
GTM & Product executive; passionate about high growth and scale. Views are my own. Doing transformative work @promptql
San Francisco Bay Area
Joined January 2009
- Tweets1.7K
- Following356
- Followers483
- Likes640
You know that feeling when everyone needs something from you, and you're the only one who knows where to find it?
Another Slack message. Another person waiting on an answer.
Sasha and Emilia at @ExcedrInc know it well.
Now they're building a shared company brain with PromptQL, where people and AI work from the same context.
Less digging through Slack. Less chasing people. More getting things done.
Makes you wonder: how much time do we spend doing the work vs. just finding the person who knows?
A look inside Excedr's AI-native journey 👇
A founder’s knowledge should be an unfair advantage, not a bottleneck.
Excedr is putting 15 years of it to work for the whole team.
This July, Excedr raised $25M to help scientists lease expensive life sciences equipment to accelerate research.
This is their PromptQL story, and of how they've become one of the most AI pilled startups in the valley.
Jon started Excedr when he was frustrated with literally having to run uphill for lab equipment at Berkeley. So he started Excedr, a company that scientists now love to help them get the equipment they need.
But after 15 years of running Excedr he had a view of the business that was hard to give to everyone else, and as science research is starting to accelerate and things are scaling rapidly, Jon was the bottleneck to scaling up.
The reasons behind an underwriting decision. What finance learned from a particular customer. How the team works. The context that's obvious only if you were in the room.
At @ExcedrInc, they're bringing those pieces together in PromptQL, so people and AI can work from shared knowledge rather than each starting with their own fragment.
Jon calls it the Excedr-verse.
And hearing @jonchee_excedr, Sasha and Emilia talking about their multiplayer AI with shared context has utterly blown my mind.
It makes me so happy to be a partner in their journey to quite literally help make the world a better place to live in ❤️ 🏥
We went inside it for the latest 4min episode of "Inside an AI-Native Company"
96
Suku Krishnaraj retweeted
Today we’re launching something for the era where every AI logo looks like a butthole,
❌ Bots
❌ Dots
✅ Pots
Supports all models. Except Muse, obv.
Suku Krishnaraj retweeted
Man. If only someone had solved all of these at real enterprise scale. 🙃
Building multiplayer AI agents for teams isn't for the faint of heart.
Single player is easy mode.
Team level permissions and control have so many layers:
• Who owns the agent vs who can edit it?
• Personal OAuth or shared key?
• Which channel context can it actually trust?
• Who shares the device it executes on?
• What memories are private vs team-wide?
• How can i use my own agents but keep them private?
• 50 other things..
It’s not a feature that can just be added on.
It’s an entirely new OS layer and the biggest open opportunity in AI right now.
Suku Krishnaraj retweeted
Just came across this and it caught my attention.
Everyone’s arguing about which AI chatbot is best, while @PromptQL is taking a different route and getting multiple humans & AI bots to work together.
@davidsacks @jason this feels pretty close to a discussion you had on All in pod last week ?
Definitely will keep an eye on these guys…
Suku Krishnaraj retweeted
When your announcement is so massive that it's pre-empted by the besties 😅
@Jason @DavidSacks @chamath @friedberg, thanks y'all!
Suku Krishnaraj retweeted
INSANE moment at PromptQL! Just closed our first 8 figure deal at a Fortune 500. Excluding token costs tyvm.
I jotted down some learnings that might be useful if you're a founder in the AI space – esp. at the application layer.
My observations are downstream of 3 big changes:
⋅Unstable reference architecture
⋅Existential chaos on both sides
⋅Security
The six lessons:
1⃣ Navigators, beyond just champions
2⃣FDXs along with FDEs (forward deployed execs)
3⃣Tokens is the budget line-item
4⃣Customers want to deepen their moat, not just understand yours
5⃣Multi-tenancy is dead
6⃣Connect to the data, don't move it
Ofc all the old rules still apply. You need a real product, provide value, build a relationship etc. No shortcuts there.
Full article: nitter.cf/tanmaigo/status/209203…
Suku Krishnaraj retweeted
Fable 5 is a BEAST!
We gave 4 frontier models the same real job: clean up a bloated page in the wiki our team + our AI use as a shared skills repo.
Fable was the only one that worked by a huge gap, with a generalizable intuition.
~40 sections of accumulated notes, duplicated & stale context, broken links.
THE SETUP
Our product (PromptQL) gives a team a shared AI agent plus a human-readable wiki the agent maintains as a shared skills and shared knowledge repo.
This particular page had bloated into ~40 sections, had the same guidance repeated in 4 places, had dated status notes that were stale for a few months, and section titles that were not named according to the principle the agent is usually supposed to follow while maintaining wiki pages.
Think of it as a dark corner of our wiki that because of a combination of human & AI actions become a particular weakness for either agent or human to maintain.
We asked each model to restructure it with identical instructions.
WHY THIS IS A GOOD MODEL TEST
Maintenance punishes eloquence and rewards diligence. The only useful output is the fully rewritten page. The task also had hidden constraints (the page format supports flat sections only, no nesting) that are typical of knowledge work. We did this work on a duplicate copy of the page intentionally, while the original page was also kept around.
WHAT FABLE 5 DID (10/10 on our rubric)
• Read all ~40 sections before judging. The only model that did.
• Found the actual redundancy: one piece of guidance repeated 3x, another 4x. Collapsed each to a single home.
• Rewrote 40 incident-named sections into 9 skill-shaped ones ("how to do X" instead of "the June 12 issue").
• Added a routing map at the top so the reader can jump straight to the right section
• Deleted stale point-in-time facts and kept the durable ones.
• Caught the broken link. Every other model missed it.
• Noticed the page looked like a duplicate, flagged it — and then did the assigned job anyway.
• Delivered the complete rewritten page in one response. No outline. No "here's what I would do."
WHAT THE OTHER THREE DID
• GLM 5.2 wrote an articulate review memo about the page instead of restructuring it, but only after reading ~15 of the 40 sections. It shipped the diagnosis as the deliverable with a headline recommendation to delete the page ignoring the fact that we'd told it about the deliberate duplicate of the page. However, it produced the single sharpest observation of the batch about keyword placement.
• Opus 4.8 preserved every word verbatim from the original page and wrapped the mess in nested sub-headings that was invalid formatting. It also needed a second prompt before it would output the full page.
• Sonnet 5 reasoned beautifully about what the page WAS but applied every correct principle at whole-page granularity to justify discarding the content, and never shipped a complete rewrite.
THE INTUITION TO TAKE AWAY
Consider a task that is: read everything → judge fact by fact (rigorously) → rewrite completely — long, unglamorous work where the model has to respect the constraints of the system it's operating in and spend real output tokens finishing the job. For such tasks, Fable 5 is currently in a class of its own.
Add that to your model routing classifier and thank me later.
Never let one lab own your AI. They can ban you, throttle you, or shut you down. PromptQL Tag: your brain, your rules.
Anthropic banned our company in March, and gave us an Al bot for support. We were a $40k/mo customer.
Lesson learnt: Never let one lab own your Al.
Today we're launching PromptQL Tag: Claude Tag on AGI steroids.
Here are 8 things it does that Claude Tag will never do:
45
Shared context is the hardest unsolved problem in enterprise AI. Not anymore.
We call ours Operational Language Wiki, essentially a pattern for building memory to fill in the delta between the “dialect” of a team vs the “world language” that the LLM is trained on.
Use it to capture knowledge + skills.
Works for us. Works for F100 banks. 🔥
Feel like it’s been a good format to capture all context. Haven’t come across anything it can’t handle well.
Markdown + JSON + coding agents is all we need after all.
130
Suku Krishnaraj retweeted
PromptQL quickly learns your company's knowledge and workings, including specific terms and rules, to create a system that improves over time.
It learns and uses this layer to create custom AI that thinks like your business.
link.promptql.io/btz-has-x
Suku Krishnaraj retweeted
🚨SHOCKING: You can now build an AI that actually thinks like your business.
PromptQL lets you train reasoning agents that don’t hallucinate and automate everything.
Here’s how it works (with live examples):👇
Congratulations to the whole team 👍🏾
Congrats to our XC Bells for finishing 2nd in the @CIFState championships! Best finish in school history! #GoBells full results here: rtspt.com/events/cif/xc2024/…
2
5
168
Suku Krishnaraj retweeted
🚀 Happening tomorrow...our most exciting product launch to date!
We can't wait to show off what we've built. Plus catch an exclusive keynote from CEO @tanmaigo detailing how truly agentic data access will transform the #AI landscape.
📅 Save the date and elevate your AI game!
hasura.info/3C1sKUP
Suku Krishnaraj retweeted
We're showing off our new AI data access tool online during Dev Day on 31st Oct!
See how you can use it to build your own intelligent Customer Support Assistant that can
✅ Increases productivity
✅ Find answers and close tickets faster
✅ Provide insights into the team’s performance
✅ Eliminate repetitive questioning
✅ Unify customer data for a comprehensive view
✅ Leverage product usage data and telemetry for proactive support
Register ➡️ bit.ly/3YEu09k
Suku Krishnaraj retweeted
Varsity Bells rake home the large school win at the Clovis Twilight Invitational!
Amazing day for the #Bells
76






















