News archive
Recruiting, HR & AI news: November 13, 2026
Friday, November 13, 2026. 62 stories, each linked to its source.
Ranked by importance within each day. Pick a topic to see it newest first.
Friday, November 13
Get the news by email
Free newsletter
Get the Work Insiders weekly
One email a week: the recruiting and HR tech news that mattered, the tools we just added to the directory, and one practical guide. No spam, and you can unsubscribe in one click.
The University of Chicago Booth School of Business.
ZoomInfo is a go-to-market data platform with one of the largest B2B databases on the market.
Reason Machines, a Y Combinator Fall 2026 startup, has launched Reason, an agent harness that runs coding sessions on cloud VMs. The company says Reason runs tasks at 30 to 80 percent lower cost than Codex, Cursor or Devin and scored 65.1 against Codexs 62.5 on the Coding Agent Index at 22 percent lower cost per solved task, figures from its own reproduction rather than an independent test.
Nebuly, based in Turin and New York, raised a $21 million Series A led by Planven, with Exor Ventures, Tenacity Capital, TGVG Capital, Fastweb+Vodafone via Founders Factory and CDP Venture Capital also taking part. Its platform connects to the AI agents a company already runs, including custom tools and assistants built on Claude, ChatGPT and Microsoft Copilot, and measures hours saved, capacity freed up, which teams get the most out of AI and, for customer-facing assistants, revenue influenced and churn or upsell signals. CEO Francesco Signorato says companies have been counting users and prompts and are now making hiring and headcount decisions on assumptions about AI productivity rather than evidence. Customers include Oura, Iveco Group, Fastweb+Vodafone, CNH and Zeiss, and the funding will expand its go-to-market team across the US and Europe.
What its credits cost per endpoint, the request rates each plan allows, and what its API and MCP server return, now that it no longer automates LinkedIn accounts.
Deepnote has released the Agent Initiative Benchmark, an open-source test that drops AI agents into the Slack of PageMantle, a synthetic software company with six years of history across 11 services including a CRM, billing, product analytics, a data warehouse and an issue tracker. Across 50 workplace situations, each agent has to decide on its own whether to help, dig up the evidence, and post to the right people, or say nothing when the problem is already handled. OpenAI's GPT-6 Astra led the 12 models tested at 8.97 out of 10, GPT-6.1 Sol followed at 8.65 for about $0.11 per task, and Claude Fable 5.2 was third at 8.17 for $1.87 per task. Deepnote found that the priciest models were not the best, that models struggled most when their help was not needed, and that the models making the most tool calls finished last.
Cognition has launched Devin Apps, live private web apps that its AI software engineer Devin builds and hosts on its own cloud computer: a team describes the tool it needs, such as a metrics dashboard, project status page or review queue, and Devin builds it and puts it online behind a stable link. Every App requires a Devin sign-in and is never public, and sharing can be limited to named people, an organization or a whole enterprise. Because the Apps run on Devin's machine, they can use a team's existing MCPs, integrations and codebase, refresh data on a schedule, and keep the last working version live if an update fails. Machines sleep when no one is using an App and billing is based on how long they are awake, and Cognition says Apps suit internal tools and dashboards rather than high-traffic production services.
LangChain has raised $300 million at a $3 billion valuation, led by existing investor CapitalG with new investors GIC and DFJ Growth and existing backers IVP and Sapphire Ventures, bringing its total funding to $460 million. The money goes into LangSmith, its framework-agnostic platform for running, improving and governing AI agents, which the company says more than 7,000 organizations use, from Harvey, Rippling and Listen Labs to NVIDIA, Toyota and Salesforce. LangChain says Workday uses LangSmith to run and govern agents across the business and has cut the time to deploy a new agent from five months to under a day. The company plans to show new LangSmith features at a release event on November 20.
Paychex Announces Tax Penalty Protection Service.
Where to send qualified LinkedIn leads from traxy: Slack, HubSpot or Salesforce via Zapier, Clay, n8n, webhooks, HeyReach and Meta, and how routes work.
All Gemini app users in the U.S. can connect Google Wallet to get responses informed by their spending habits.
A listing says $39, the vendor says $199: an Interviewer AI review built on its pricing page, terms, privacy policy, help center and all 20 Capterra reviews, with the cost of 150 applicants a month…
antimattr, a Y Combinator Fall 2026 startup, is taking pre orders for Project Mnemosyne, earbuds and a smart ring for working with AI agents. The company says the earbuds filter out close to 90 dB of background noise and cut word error rate from about 30 percent on AirPods to under 5 percent. antimattr says it has more than 2,000 pre orders, with pricing starting at 299 dollars.
, a secure way for agents in ChatGPT Work, Codex, and dots to complete purchases with user approval and existing card protections.
New York startup Concentrate AI launched a healthcare AI gateway that gives companies handling patient or other sensitive data access to OpenAI's GPT, Anthropic's Claude, Google's Gemini through Vertex AI and open-weight models through one API and a single Business Associate Agreement, with Concentrate holding the agreements with each provider. In HIPAA mode, zero data retention is locked on, prompt and response logging is locked off, and requests only go to covered routes, including fallbacks and its automatic model router. Access costs $499 a month with inference billed at provider rates and no token markup, and the company says switching suitable workloads to cheaper open-weight models can cut inference costs by up to 70%. It is open to healthcare, legal tech, insurtech and fintech teams, with early customers including Pocket, Kinstead and Hummingbird Healthcare.
Hark, the AI lab behind the Hark Pro personal assistant, released Handoff², the second version of the computer-use agent that operates a browser to finish tasks like booking flights, buying groceries, scheduling appointments, claiming refunds and applying for jobs. Hark says it averages twice the speed of the original, with flight bookings up to 3.8 times faster and long research tasks more than five times faster, and it completed 82% of the company's internal 478-task, 238-website benchmark, up from 73%. The company says Handoff² lands within 5 points of Claude Opus 5.5 at its max setting while finishing tasks about twice as fast, though those figures come from Hark's own tests. Since Hark Pro launched in October, Handoff has run more than 5.6 million browsing sessions and handled over $20 million in purchases, and Handoff² is rolling out to all Hark Pro users now.
Fish Audio has opened Fish Agent, a no-code platform where you describe a voice agent, attach knowledge and tools, pick a voice and publish it to a phone number, a website widget or the API, with a median deploy time of 32 seconds according to the company. The agents run on Live 3, Fish Audio's new full-duplex speech model, which keeps listening while it speaks so a pause, a quick "mm-hmm" or a mid-sentence correction can shape the reply, and the company says it can start a short reaction in as little as 186 milliseconds. Fish Agent is in public beta for every Fish Audio account, Live 3 speaks English only for now, and enterprises can deploy it on-premises. Launch pricing on Fish Audio's pages is listed at $0.035 per minute (its docs show $0.06), plus phone-call charges.
Shield uses OpenAI to help IT technicians hand off work in legacy portals.
No stories on this topic yet.





