Alexander Christie retweeted
We just signed an 8-figure compute contract.
@attio is increasingly becoming the place agents do work. That takes compute and tokens, a lot of them!
The next platform war isn't just about who stores your data, it's about where agents are allowed to act on it. Owning that is worth a whole lot of compute!
9
5
125
12,140
Looking forward to chatting with @richardmuirhead about what it takes to scale out CRM in the AI era!
The first [O]pen House door opens.
A fireside chat with Alexander Christie ( @byteofbits), Co-founder & CTO of @Attio, and @richardmuirhead.
Attio is building the fastest-growing CRM in one of software's most crowded categories, now leading the shift to agentic CRM. We'll get into the conviction to build anyway and what it takes to scale a category leader.
The conversations behind R[3]sidency x Construct have always been off the record. For the first time, the door's open.
Apply to join the room, 8 Sept → luma.com/attio-residency
1
1
13
611
Hey @BenyaminHolley,
We have a bunch of our customers (including ourselves) who are using Attio at way past 10m in revenue!
Appreciate that there are lots of areas that we need to improve but would love to get your detailed feedback on where things like our automations break!
1
7
496
I haven’t seen a convincing solve for having so many schemas at play in the per tenant model. Agree it’s architecturally elegant, but the overhead of migrations and schema management hasn’t made sense to me compared to the dSQL engines.
11
1,778
We use Google Spanner. I agree that giving agents SQL access is incredibly powerful, but I wouldn’t personally give them access to the actual database to do so.
We use a custom SQL engine with our agents which allows us to more granularly enforce certain conditions and prevent performance issues on the main datastore.
7
1,603
Just because you can do it, doesn’t mean it’s optimal. If you reread their blog post on this and think that’s a system you want to operationalise then we disagree about what to optimise for!
2
3
1,797
We’re fully on Google Cloud Spanner now, but if we couldn’t use that we would use one of the alternative dSQL engines instead
1
1
3
871
We migrated away from Postgres in our stack back in 2022, when a single large customer onboarding managed to irreparably brick the query planner.
We haven’t regretted that decision once.
The postmortem for our incident yesterday has been published, including the root cause analysis and our planned remediations.
We are deeply sorry for the disruption this incident caused to your team, your business, and your users. We understand that you depend on Clerk to be available, and we failed to meet that expectation yesterday, and too many times in recent months.
This is unacceptable, and we will be bringing increased attention to proactively adding monitors, redundancies, and failovers to our overall system. While we've added many throughout the past year, recent events have made it clear that we are not moving fast enough.
Thank you for your patience and continued partnership.
clerk.com/blog/2026-02-19-sy…
24
2
194
128,747
Auth is the one service I would never outsource. It’s not as hard to build and secure as the vendors pretend, it changes once a decade at best and historically speaking you are far more likely to have issues with your provider making changes than you are the auth layer changing!
1
1
19
1,748
This is a really clever abuse of Object Storage, but I can’t help but feel that this is actually a fairly complex wrapper around Spanner.
You get the same guarantees if you just write the whole thing to Spanner (as it powers the metadata of GCS) and you don’t need the buffer.
queue.json on object storage is all you need to build a reliable distributed job queue
→ FIFO execution
→ at-least-once delivery
→ 10x lower tail latencies
tpuf.link/queue
1
1
29
6,588
Hey Wayne,
We delay outbound sending to protect the reputation of your domain and mailbox (as Attio outreach goes through your mailbox directly). This is similar to other platforms like Outreach and Apollo and is best practise for delivery long term.
We’re exploring adding other email delivery options for sending larger batches of emails for marketing purposes! Would love to understand more about your use case so we can design a solution that works for you!
1
3
96
We’re somewhat protective over the IP of exactly how the data layer works but it’s a columnar storage engine built around Spanner with a custom in memory query engine on top. The query engine supports a few different ways of expressing the query depending on the workload.
No JSON involved!
1
7
2,424
It’s available on all paid plans - if you’ve refreshed the page or reloaded the app then please reach out to support and we’ll figure out what went wrong!
1
3
52
Today we're launching Ask Attio, our most powerful product update ever. It's a reimagining of CRM that puts agentic workloads at the heart of the platform.
Ask Attio builds upon months of deep technical work to unlock the full potential of @attio and enable agents that operate with unparalleled scale, accuracy and insight.
We call this system Universal Context: it combines semantic, full text and boolean capabilities to provide a transactionally consistent data model built to run game changing agentic workloads at scale.
And even better, Universal Context isn't constrained to Ask Attio - you can access all of the power of Universal Context through any tool you choose through our brand new MCP server!
We're so excited by the possibilities that Ask Attio and Universal Context open up for our customers - let us know what you're excited to try first?
attio.xyz/ask_now
10
13
96
18,491
We have some pretty cool things coming in this space 👀 Keep your eyes peeled for some updates in the near future!
3
49
Hey @AdvaitRaykar ,
Shoot me a DM - would love to get your feedback on something!
2
94
We're investing in files right now and expect to have endpoints available for the system soon!
1
2
75