Why AI agents are buzzy again, and more AI and tech news from this week

Welcome back to LinkedIn News Tech Stack, which brings you news, insights and trends involving the founders, investors and companies on the cutting edge of technology, by Tech Editor Tanya Dua. You can check out our previous editions here.

Today’s edition dives into how the buzz around AI agents is picking up again, and explores where we are in terms of their development and adoption. We also have a rundown on the latest advancements from the world of AI and other tech news you need to know.?

Don’t miss the latest edition of VC Wednesdays, where Battery Ventures ’ Neeraj Agrawal joined me to discuss his T2D3 growth metric for scaling SaaS startups, why the second generation of software is all about automation and more.

Pitch me the interesting investors, founders, ideas and companies powering emerging technologies like AI to reach the inboxes of nearly 1 million subscribers plus thousands more on LinkedIn. Follow me for other tech updates. And click 'Subscribe' to be notified of future editions.

A deep dive into one big theme or news story every week.

Missed the first five minutes of your meeting? You can already use Copilot or Gemini to summarize what you missed, and help you draft a follow-up email to your team.

But in the near future, AI software known as agents could take things up a notch and make life far more seamless: automatically reserving you a conference room, preparing you an agenda and even accessing your files to make relevant changes to your project documents after the meeting.?

While talk of “AI agents” has been swirling about for a year now, recent weeks have seen renewed buzz around these AI systems that can autonomously execute on a range of tasks by interacting with and learning from their environment, with startups and incumbents both racing to unlock this next frontier in AI as usage continues to climb.

• Just this week, LinkedIn parent 微软 released new features that will let companies build AI agents in Copilot Studio, even as OpenAI , 谷歌 and Meta tinker on their own versions of AI agents.?
• Meanwhile, a new Paris-based startup called The H Company has raised a $220 million seed round to build “agentic” foundational models that it says will be more capable of reasoning, planning and collaborating than its competitors.
• And a third company called AILY LABS has rolled out an add-on “agent” feature to its mobile app, which founder Bianca Diana Anghelina says provides personalized guidance to enterprises to optimize their business goals.

Glasswing Ventures ’ Rudina Seseri attributes the growing number of agents to increased experimentation by businesses, which will only increase — though it’s still very early days.

“Consider Klarna ’s big announcement that two-thirds of their customer interactions are now being handled with AI agents,” she said. “With creativity, the use cases for which agents are best suited are endless. In 2024, I believe we will continue to see similar headlines as enterprises transition from areas of experimentation to full implementations leveraging AI to create strong business value.”

Another reason such agents are taking off now is simply because developers are getting better at system design, or getting LLMs to work more efficiently with other softwares, said Sandhya Hegde , partner at Unusual Ventures .?

Most AI agents currently in production work by “tool use,” or combining a large language model with other software via APIs to be able to execute a certain task, she added. A travel company, for example, could use GPT, its own aggregator and your credit card on file to then potentially create hyper-personalized recommendations and further, even book your trip.?

For enterprises, AI agents mark a step-function change in realizing AI’s true potential, as they can theoretically make workers more productive by asynchronously taking over entire tasks, also saving time and money. That explains why a slew of startups are developing AI agents across a range of functions, from sales development to customer service.?

Former Salesforce CEO Bret Taylor ’s new company Sierra is building AI agents for customer support for companies including WeightWatchers. Then there’s 11x ’s autonomous AI agent Alice, which can research potential clients on behalf of sales professionals and reach out to them all on its own. Another one is Sema4.ai, which can create ‘digital twins,’ helping banks, for example, detect, troubleshoot and remediate fraud.

“It is complex to automate enterprise workflows, and augmenting the people working on them is even harder,” said Mayfield ’s Navin Chaddha . “AI agents combine actions, intelligence and enterprise context, helping teams interact with LLMs to complete meaningful work faster.”

Indeed, AI agents are starting to become useful in some business contexts, according to a new report from Amsterdam-based software developer and investor Prosus Group. But agents are still far from perfect, as many continue to get caught up in loops of actions, making their execution unreliable.

And that’s why, while some businesses have started to pilot AI agents, “back-office operations that are not customer-facing are where enterprises will feel more comfortable experimenting with agents first,” said Unusual's Hegde.

These issues could soon be addressed as techniques such as grounding take off, where AI systems can fact-check the results that AI models produce, improving their problem-solving and reasoning. Plus, the likes of H and Adept AI are trying to build fundamentally new models that go beyond API-based tool use, where agents are able to actually learn how to execute on tasks, said Hegde.?

Still, not everyone is convinced that AI agents are ready for the mainstream.?

“AI agents are self-aware, they have memory and context, and they interact autonomously with other agents,” Norwest ’ Scott Beechuk told me recently. “We’re probably a year away from the first set of AI agents to really hit the market.”?

Here’s where we bring you up to speed with the latest advancements from the world of AI.

• AI is coming to Alexa. 亚马逊 is adding generative AI to its 10-year-old voice assistant — and the upgrade will be accompanied by a monthly subscription fee, CNBC reports, citing unnamed sources. The "more conversational" Alexa is set to debut later this year and help the e-commerce giant better compete with new chatbots from Google and OpenAI, which some have called "Alexa and Siri killers."
• OpenAI has dropped its Johansson-like voice. The ChatGPT-maker has pulled one of the voices of its viral ChatGPT-4o assistant following accusations that it sounded too much like Scarlett Johansson in “Her” — a 2013 film in which Johansson played an AI system. Johansson released a statement Monday evening saying she was approached to be one of the voices for ChatGPT, but declined.?

• Humane is exploring a sale, Bloomberg reports, citing unnamed sources. While still early in the process, it's looking for between $750 million and $1 billion. The move follows the rocky public launch of its wearable AI device last month, which was meant to rival Apple devices. Priced at $699, plus a monthly subscription of $24, the palm-sized wearable was pitched as a computer-and-voice-assistant hybrid that could reduce screen time. But critics and users had an extensive list of complaints, including overheating and a brief battery life.
• 英伟达 blew past Wall Street forecasts in its first-quarter earnings report. The company posted $26 billion in sales on Wednesday — a 262% increase from one year ago. Nvidia, whose chips power everything from Google’s Gemini to OpenAI’s ChatGPT, is also projecting $28 billion in sales for the current quarter. This was the third consecutive quarter in which Nvidia has seen growth exceeding 200%, signaling that spending on AI remains strong.
• AI know-how can earn you more money. A new study found that jobs seeking candidates with AI skills offer an average of 25% higher pay than traditional roles. PwC's AI Jobs Barometer is based on more than half a billion job postings in 15 countries. The premium for AI-related roles is higher in the U.S. than Canada (11%) or the UK (14%). Professions such as law and finance have potential for even bigger increases. The likely reason: Productivity is growing nearly five times faster in industries "most exposed to AI," the study found.?

Here’s a list of other notable AI developments from this week:

• Scale AI , which helps companies label and test data for AI model training, has raised $1 billion in Series F financing, valuing the six-year-old startup at nearly $14 billion, CEO Alexandr Wang shared on LinkedIn. The round was led by Accel , with existing investors including Spark Capital , Founders Fund and Tiger Global Management and new investors such as Elad Gil , Intel Capital and AMD Ventures also participating.
• Patronus AI , a builder of governance and compliance tools for LLMs, has raised $17 million in Series A funding, co-founder and CEO Anand Kannappan shared on LinkedIn. The funding comes just eight months after the company announced a $3 million seed round.
• Coactive AI , which helps companies make sense of their unstructured visual data, has raised $30 million in Series B funding, co-founder and CEO Cody Coleman shared on LinkedIn. The round was led by Laurene Powell Jobs’ Emerson Collective and Cherryrock Capital , a new venture fund from former TaskRabbit CEO Stacy Brown-Philpot . Other investors included Andreessen Horowitz , Bessemer Venture Partners and Greycroft .?
• Abu Dhabi-based G42 and Microsoft are investing $1 billion in Kenya to fund a Microsoft Azure data center and an open-source large language model in Swahili and English.

Catch up on the tech headlines you may have missed this week and what our members are saying about them on LinkedIn.

• TikTok is planning ‘significant layoffs,’ The Information reports, citing unnamed sources. The layoffs will reportedly affect "a large percentage" of employees in operations, marketing and content, with employees being notified this week. The global layoffs come amid ongoing reorganization and suggest mounting pressure on the company to become profitable. TikTok is also facing a potential U.S. ban if it doesn't sever ties with its Chinese parent company, ByteDance. Billionaire Frank McCourt says his Project Liberty is forming a team with Guggenheim Securities to launch a bid for TikTok’s U.S. business.
• 苹果 is reportedly planning a slimmer iPhone. The company, which recently unveiled a thinner iPad, will unveil the device, code-named D23, in September 2025. It? will use Apple’s latest-generation processor and could include an improved front camera for selfies and video calls. The phone’s arrival would come as Apple faces increased competition from Huawei and Honor in China and Samsung Electronics. In related news, Apple has also announced eye-tracking technology to help people with physical disabilities navigate its devices.
• The FDA has approved a second Neuralink test. The agency has signed off on improvements to the device after some of its data-transmitting threads came loose in the first test patient, Noland Arbaugh. Neuralink plans to implant the threads deeper in the brain of the next patient, in the hopes they'll stay put. More than 1,000 people have signed up for the study, but fewer than 100 qualify, per The Wall Street Journal.?
• Airbnb is taking steps to meet surging demand for EV-charging points. The company has teamed up with ChargePoint to offer hosts up to a 36% discount on EV-charging hardware, which starts at $399. The promo also includes $100 off installation services. Additionally, the first 1,000 hosts to buy ChargePoint hardware get a $200 “early bird” discount, The Verge reports. Airbnb says the move is an extension of its in-home sustainability push, though it benefits from it, too: Searches for rentals with EV chargers jumped 80% between 2022 and 2023, and those listings are usually booked for more nights.

Here’s keeping tabs on key executives on the move and other big pivots in the tech industry. Please send me personnel moves within emerging tech.

• Techstars CEO Maelle Gavet is leaving the startup accelerator after nearly four years due to health reasons, she shared on LinkedIn.
• Karl Tuyls , formerly of Google DeepMind , has joined H Company as co-founder and chief research and operations officer, he shared on LinkedIn.

Thanks for reading. Please share Tech Stack and forward it around if you like it! And if you have any news tips, find me on InMail.