we can achieve these accuracies, we are a better choice for mission-critical applications such as controlling your car.
To be clear, we do use ML… for tasks that it is best at such as speech recognition, question answering, summarization, and the generation of synthetic speech. Like any good engineer, we ignore fashion and use the right tool for the task!
That's right. Our Beta test program will be launching in July. If you invest $500 or more in SapientX, you are automatically enrolled as one of our test pilots. Test pilots help us to identify strengths and weaknesses in our conversational characters so that we can make SapientX better and better. Want to be part of our exciting future... invest now. Already an investor... sit back, you will get your invite this summer as we add more and more of you to the test!
Xander's unique. I'm not just referring to his looks, his speech or the game level and characters that he keeps company with... I'm talking about that as near as we can tell... he's the first commercial use of a life sized, 3D conversational character. At Xander's launch, a few weeks ago at AWE, we made videos of people using him, but it's a bit hard to hear what was being said. I decided to just talk to him in my office with a good mic so that you could see what a conversation with him is like.
Our whole team is excited about Xander and how we are going to use his siblings to power digital signage applications. I went to InfoComm, a huge digital signage show in Las Vegas two weeks ago. Signage companies really took notice and we have begun conversations with several of them. We have also agreed to use Xander at three more shows in a few month and we are working on an airport guide name Edelle (she's kind of a cowgirl) for use at two airports. This feels big to us. Enjoy the video. David
Interview of SapientX CEO David Colleen by Dominque Wu, a consultant at Facebook. David discusses the history of high resolution character development and the next steps that the industry needs to follow to make successful conversational characters.
SapientX had a great reception at the recent TC Mobility show in Silicon Valley. We showed Visteon Via, a digital assistant for cars. Visteon is one of the largest auto components suppliers in the world... selling to all of the major brands.
{{ profileCtrl.commentsLoading ? 'Loading...' : 'Show More Comments' }}