News Analysis
Research Report
The ROI of Employee Experience
An AI startup has emerged from stealth with one of technology's largest seed funding rounds, though whether its vision survives contact with enterprise reality remains an open question.
Humans&, founded in September 2025 by researchers from Anthropic, xAI, Google, OpenAI and Meta, announced in January that it had secured $480 million at a $4.48 billion valuation. The all-cash, unstructured round was led by SV Angel and co-founder Georges Harik, with participation from NVIDIA, Jeff Bezos, Google Ventures, Emerson Collective and numerous other prominent investors.
The company's founding team brings pedigree: Andi Peng from Anthropic, Eric Zelikman from xAI, early Google employee Harik, alongside Yuchen He and Noah Goodman. Beyond this roster, however, details about what Humans& plans to build remain scarce.
Table of Contents
- The Collaborative AI Thesis
- An Outlier Even for AI
- Collaboration Always Becomes Automation
- Execution Determines Success
- The Road Ahead for Humans&
The Collaborative AI Thesis
The philosophical framework driving this investment emerged during an interview on the No Priors podcast, where co-founder Zelikman outlined a vision that challenges current industry assumptions. AI has become trapped in what he calls a "task-centric trap," where most labs prioritize reasoning power whilst neglecting the emotional intelligence required for human collaboration.
The technical challenge at the heart of this vision is what Zelikman describes as the "stranger" problem. AI interactions feel like repeatedly meeting someone new because models lack long-term memory and fail to learn users' values or ambitions. They treat every interaction as a single-turn game, forcing users to repeatedly dump context into prompts.
Humans& is betting its $480 million war chest that it can solve this through compute investment, training models that move beyond automation to become partners. In this vision, AI doesn't merely book a hotel, but also understands constraints and goals that make that hotel the right choice.
The company's stated focus on long-horizon tasks and multi-agent systems suggests an architectural approach designed to coordinate complex work over time rather than automate isolated steps. This could theoretically support human collaboration, but the gap between technical capability and real-world deployment remains substantial.
An Outlier Even for AI
The funding round represents an anomaly, according to Lindsey Mignano, an attorney at SSM who works with AI startups from seed through Series B. "Humans&'s funding and valuation are highly unusual even by the standards of the current AI boom because of both the scale and the opacity of what it's achieved so early in its lifecycle," she said.
The $4.48 billion valuation makes Humans& a unicorn from day one. Most startups reach unicorn status after multiple rounds of funding and demonstrated product or revenue traction, not at the seed stage, Mignano explained. While large seed rounds exist in AI, this represents the extreme end of that trend, exceeded only by examples such as Thinking Machines Lab's record-breaking $2 billion seed round from 2025.
The valuation reflects betting on team pedigree and strategic positioning rather than proven market fit. In early-stage fundraising, investor confidence often stems from the long-term influence of founders and backers rather than capabilities. The involvement of strategic investors like NVIDIA positions Humans& within the broader enterprise AI infrastructure, though it also raises questions about whose priorities will shape the technology's deployment.
Collaboration Always Becomes Automation
The challenge facing Humans& extends beyond product development. Claims of "human-centric" AI hold only if deployment incentives align toward augmentation rather than efficiency extraction, and enterprise conditions suggest this represents the steeper uphill battle.
Most enterprise AI initiatives are measured by cost reduction and throughput. Under those conditions, even well-designed collaborative systems tend to drift toward headcount reduction or role compression. Increasing productivity without reducing headcount remains theoretically possible but historically rare, requiring choices about how work is measured, how accountability is shared between humans and AI systems and what success actually means for teams.
The tension is structural rather than technical, said Martin Mehl, a scholar-in-residence at California Polytechnic State University studying algorithmic management. "The architecture of profit will always find the path of least resistance — and right now, that path runs straight through labor cost reduction, not capability augmentation," he said.
Collaboration requires investment in human development, continuous retraining, interface design that respects worker autonomy and business models that share productivity gains, Mehl said. Surveillance and replacement require only deployment and extraction.
The incentive gradient proves brutal in practice. Companies using AI to monitor, measure and eventually minimize human involvement show immediate margin improvement. Companies using AI to help employees work better, make better decisions and be more creative show more engaged workforces and potentially better long-term outcomes that are harder to quantify on quarterly earnings calls.
The legal infrastructure compounds this problem. Labor law lags behind algorithmic management, with few data rights for employees and a dearth of frameworks that might require profit-sharing from automation gains or mandate transparency in AI-driven personnel decisions.
The predictable outcome follows: every "copilot" becomes a surveillance system, then a replacement roadmap — not because the technology couldn't augment human work, but because the only question being asked is "how much of this job can we eliminate?" rather than "how much more valuable can we make this person?"
Execution Determines Success
Natalie Spiro, CEO at Blue Fire Leadership, sees potential in Humans&'s technical approach while emphasizing that outcomes depend on implementation choices. "If Humans& is able to create an AI system that preserves context, supports human agency and keeps humans meaningfully in the picture, this could potentially revolutionize the way that we think about AI in the enterprise — not as a replacement layer, but as an intelligence partner," she said.
The focus on long-horizon tasks and multi-agent systems indicates an attempt to handle complex tasks over time rather than respond to one-off inputs. This could support human collaboration, but only if humans remain involved.
The central problem with "human-centric AI" is that efficiency pressures often cause companies to automate even when their stated goal is augmentation. AI increases efficiency without requiring immediate headcount reductions, but this often creates unrealistic expectations rather than reducing workload, unless managed well by leadership.
The real danger isn't near-term job loss but deskilling, overreliance on AI systems that humans don't understand and eventually losing human judgment altogether, Spiro said. If Humans& succeeds, the effect won't be primarily technical. It would force organizations to rethink workflows, incentives and ownership of outcomes. The technology supports collaboration, but only organizations decide whether to preserve it.
The Road Ahead for Humans&
As Humans& enters a crowded arena fighting for talent and hardware, the company remains focused on recruitment, seeking engineers and designers who bridge the gap between high-level reasoning and product interaction.
Webinar
On demand
Fuel Your Front Line: Workforce Trends Shaping 2026
Get expert insight into the tools and tactics top frontline teams will need in 2026.
Watch Now
Webinar
On demand
Fix the Disconnect Between Projects and Profit
Your margins are hiding in disconnected data. Let's find them.
Watch Now
Webinar
On demand
AI in Customer Service: Faster Resolutions, Happier Customers
Don’t let rising demand burn out your team. See how to build a smarter, more resilient support org.
Watch Now
Webinar
On demand
Accelerating Healthcare Ops with AI
Sign up to see how your peers are putting AI to work — and making it stick.
Watch Now
Webinar
On demand
From Nice-to-Have to Non-Negotiable: Prove the Value of Your EX Stack
Transform your tech stack: prove its value, secure your future.
Watch Now
Webinar
On demand
The Great Workforce Breakdown (And How to Stop It)
Disconnected employees. Failing communication. It’s time to turn things around.
Watch Now
Whether this philosophical pivot defines the next chapter of the AI era or becomes its most expensive experiment yet, the arrival of Humans& signals that the battle for AI's future isn't solely about technical capability. The $480 million question is whether "trust" and "connection" win in a market obsessed with automation—and whether any amount of capital overcomes the structural incentives that push collaborative systems towards replacement.
The answer likely depends less on what Humans& builds and more on whether enterprises are willing to reorganize themselves around augmented work rather than extracting efficiency from automated processes.
Editor's Note: Catch up on more action in the never-dull enterprise AI space:
- SharePoint Document Libraries Update Puts Copilot Front and Center — SharePoint document libraries got an AI update, but the change is less about UX and more about putting Copilot at the center of every workflow.
- Claude Cowork Is a Productivity Test Enterprises May Fail — Claude Cowork can organize files and write reports on its own. Enterprises still have to figure out who’s responsible when it gets things wrong.
- 3 Years After ChatGPT's Launch, Has Work Improved? — Three years after ChatGPT launched, generative AI has made its way into every facet of our working lives, to mixed results.