@ZooWorkAI

The Open Enterprise AI Stack. Put frontier intelligence in every employee's hands and every business workflow.

zoowork.ai
Joined June 2026
  • Tweets85
  • Following75
  • Followers46
29Media
people are now arguing about whether an agent should be allowed to approve the refund. a year ago the fight was over letting it answer the ticket at all, and nobody brings that one up anymore
Sebastian Siemiatkowski
@klarnasebOct 19
When you have been in the game for 20 years... stopping agents to do work for the customer, is the same I have seen for such a long time. But they will not be stopped.
1
33
Building an agent is easy now. Handing it to a client is the hard part. @ZooWorkAI lets you build one by chat (or with Codex or Claude and their SDK) and deliver it as a link, an API, or a Slack or WhatsApp bot. Your expertise, deployed in minutes: zoowork.ai/?utm_source=x&utm…
🤖 Made with AI
3
2
3
296
3.8x on cost for five points of intelligence, and on everyday work those five points barely show. we route the boring stuff to the cheaper models, and that routing is where most of our model savings come from
Tomasz Tunguz
@ttunguzOct 14
Five frontier models now sit within five points of each other in terms of intelligence, but their cost per unit differs by 3.8x. Why could this be?

Price War in the Chip Aisle

Article · @ttunguz
1
70
10 years sounds about right from where we sit. getting an agent to run inside someone's actual POS system for months is where most of them stall out.
Amjad Masad
@amasadOct 13
You can pause frontier research today, and it would take AI another 10 years to diffuse into the economy. We're working with solutions partners to accelerate that.
53
In retail, your AI agent can look up a customer’s order, but should it also be able to issue a refund? Here’s how we decide what an agent can see and do:
1
1
56
more replies
Data access and permissions belong in the solution design, and the client’s own systems should enforce them. Human approval is one gate. It does not replace backend authorization.
1
5
Where does your team draw the line between what an AI agent can do on its own and what needs approval?
7
what's even left for the gateway once routing moves into the harness? just the cost and latency stuff?
Harrison Chase
@hwchase17Oct 13
router lives in the harness, not a gateway, bc the harness knows the task blog: langchain.com/blog/how-to-bu… xitter.cf/neil_xbt/status/210997…

How to Build a Model Router in the Harness

How we built a model router into Open SWE's harness that cut median cost per coding task by 64% with no measurable drop in quality, and how to build your own.

langchain.com
1
31
A stylist that knows your closet. A bot that reads every filing before the market opens. How much can you ship between breakfast and 6 PM? Nov 7 · NYC · with @AITinkerers · $100 in credits per builder · 100 spots Apply 👇 nyc.aitinkerers.org/p/fashio…

Fashion and Finance Hackathon with ZooWork [AI Tinkerers - New York City]

Fashion and Finance Hackathon with ZooWork

nyc.aitinkerers.org
1
1
65
Before we build an AI agent, we agree with the client what “done” looks like. Here’s why, and what we agree on before any code gets written:
1
1
1
39
more replies
A demo proves the happy path. Acceptance criteria determine whether the system should scale, keep piloting or stop.
1
10
What’s the first thing you’d want agreed before an AI agent touches your business systems?
7
Load more