Connect with us

Startups

How AI will drive the hybrid work environment

Published

on

The future of remote work is a hotly debated topic, with employees pulling for more time away from the office and many employers preferring to keep workers where they can see them. In the end, we will most likely settle on a hybrid work environment that blends various degrees of work-from-home (WFH) and in-office activity.

As vital as it proved to be during the pandemic, however, WFH was not without its drawbacks. Poor connectivity, potential security issues, and convoluted workflows plagued many businesses as they tried to keep afloat during the lock-downs. So if this model is to flourish in the post-pandemic world, new technologies will have to step up to satisfy the needs of workers and bosses alike.

AI drives smart support

Artificial intelligence has the capacity to vastly improve work-from-home environments, bringing much-needed support to communications, collaboration, workflow management, and even security. But getting there will not be as simple as loading new software and letting it go to work. Enterprises across the board will have to carefully consider the ways in which AI should support the workforce of the future and the level to which it can influence workflows and business models.

The best way to begin is to establish a strong AI foundation, says Alex Smith, global AI product lead for knowledge work platform iManage. Since AI thrives on data, a central repository for all enterprise data is essential, and this can only be done in the cloud. In a world where access to data must be maintained for workers at home, in the office and anywhere in between, only the cloud has the capacity to deliver such broad connectivity. At the same time, the cloud makes it easier to search and share documents, email and other files, plus it provides advanced security, zero-touch architectures, threat analysis and other means to ensure access to data is managed properly – all of which can be augmented by AI as the data ecosystem scales in both size and complexity.

Once this foundation is established, organizations can strategically implement AI across a range of processes to help ensure the work gets done, no matter where the employee is sitting. Knowledge management, for one, benefits tremendously from AI to help identify those with the needed experience and skillsets to accomplish a particular project.

In a way, AI’s capability to support hybrid work environments upends one of the biggest fears surrounding the technology. Not only will it not take away your job, it provides the means to work where and how you want. In Cisco’s inaugural Global Hybrid Work Index, AI features prominently as one of the key drivers of work in the future. From July to September 2021 alone, the enterprise saw a 200 percent jump in the use of AI for key aspects of remote meeting engagement. These included noise reduction, automatic translation and transcription, polling, gesture recognition and other tools needed to maintain conversational workflows both in-person and in virtual settings.

A time of need

One thing the pandemic did accomplish for the hybrid work environment is to demonstrate that necessity does, in fact, breed innovation. All of the leading work and collaboration platforms jump-started AI development in response to the rapid rise of remote work. Webex’ Chris Rowen recently highlighted the myriad ways AI was used to improve the platform’s performance and incorporate new features. Audio intelligence, for one, helps ensure that only relevant speech is getting through, not barking dogs or sirens, regardless of how far the speaker is to the microphone. Real-time translation and closed-captioning has also been added to Webex Assistant, giving users the ability to converse in more than 100 languages.

As hybrid work becomes normalized in the digital workplace, we can expect AI to provide more than just a supporting role but to become an active participant. Platforms like Moveworks are forging ahead with conversational AI, which gives chatbots the ability to mimic human speech. As CEO Bhavin Shah explained at VentureBeat’s recent Future of Work Summit, conversational AI is driven by three key transitions in the enterprise: SaaS integration, enterprise messaging and advancements in natural language understanding (NLU). While the technology is by no means perfect, Shah says emerging techniques like advanced spell correction and statistical grammar models will allow chatbots to become more adept at reacting to conversation rather than predetermining it.

To be sure, some enterprises holdouts will resist the transition to hybrid work, largely on fears that employees who aren’t monitored will lack the incentive to give it their all. But while employees’ time at their home workstations may be difficult to monitor, their productivity is not. And with AI at the ready to analyze performance throughout the enterprise at a granular level, most organizations will likely find that work flexibility will be a boon, not a bust, to the business model.

Source link

Startups

With Pickme, your neighbor can receive your packages for you • TechCrunch

Published

on

French startup Pickme has raised a $3.7 million (€3.5 million) funding round to build a new network of pick-up and drop-off points for e-commerce retailers and logistics companies. With Pickme, a customer can order something and let their neighbor receive the package directly. It’s something that many people already do, but with a more formal relationship. Customers can then collect the package later when they have time.

OneRagtime is leading the funding round with Founders Future, FrenchFounders, Kima Ventures and several business angels participating.

If you’re usually not a home during the day, chances are you get a lot of missed delivery notifications. That’s why many people choose to receive their packages to a pick-up point near their home.

Pick-up points tend to be regular stores that want to increase their revenue by becoming tiny logistics hubs. Popular pick-up point carriers in France include Mondial Relay, Relais Colis and La Poste’s Pickup division. Some of them also operate automated lockers.

These networks have been thriving for several reasons. In addition to a growing volume of online orders, peer-to-peer marketplaces like Vinted, eBay and Leboncoin have also been doing well.

Pick-up locations can also act as drop-off points. Delivery persons can collect outgoing packages as they hand off some packages. As a bonus, using pick-up points reduces CO2 emissions as delivery trips are shorter and more efficient.

Pickme wants to offer a third way. Instead of choosing between home delivery and a regular pick-up point, they could choose to get their packages delivered to their neighbor’s home.

So far, 96,000 people have created a Pickme account and the startup has signed partnerships with DHL, GLS, Colissimo and Geodis. Individuals decide when they are at home so that they can handle packages. They can also define a maximum limit of packages depending on the size of their home — for example, it can be 25 packages or 50 packages.

When a customer collects a package, the neighbor receives a bit of money (less than €1 per package). Right now, Pickme processes 30,000 packages per month. It competes with Welco, another startup with a similar positioning.

In dense urban areas, there are already plenty of retail stores that act as pick-up points. But startups like Pickme and Welco could be more useful in rural areas where there are fewer shops. Right now, 30% of Pickme’s network members live in rural areas.

Source link

Continue Reading

Startups

Down rounds are still rare by historical standards • TechCrunch

Published

on

If you thought that the recent venture capital market was tough, let me tell you about 2016, 2017, 2018, 2019 and 2020.

With the first week of December under our belts, we’re not too far away from the end of the year. And that means that 2022’s venture capital story has largely been written. It’s not a single narrative; instead, this year started on a high, with momentum from the monstrous 2021 funding period persisting into the new year. From that point, we’ve seen a slowdown accelerate into what some consider a downturn.


The Exchange explores startups, markets and money.

Read it every morning on TechCrunch+ or get The Exchange newsletter every Saturday.


Startups raised lots of capital this year. Less, yes, than last year, but more than in nearly any year in recent memory. It’s still a good time to build a tech upstart.

Does that perspective feel too sunny when we hear so much doom and gloom on Twitter regarding startup prospects in a more conservative investing climate?

Source link

Continue Reading

Startups

Howie Mandel gets a digital twin from DeepBrain AI

Published

on

Howie Mandel is stepping into the metaverse. DeepBrain AI has created a pretty realistic AI version of comedian and actor Mandel.

Deepbrain AI, based in South Korea and Palo Alto, California, calls its creation “AI Howie,” and it’s an interactive virtual human and digital twin for immersive and personalized fan experiences. AI Howie mentions VentureBeat and talks to me in the attached videos.

Unlike the “deepfakes” of Tom Cruise and other actors, the real Howie Mandel cooperated with DeepBrain AI to create the virtual human AI replica of the famous comedian, actor, host, and technology enthusiast. We used “virtual Paris” AI character at our recent MetaBeat event in San Francisco.

“I am equally thrilled, excited, and terrified to finally have the ability of showing up and doing things without going anywhere or doing anything,” said Mandel, in a statement. “Thank you, DeepBrain.”

DeepBrain AI applies deep learning technology to create hyper-realistic virtual humans through its AI Studios and the AI Human platforms. These virtual humans are digital twins of the real person, with the same appearance, voice, gestures, and subtle mannerisms. The AI Studios platform enables script-to-video software that synthesizes dynamic video content in seconds, producing the quickest and most
realistic AI-generated videos. The script-to-video editor makes it easy for customers to select a model and then make it say something based on a script. Within a minute or so the video is made.

This is a powerful communication and marketing tool for celebrities, professional athletes, news anchors, and even politicians. Before working with Howie Mandel, the DeepBrain AI team created digital twins of Premier League soccer superstar Son Heung-Min, multiple news anchors across Asia, and South Korean president Yoon Suk-yeol.

Joe Murphy, business development manager for DeepBrain AI, said in an interview with VentureBeat that the virtual Howie is a conversational model that you pepper with questions. DeepBrain AI designs and develops these virtual humans for the purpose of creating digital twins (like Howie Mandel), digital people, and avatars.

It takes about four weeks of machine learning work to create a Howie Mandel digital twin.

“We create models of real people,” Murphy said. “We also have completely synthetic virtual humans. That is what we’ll call digital people. And then avatars are just the basic Roblox type of avatars. But where our technology comes in with the digital twins is we go through a deep learning process to clone the person’s voice, their mannerisms, their face, the way their eyes move, the way their lips move.”

He added, “So we create what we call the digital twin of the real person with all the uniqueness of that person. Our mission is to use this technology that we’ve developed throughout Asia and bring it to America.”

In addition to the script-to-video capabilities, the company provides fully conversational experiences with its AI Human software. The AI Human solution enables fans to interact and engage with AI Howie by simply asking questions. For example, when asked, “What was your favorite act on AGT this season?” the AI Howie model responds in real-time to support interactive, fun, and engaging fan experiences.

AI Humans are available within mobile apps, web browsers, or voice-activated kiosks.

“Our vision is to humanize digital experiences and empower creative teams to generate immersive content at scale,” said Eric Jang, DeepBrain AI CEO. “Working with Howie Mandel was a fun experience, and we are excited to see how the AI Howie collaboration will connect with his fans worldwide.”

DeepBrain AI, (formerly Moneybrain), a conversational AI startup based in Seoul, South Korea, has raised $44 million in a series B round led by Korea Development Bank at a post-money valuation of $180 million. The company started in 2016 and it has raised $54 million to date. The company has 130 employees.

The AI is being used for AI news anchors in South Korea and China at four different television networks. The networks flag that the anchor is an AI avatar so that no one gets confused.

The real Howie Mandel spent about a day shooting video with DeepBrain AI.

While multiple companies are working on virtual humans, DeepBrain AI’s avatars are hyperrealistic. One of Asia’s largest insurance companies is also using it, as is a “brand ambassador” for a soccer team.

“When we worked with Howie Mandel, we went down to his studio in Los Angeles, we provided a script, and fed our training data into our neural network,” Murphy said.

It took about a day to do a video shoot with Mandel and about three to four weeks of machine learning time on the computers to generate the first AI model.

Back in January, DeepBrain AI opened its office in Palo Alto, California, and it is talking to partners in Silicon Valley and the rest of the U.S. Over time, Murphy said that the hope is to create AI avatars in realistic 3D for the metaverse. In South Korea, kiosks are appearing in places like banks with both 2D avatars and 3D avatars.

Over time, Murphy said the avatars have gotten better at mannerisms, lip sync, and subtle gestures. The speed of real-time responses in conversations has also gotten faster. The company is talking about doing more with game companies and major brands.

Source link

Continue Reading

Trending

URGENT: CYBER SECURITY UPDATE