Feels like as an industry we are going down a path of trading bloated Eng teams for “token bloat”, where “slop” (writing, technical systems, product) is a clear symptom.
Been feeling that agents have broken planning projects w/ cycles in @linear. Cycle view only affords for tickets assigned across team, but planning out projects at ticket level feels antiquated. Seeing who is owning which milestones feels closer to what I'd want. Curious how @karri and @linear team or other teams are planning these days.
Luis Ramirez retweeted
If 2025 was the year of agentic coding, 2026 will be defined by computer use agents.
Insurance is the #1 market for computer use. Many carriers and brokers run on hundreds of systems: desktop apps with no APIs, green screen terminals that can't be accessed outside the mainframe environment, and third party portals.
This is where most of the manual back-office work lives, yet it's also where most automation has broken down. Insurers became the largest buyers of RPA for exactly this reason, but bots built to mimic clicks created a fragile maintenance treadmill.
Computer Use Agents (CUA) change that by operating any system a person can, logging in with 2FA, viewing the screens, reasoning through edge case updates, and acting in real time.
Until March, CUA was too slow, costly, and inaccurate to be run in production. With the release of Sonnet 4.6 and GPT-5.4, accuracy on our internal evals jumped from 30% to 94% overnight. Today, long-running tasks spanning >30 min run at 99%+ accuracy in our computer use harness.
For our customers, computer use agents that were completely out of reach six months ago are now in production.
Most insurance work still lives in systems that were never built to talk to anything else: policy admin platforms, claims systems, the mainframe you reach through a green screen. Still, these are the platforms that hold the data the mission-critical work depends on.
Computer use removes that limit. If a person can operate a system, a @pacecom agent can too: it sees the screen, reasons about what's on it, and acts. We've pointed it at the trickiest UIs out there - cluttered policy platforms, 40-year-old mainframes, systems behind two-factor logins - and because it reasons over screens instead of following hardcoded clicks, it doesn't break when a field moves.
This is the work nobody could automate, till now. See how we build production-grade multi-modal agents here
Luis Ramirez retweeted
These guys will probably be top 3 fastest growing autopilot company this year.
Today, Pace has raised a $46 million Series B, co-led by @ThriveCapital and @Sequoia, with participation from @emergencecap and @pruvencapital, to help our customers insure more of the world’s risk.
@pacecom agents have completed more than 250,000 critical insurance operations, growing 3x every quarter. The world’s leading insurers, like Prudential, WTW and Convex, trust Pace to scale back-office operations.
60% of the world's losses last year went uninsured. Closing this $9 trillion protection gap starts with AI-native operations.
I joined Pace last year with two goals: work with a top-tier team and build at the frontier of agents.
Since then, we’ve built an agent harness that can reason over extremely complex documents and rules, handle long customer interactions via phone, use browsers and desktops, and use tools to move work forward. These capabilities work seamlessly together to ultimately complete insurance workflows end to end. Pace agents have now completed over a quarter-million tasks for our customers and growing fast.
Entering this next phase of the company I’ll be doubling down on what makes this all possible: building out our product & engineering team. If you’re a designer or engineer also looking to work at the agent frontier alongside a world class team, we’re hiring.
ats.rippling.com/pace/jobs
Today, Pace has raised a $46 million Series B, co-led by @ThriveCapital and @Sequoia, with participation from @emergencecap and @pruvencapital, to help our customers insure more of the world’s risk.
@pacecom agents have completed more than 250,000 critical insurance operations, growing 3x every quarter. The world’s leading insurers, like Prudential, WTW and Convex, trust Pace to scale back-office operations.
60% of the world's losses last year went uninsured. Closing this $9 trillion protection gap starts with AI-native operations.
🫡
The 2026 Enterprise Tech 30 is live! 🚀
60 companies shaping the future of enterprise.
In partnership with @EricNewcomer at @NewcomerMedia
See the full list: wing.vc/et30 #ET30
After 1 year of working @pacecom, I feel more and more bullish every day that we're building something special: constantly at the edges of model capabilities, squarely at the intersection of services and software, and in an industry ripe for this kind of transformation.
We're hiring across FDE, GTM, and Eng. DMs open.
Sequoia’s @JulienBek says many of their founders are now wondering if they’re “just an iteration away” from AI labs destroying their business.
He says the most defensible companies - and potentially the next trillion-dollar company - will be “a software business that masquerades as a services firm.”
“If you sell tools today, you’re really in the line of sight for the models and you’re effectively competing with the next generation that they’re going to launch.”
“Whereas if you sell the work, you’re actually benefiting from what the models are doing and all the billions of dollars that are going towards AI.”
Luis Ramirez retweeted
Anthropic just released Sonnet 4.6, their most capable model for computer use. We've been testing it with the @claudeai team ahead of today's launch.
It hit 94% on our Complex Insurance Computer Use Benchmark. Early Claude models scored around 30% on the same tasks.
For the first time, the "mid-sized" Sonnet model has outperformed the heavier Opus 4.6 on our hardest tasks.
Pace agents navigate carrier portals, legacy systems, and dense insurance interfaces thousands of times a day. This requires reasoning through edge cases in environments that weren't built for automation.
We partnered with the @AnthropicAI team to improve performance on exactly these kinds of workflows. We've seen Sonnet 4.6 self-correct through failures in ways previous models couldn't — and we're already shipping it to our customers.
For our insurance partners, this means higher straight-through processing rates and faster turnaround for their policyholders.
Grateful to the team for the partnership and the shoutout in the launch blog post.
One thing I love about the team and culture that we've built so far is that getting every single one of our pilots to production is a team effort. Our FDEs lead the charge and everyone across the org swarms when needed, working closely with customers to ensure success. Some of this is structured processes, but most is just deeply embedded into the culture.
The insights we glean from these direct interactions get fed back into our next product iteration, GTM playbooks, etc and compound over time.
"How did you pull that off? That's not normal."
That was @gradypb's reaction when @jamiecuffe shared that 100% of our pilots have gone to full production.
In the current AI hype cycle, it's easy to build a demo that dazzles.
It is much harder to build a product that survives real insurance operations: messy data, complex edge cases, and zero tolerance for error.
We don't just ship software. We do the hard work upfront to understand the domain as deeply as the customers we serve.
Jamie joined Pat and @laurenmhreeder on @sequoia's Training Data to discuss what it takes to become the AI operations partner for the world's leading insurers.
Watch the clip below. 👇
Luis Ramirez retweeted
$10M led by @sequoia to build the AI operations partner for insurance
Tens of thousands of complex transactions completed
1/100th the workforce of traditional BPOs
Read the full post from our founder below ⬇️
Today we’re announcing Pace has raised $10M led by Sequoia to build the AI operations partner for the world’s leading insurers.
Insurance exists to do something remarkable: turn uncertainty into security. But between that sits an enormous amount of operational friction: processing submissions, servicing policies, and adjudicating claims.
Leading insurers have already completed tens of thousands of these complex insurance transactions with @pacecom agents in production. Check out our team showing what this looks like in practice: agents which work across systems and modalities to complete work end to end.
Luis Ramirez retweeted
Building the agentic workforce for the insurance industry is a massive technical challenge.
It requires the best engineering talent in the world.
We are proud to have Kim on board to help us automate operations for the world's leading insurers.
The bar is high, but the mission is worth it.
The best engineers get their pick of where to build the future
When you're building agents for the world's leading insurers, talent density isn't a luxury
That is why I am thrilled that Kim is on our team at @pacecom
She was previously Employee #1 at Captions/Mirage. She knows how to ship AI at massive scale.
Now, she's bringing that firepower to the insurance industry
Thank you Kim, excited to build the future of insurance operations with you and this team 🚀
Getting agents reliably into production and gaining trust from leading customers in insurance is a huge task. Extremely proud of the team and product we've assembled to make this possible.
From pilot to live in production: Prudential has chosen Pace to automate its insurance operations. Pace AI agents are now streamlining policy servicing & QA for Prudential’s ILI business, delivering speed & operational efficiency. With thousands of hours saved on these back-office tasks, the Prudential team is spending more time focused on what matters: building customer relationships and delivering industry-leading experiences.
Pace has already built an insanely talented team across Eng, GTM, and FDE. Huge opportunity for anyone looking to learn from some of the best across the entire business.
Forward deployed eng at @pacecom means a lot to me — it's where I started at Retool.
Retool's FDE team launched 10+ startups: @apurvas96, @domwhyte42, @SushanthRaman, @ryan__jiang, @theodormarcu, @justoutquan, @christieh97, @realchillben, & more.
FDE is the path to founding 👇
Luis Ramirez retweeted
At Pace, we’re building the AI-native BPO for insurance.
Insurance carriers process hundreds of billions of documents every year, each with its own nuance and structure. We’ve worked with OpenAI to apply GPT-5 to mission critical insurance operations tasks and it is setting a new standard for performance on every benchmark.
