Erik Munson retweeted
Try this product and you will never go back to the dark ages. Day AI is built for how founders & leaders actually think: Ask any question about the state of sales, customer sentiment, product requests - you name it - and receive a data-rich answer in seconds. Proud to support @markitecht , @MichaelPici , and the Day AI team on the board as Team Sequoia doubles down to lead the Series A.
Today we're announcing Day AI's $20M Series A led by @sequoia, and that we're now generally available.
We've spent 18 months building what we believe is the Cursor of CRM.
Here's what that means 🧵
Erik Munson retweeted
extremely grateful to work with the @Day_ai_app folks and power their AI.
we use Day for every call, and it's transformed the way we work. selfishly, it's also great to see the results of our own product being put to use!
absolutely well deserved
Exclusive: startup Day AI raised $20M from Sequoia to build the ‘Cursor for CRM’
CEO @markitecht tells @UpstartsMediaCo:
“Salesforce as we think of it today is not going to exist in 2 or 3 years. I don’t know if it’ll be us, but I really hope it’s us.”
upstartsmedia.com/p/day-ai-s…
Erik Munson retweeted
There are two types of AI startups: ones making incremental improvements, and others building patiently from the ground up for the future. Day.ai is in the latter, and they just raised $20M from Sequoia.
Congrats to @markitecht and Michael Pici. I was lucky enough to work with them in the early days of the HubSpot CRM and Sales Hub. Watching them attack this category again has been awesome.
Instead of building a typical CRM (values in fields) and slapping AI on top, they've built from the ground up with the idea that agents will be a core user. The foundation is built on a context graph where every datapoint has a chain of reasoning and sources behind it. An LLM and agent can understand not just what happened, but why which leads to better actions taken by agents.
Congrats again to the team!
Today we're announcing Day AI's $20M Series A led by @sequoia, and that we're now generally available.
We've spent 18 months building what we believe is the Cursor of CRM.
Here's what that means 🧵
Erik Munson retweeted
Exclusive: startup Day AI raised $20M from Sequoia to build the ‘Cursor for CRM’
CEO @markitecht tells @UpstartsMediaCo:
“Salesforce as we think of it today is not going to exist in 2 or 3 years. I don’t know if it’ll be us, but I really hope it’s us.”
upstartsmedia.com/p/day-ai-s…
this has been the biggest adjustment in mindset i've seen our early customers make... the realization that their CRM has *answers*, and all they have to do is ask. once you've seen it, it's impossible go to back.
our last Board meeting was a major turning point. every single question... "let's ask the product". every single answer made sense. it's the closest thing to "God view" I've seen in a business.
just watch the video... it's hard to put into words what @markitecht and @MichaelPici are building at @Day_ai_app
it's launch day! @Day_ai_app is now generally available, and we've raised a $20M Series A led by @sequoia.
we think you should expect much, much more from a CRM than tabs and lists. Day AI's customer context graph and long-running agent harness has unlocked an entirely new world of possibilities for our early customer partners. they're learning that when they have a question about their business, they can just ask.
i'm so excited to share it with more of you. if you think it sounds fun to work on wildly ambitious data platform and agent orchestration challenges, i'm hiring engineers in SF, Boston, and remotely in the US.
Today we're announcing Day AI's $20M Series A led by @sequoia, and that we're now generally available.
We've spent 18 months building what we believe is the Cursor of CRM.
Here's what that means 🧵
imo, the latent capability in today’s frontier models is where much of the real world economic impact will come from in 2026. continued model scaling will make a difference, but the unlock for most domains of work is now agent harnesses that bring out the skills models already have.
the winning application layer companies are obsessive about their harnesses. here’s why 🧵
@sequoia nails it: "Some of the most beloved products are known for exceptionally engineered agent harnesses"
they call out Manus, Claude Code, Factory.
the harness is the competitive moat now. not prompt engineering. application-aware infrastructure engineering.
if you’re building in this space, are you treating the world model around your LLM calls as core, mission-critical infrastructure? do you tune the minutiae of context compaction like your revenue depends on it? is everything your agent needs to complete a human-level task clearly modeled and explorable as data with correct access controls?
these are the new table stakes for winning companies at the app layer.
sequoiacap.com/article/2026-…
i truly wish cloudfront would compress responses when proxying (ttl = 0)... with the recent k/v upgrades it is so close to replacing nginx. really don't want to spin nginx up just to offload gzipping from my node api container.
building software in pairs is a lot more effective than i thought it was earlier in my career. haven’t quite figured out why i did such a 180 on it but i think part of it is experience. i’ve noticed i’m better at learning new stuff solo but can exploit stuff i’ve already learned better working with a buddy.
also perhaps a confidence thing - in my earlier career i was sometimes nervous folks might find my work less valuable if immersed in it up close. nowadays i appreciate the accountability and momentum of someone else being directly in my flow of work.
feels like there’s a lot of advantage in building a team where even folks with less experience and intrinsic confidence in their performance can experience pairing the way i do now, vs how i used to experience it.
the speed of this thing is unreal. it's unlocked new ways of exploring LLM data that were previously too slow to be practical, like pasting some random string from a customer report into the search box to see where it originally came from.
and equally as important, operating Brainstore inside our cloud account has been zero-drama, it's stateless and autoscales and has not had even one hiccup. not something you usually expect from large scale datastores. great to be able to let it do its thing in the background and focus on other parts of our infrastructure. very impressed.
