Month: April 2018

Can Artificial Intelligence replace lawyers? Perhaps sometime in the distant future, but in the meantime AI is already augmenting the work done by legal professionals as startups race to reach that ultimate goal.

One burgeoning player in the AI-powered legal tech space is Tel Aviv-based LawGeex, which has developed automated contract review technology to help companies sift through things like NDAs, supply agreements, purchase orders, and SaaS licenses, to ensure they’re aren’t any unsanctioned legal gotchas buried deep in legalise. Today, the company is announcing that it has closed $12 million in new investment.

Led by VC fund Aleph, with participation from previous backers, including Lool Ventures, the new round of funding will be used by LawGeex to further develop its product, and build a bigger presence in the U.S. where it recently opened a New York office. It brings the startup’s total funding to date to $21.5 million.

Designed to answer the question ‘Can I sign this?’ the LawGeex contract review system aims to significantly speed up and cut costs inherent with the contract approval process. The idea is that once a new contract is sent to a business, it is uploaded to LawGeex where a “first-pass review” of the contract is undertaken using the startup’s AI. This checks the contract against a company’s predefined legal policies.

“If everything looks good, we can automatically approve the contract for signing right then and there,” explains LawGeex VP Marketing Shmuli Goldberg. “If we spot any issues that need to be corrected, we escalate the contract to the legal team, and highlight the exact sentence they need to fix, and what they need to do to fix it”.

The desired outcome is that legal professionals no longer need to spend time reviewing problem-free contracts, and only spend a few minutes, instead of hours, on problematic ones. “We free up the time of whoever does that first review of the contract, be it a paralegal who takes a first look before sending issues on to a lawyer, or a contract review team who triage incoming contracts,” Goldberg says.

Put more simply, the LawGeex product operates a little like a spelling or grammar checker (see screenshot above). But instead of looking for specific keywords or language, the AI has been trained to understand technical legal language or so-called legalese. “It actively reads the contracts and “understands” the legal concepts. This means we can find and flag provisions even if they’re written in a way we’ve never seen before,” says the LawGeex VP.

To make all of this possible, over the last four years the company’s “recursive neural network”-based AI has been trained by feeding it hundreds of thousands of legal contracts, and having experienced U.S. lawyers annotate those contracts along the way. “We’ve now reached the point we can say that in certain cases, for example reviewing standard NDAs, our AI is actually more accurate than a human, as a recent study led by several academics at leading universities showed,” claims Goldberg.

from TechCrunch

Harbor helps businesses legally issue cryptocurrency tokens that represent ownership of real-world assets like real estate, fine art, company equity, and investment funds. This “tokenization” might sound boring, but it could be a big business that unlocks trading of illiquid property.

Harbor‘s intention to become a fundamental bridge between the offline and crypto economies has attracted a $28 million strategic round led by Founders Fund and joined by Andreessen Horowitz, Pantera Capital, and more. Following its $10 million Series A in February, Harbor has now raised over $40 million to dissolve the legal barriers to private securities tokenization.

“We think there’s going be a far greater appetite for owning real-world assets using the blockchain” than digital only cryptocurrencies, Harbor CEO Joshua Stein tells me. He expects it be like the impact “email had on snail mail”, but with value instead of content being sent back and forth. Once someone like Harbor handles the technical necessities to make transfers instant, free, and secure, people will exchange a lot more frequently.

The Harbor team

Here’s how Harbor works. Clients pay it in cash to make their tokenization of an IRL private security legal. Traditional trading of these assets can be complicated and expensive given there are often financial regulations or licensing requirements restricting who can buy and sell them. For example, foreigners or unaccredited investors without enough net worth aren’t allowed to own certain securities. The lawyers to handle these sales can be expensive, and the process can take weeks.

Normally, businesses have to be very careful about who they let buy these securities because they’re liable for a 20-year criminal sentence if they violate SEC law. With Harbor, a white list of eligible owners is established by an outside law firm that takes responsibility, and Harbor’s smart contracts refuse to process an illegal sale. Harbor effectively bakes securities law compliance like know-your-customer and anti-fraud/money-laundering into the tokens themselves so trades can happen instantaneously without legal assistance on every sale.

Harbor is hoping to launch this Regulated Token (R-Token) system with its first client this summer. The tokens are ERC-20 compatible so they can be sold on lots of cryptocurrency exchanges and stored in popular wallets. Stein stresses that investors will have to trust the underlying securities they’re buying. But they’ll get more trust in who owns something through blockchain transparency rather than some signed contract locked in a desk or vault somewhere. And they won’t have to trust who they’re selling to since the smart contracts only execute the trade if its legal.

The idea of making the way hugely valuable assets trade faster, easier, and cheaper led Harbor’s latest round to be oversubscribed. That’s even though it only came out of stealth two months ago from Craft Ventures, the fund and incubator run by PayPal mafioso David Sacks who sold Yammer to Microsoft.

Craft Ventures, Vy Capital and Valor Equity Partners joined this that included other new investors like Future Perfect Ventures, 1confirmation, Abstract Ventures, and Signia Venture Partners. Nicolas Berggruen of Berggruen Holdings, Napoleon Ta of Founders Fund, and Kyle Samani and Tushar Jain of Multicoin Capital also put in their personal money.  Sacks knew Ta, which set up Founders Fund to lead the round. Meanwhile, Stein says Harbor wanted to team up with Andreessen Horowitz partner and crypto thought leader Chris Dixon.

Harbor will have to compete with the other blockchain-for-securities startups like Polymath, which runs entirely decentralized and trustless infrastrucutre to the point that you have to hope strangers want their deposit back enough not to screw you on legal compliance, and tZERO, which is building its own full-stack compliance system. Harbor’s reliance on outside legal firms to build the smart contract white lists makes it more akin to a traditional financial player.

Harbor could make a lucrative business out of letting clients sell American securities to the Chinese market, which has shown a strong interest in crypto assets. Stein talks about “a crypto nirvana of a trustless environment” like a true Bitcoin bro. But his new A-list investors show Harbor is no pump-and-dump.

from TechCrunch

Lingumi, the London and Cardiff-based edtech startup that teaches English to kids aged between 2 and 6 using an app and a range of physical products, has picked up £1.2 million in seed funding.

Leading the round is ADV, with participation from existing backers LocalGlobe, and company builder Entrepreneur First (Lingumi was part of EF’s 5th cohort, which is turning out to be quite a vintage year). A number of unnamed angel investors also took part in the round.

Founded by CEO Toby Mather and CTO Adit Trivedi after they paired up at EF in late 2015, Lingumi has built a language learning platform for pre-school kids, initially targeting the teaching of English. Described as “digital-first,” it consists of an app designed around a curriculum of daily lessons, which can then be augmented with Lingumi’s physical products, such as ‘Play Cubes’ and ‘Jumbo Word Cards’.

In a call with Mather he told me that the Lingumi product that exists today is very different to the one he and Trivedi originally pitched at EF Demo Day back in early 2016, even if the mission remains the same: to increase access to learning a second language, based on a belief that a child’s earliest years present a “magic window” to do so.

Initially, the pair had developed a concept for a connected toy that controlled a learning app but pivoted to a subscription model after families kept hitting the end of the startup’s curriculum and requested more. In July last year, Lingumi ditched the connected toy entirely and switched to a “pure digital subscription,” up selling its now much simpler physical products separately.

“We’re increasingly aware of the exceptional ability of infants to learn a second language from very early in childhood, but access to English is typically restricted to the super rich, or to children aged 7 or older, as they begin to attend school,” says Mather. “Even there, they are taught badly and infrequently. We’re building an English learning methodology that is low-cost, can be used in the earliest years, and is effective, even if the parents themselves don’t speak English”.

The Lingumi platform works best when parents or caregivers participate, too. The app delivers a single lesson per day of around 20 minutes and purposefully limits screen time, hence the range of supplementary non-digital prodicts. “Children receive everything in English, through a playful learning programme built for pre-schoolers, but we encourage and train parents in their native language to play and learn with their children,” explains the Lingumi CEO.

“Multiple studies have shown the impact of this style of co-learning on outcomes. The learning method is also unique: unlike most curriculums, which focus around ‘edutainment’ or reading and writing skills, or teach older children via live video, ours is focused on constructive, natural spoken English in the earliest years. As we develop the curriculum, we’re continually leveraging our data on each child to improve the experience and learning trajectory for them”.

To date, Lingumi claims 10,000 users and Mather says typical customers are families with one or two working parents, “usually middle or working-class, aspirational families who understand both the fun, and the educational benefit of beginning a second language with their child”. It has customers in over 40 countries, but is mainly focussed on Western Europe, Taiwan, and, increasingly, China.

In fact, Mather says China is a potentially huge market and is in part seeing the startup pilot a version of Lingumi for kindergartens that want to begin teaching English, leveraging the company’s existing learning method and content.

To that end, the company plans to use the new seed funding to further develop its “digital and physical product ecosystem,” and scale the learning platform into new markets.

from TechCrunch

Kdan Mobile founder and CEO Kenny Su

Kdan Mobile, a Taiwanese startup that makes cloud-based software for content creators, announced a $5 million Series A today, raised from investors including W.I. Harper Group, Darwin Venture Management and Accord Ventures. Founded in 2009, the Tainan City startup says its products have been downloaded more 120 million times, with about 40% of its customers located in the United States.

Its Series A takes Kdan Mobile’s total funding so far to $6.5 million. The capital will be used for product development, including blockchain-based encryption for documents and real-time collaboration features, to appeal to enterprise and education users. The company also plans to spend more on user acquisition in the U.S. and China, two of its growth markets.

Kdan Mobile’s products include Creativity 365, a software suite with a mobile animation creator and video editor, and Document 365, launched last year to attract enterprise users. The company also recently began offering new subscription plans for businesses and educational organizations and claims that its cloud platform, called Kdan Cloud, now counts over 3.5 million members.

Founder and chief executive officer Kenny Su says Kdan Mobile is seeking new partners that will allow it to establish a bigger presence in markets like Japan. One of its Series A investors, Accord Ventures, is based in Tokyo, and Kdan Mobile may start marketing to the country’s animation industry, Su tells TechCrunch. The company already has partnerships with Taiwanese mobile services provider GMobi, Jot Stylus maker Adonit and Ningbo, China-based design sharing platform LKKER.

Su says one of the ways Kdan’s products differentiate from cloud-based software by Google, Microsoft, Adobe and other major competitors is its focus on artists, designers and other creative professionals. Kdan’s products were also created to allow users to start projects on mobile devices before moving onto desktop apps. As many users of Google Docs, Office 365 or Adobe Creative Cloud have discovered, accessing them on mobile devices feels much more awkward than on desktop. Kdan Mobile, however, was founded just as smartphones and tablets usage was becoming widespread, and its products were created specifically for mobile.

“We are trying to fill the gap, helping users create content on mobile and then allowing them to finish it in a desktop environment, not only with our own tools, but also by exporting to other places including Adobe,” says Su.

Part of Kdan Mobile’s Series A financing will also be used to figure out how to the company can increase the use of artificial intelligence in its products. Kdan Mobile already uses machine learning algorithms to improve its software by analyzing what users upload and recommend on its content sharing platform.

In a press statement, W.I. Harper Group managing director Y.K. Chu said “We are stunned by Kdan’s leading development technology and global vision. We are glad to be part of their development plan and expect to grow with them.”

from TechCrunch

Pluralsight, the Utah-based education technology company, has revealed its IPO filing. 

Given the timing of the unveiling, the company is likely targeting a May public debut.

Its core business is online software development courses, helping people improve their skills in categories like IT, data and security. Businesses small and large pay Pluralsight to help train their employees. It also has offerings for individual subscribers.

In the filing, the company acknowledges that it is a competitive landscape, and names Cornerstone OnDemand, Udacity, Udemy, LinkedIn Learning as others in a comparable market. It also mentions General Assembly, which was recently acquired by Adecco for $413 million. 

This is the first glimpse we get at Pluralsight’s financials. For 2017, the company brought in $166.8 million in revenue, up from $131.8 million in 2016 and $108.4 million in 2015.

Losses are growing, however. This is partly due to a sizeable increase in sales and marketing expenditures. For 2017, the company lost $96.5 million. This is up from losses of $20.6 million in 2016 and $26.4 million in 2015.

Pluralsight has been around since 2004.  Like many startups outside of the San Francisco Bay Area, the company bootstrapped its business and didn’t raise significant outside funding until 2013. Pluralsight previously raised nearly $200 million in financing.

The largest shareholder is Insight Venture Partners, which owned 46.1% of the shares prior to the IPO, an unusually high percentage. Co-founder and CEO Aaron Skonnard owned 13.4% and investment group ICONIQ owned 8.1%.

Morgan Stanley and J.P. Morgan served as lead underwriters. Wilson Sonsini and Goodwin Procter served as counsel.

Pluralsight plans to list on the Nasdaq, under the ticker “PS.”

A provision in the JOBS Act from 2012 helped make it so that companies could file confidentially and then reveal financials and other business information just weeks before making public debuts. This helps companies avoid too much scrutiny in the months leading up to an IPO. There is also a quiet period in this time, meaning that companies are limited in what they can say publicly about their businesses.

Like most tech companies, Pluralsight chose to take advantage of this confidential filing provision. But it also announced that it filed, something that companies don’t usually do. Most choose to stay quiet about IPO plans until they make the filings public, unless reporters break the news first.

It was no surprise to those who have been following Utah’s tech scene that Pluralsight is planning to list on the stock market this year. The venture-backed “unicorn” has been a late-stage company for several years now, with a reported valuation of $1 billion as of 2014. 

After a slow first couple months, there has been a flurry of tech IPO activity in recent weeks. Dropbox, Spotify and Zuora recently debuted. Pivotal, Smartsheet and Carbon Black are amongst the companies expected to list in the coming weeks.

from TechCrunch

“The percentage of Top 40 music made with our platform blows my mind” says Splice co-founder Steve Martocci. He tells me about some bedroom music producers who were “working at Olive Garden until they put sounds on Splice.” Soon they quit their jobs since they were earning enough from artists downloading those sounds to use in their songs. That led them to collaborate with famous DJ Zedd, resulting in the Billboard #12 hit “Starving”.

Splice has attracted $47 million in funding to power this all-new music economy. That might be a shock considering Martocci estimates that 95% of digital instruments and sample packs are pirated since they’re often expensive with no try-before-you-buy option. Even Kanye West got caught stealing the trendy Serum digital synthesizer.

But Splice lets artists pay $7.99 per month to download up to 100 samples they can use royalty-free to create music. That’s cheaper than it costs to listen to music on Spotify. Splice then compensates artists based on how frequently their sounds are downloaded, and has already paid out over $7 million.

Splice Sounds is like an iTunes Store for samples

“We try to make more seats at the table in the music business” says Martocci, who previously founded messaging app GroupMe which sold to Skype for between $50 million and $80 million in 2011. “GroupMe was made to go to concerts with our friends. Music has always been my motivator, but code is my canvas. Artists come up to me and hug me because I’m changing the creative process.”

Splice co-founder Steve Martocci

But now he’s getting some big name assistance, attracted by Splice’s success in the stubborn musician community and its $35 million Series B from December. Splice has just hired former Facebook product manager Matt Pake as VP of product to lead core teams in New York, and former Secret co-founder Chrys Bader to build out a new squad in Los Angeles. [Disclosure: I knew both from before they moved out of the SF social scene]

Splice now has 100 staffers, mostly hobbyist musicians themselves, but “I don’t think I have one bay area employee” says Martocci. He wants his offices where the artists live. “Everyone has a genuine passion for music. It doesn’t feel like a tech company as much” says Bader. Martocci apparently takes feedback well, which is different because “I’ve had some pretty fucking hard people to work with in the past…” Bader notes, likely referring to disagreements with his co-founder at Secret. “I have zero tolerance for bullshit at this point in my life and there’s zero bullshit on this team.”

While the Sounds marketplace has blown up recently, pushing Splice to 1.5 million users, the startup has a grander vision for software to eat instruments. That means creating the same kind of tools that help programmers code apps, but for musicians to compose songs. Splice Studio integrates with composition software like GarageBand, Logic, and Ableton to offer cloud-synced version control.

This might sound nerdy, but it’s a lifesaver. Splice Studio automatically backs up the artist’s work-in-progress song after every single edit so they can always reverse changes and safely work with collaborators without having to nervously save manually and fret about keeping all the copies organized.

Splice saves every edit to a song-in-progress so you can experiment but always reverse changes

Since Splice’s staffers actually make music themselves rather than parachuting into a foreign space, they intimately understand the frustrations they’re trying to solve. Knowing income can be unpredictable, Splice lets musicians access plugins, software, and instruments on a rent-to-own basis where they can pause payment and resume later. That’s the kind of convenience that Bader says makes Splice “easier than piracy”, echoing Spotify director Sean Parker’s plan to beat bootleg MP3s with a simple streaming service. “I wanted to build something even Reddit couldn’t complain about”, Martocci laughs.

But where Splice goes next could addresses the biggest, most insidious barrier to creative output: writer’s block. Ask most modern musicians, and they’ll tell you about their giant folders of unfinished songs. Getting from a melody rattling around in you head to a few tracks laid out in your preferred composition software is the easy part. Polishing those parts, ditching the unnecessary ones, finding the rights sounds, and tieing it all together into something listenable can be agonizingly difficult.

Creative Companion is Splice’s solution. Currently being built by Bader’s LA team, it’s a songwriting assistant that can suggest a next step and surface samples that fit well with those you’re already using. Martocci explains how Splice uses “cool machine learning stuff” to recommend ‘Hey, you should add a bass line. You should add some mastering.”

Splice just hired Chrys Bader, previously the co-founder of Secret

The question for Splice will be how many music producers out there are willing to pay. “There’s an upper bound. This is not a consumer product” Bader admits. Citing internal research, he says there 30 million music producers in the world. Many might not even know about Splice, “but at $8 a month, that’s not really breaking the bank. You might pay $200 for a plugin or $700 for Ableton. That’s insane. Musicans can’t afford that. Yet a musician friend tells me all the time ‘I’m broke, I’m broke…but I live or die by Splice.’”

Splice’s heavy-duty funding from Union Square Ventures, True Ventures, and DFJ could also attract competition. It might awake the interest of big creative services corporations like Adobe, or more established music production tool companies like Native Instruments which just launched a direct competitor called Sounds.com. But Splice is digging in for a long fight, giving away Splice Studio to lure in users and commissioning exclusive sample packs from top creators. In that sense, Splice is almost like a record label.

“I want to see a world with more transcendent musical highs” where “you have more music that’s ready for moment” Martocci opines. “If we build something that makes musicians lives better, that makes our lives better because a lot of us are musicians, what else is there in life?” Bader explains.

Computers democratized music-making, leading to a flood of amateurs sharing their content with the world. But all good democratizations necessitate layers of curation to sort through all the output, which social networks have become, and tools to let the most talented artists create what’s worth everyone’s attention.

Martocci concludes “Software is a great instrument. One-third of the world tries to make music at some point. They’re not going to pick up guitars and recorders any more.” Whatever app they choose, Splice wants to keep them in the creative flow.

from TechCrunch