Zip Engineering @ZipEngineering

Still distraught over having missed Zip Forward in NY? *TUNE IN LIVE RIGHT NOW* to catch Zip Forward Rewind—now streaming top sessions all morning! LIVESTREAM: events.zip.com/zip-forward-re…

hacking the @HackTheNorth badge but it's the linkedin zip game themed with @theziphq

they don't know admin is at @HackTheNorth .... actually they DO and SHOULD already know. come visit us anytime today in sponsor bay and maybe do a zip.com API prize 👀 👀 👀 👀

keyboard_arrow_left Previous keyboard_arrow_right Next

SaaS-poccalypse was never real

Zip @theziphq

2 months ago

Frontier AI models are becoming more powerful every day. Yet even among senior leaders, 57% told us they use AI tools their company never sanctioned. How much of that work would survive an audit? Enterprise AI *must* be built on trust and governance. That's what we're

0 0 1 358 1

Touchless invoice processing has long been the dream in AP, but only works with coding you can trust. Eishan Vijay and Xande Macedo share how Zip's AP Superagent can read your team's playbook, turn it into instructions you approve, and code invoices exactly the way your AP team would. Every change is logged and nothing ships without human approval. That's governed AI. zip.com/capabilities/s…

Aish, an engineer working on building AI @ Zip, has been tinkering a lot with LLM behaviors. She published her work titled "Tone Matters: Personality Trait Shifts in Large Language Models" on IEEE Xplore. Check it out here - ieeexplore.ieee.org/document/11620…

Shipping AI products is easy. Shipping AI products that customers trust and works in complex enterprise workflows is tough. Check out how we ship more than just Vibes and evaluate AI agents across different products and technologies zip.com/engineering-bl…

Everyone is using AI. Not everyone is getting the same return. This year we asked more than 1,000 leaders across procurement, finance, IT, and operations how AI is actually changing their work. Among the teams seeing the most return, the same patterns show up again and again. We call them the Builders. Our new report shows what they do differently. Get the full State of AI in Spend 2026 report (plus a bonus AI skill): zip.com/stateofai

Alicia and Xinyue built a super cool configless intake flow — paste a policy in plain English, and AI turns it into fields, routing logic, and a guided request flow. Really interesting glimpse of a future where compliance and governance steps can all just be plain english. ⚡️

We have started hosting super casual hack nights - a smaller one day lightweight hackathon for zipsters to vibe build and tokenmax something that helps internal tooling across all functions or customers. the difference? we actually ship what we prototype shoutout fish in the box for the fire sushi 🧵check out some cool stuff we shipped last week!

keyboard_arrow_left Previous keyboard_arrow_right Next

Enqriue built an approval-routing sentinel — because the trickiest workflow bugs don’t crash - they quietly route to the wrong person 😔 AI generates complicated edge cases -> deterministic tests check the real approval chain, and seeded bugs prove the tests actually work. Uses AI for finding the gaps and a deterministic test suite to validate them. 🚀

the future of enterprise procurement is here 👀

Zip @theziphq

4 months ago

Superagents are the future, and our Principal Product Manager Natasha Berk has been waiting for this one. AI teammates that handle real finance and procurement work end-to-end, with every action logged, permissioned, and policy-enforced so the work gets done and the audit trail

0 1 4 409 0

I made a small tool called Gloss because my agent code review loop was annoying me. Gloss opens your local diff in a browser, lets you leave comments, and returns them to the agent so it can fix + validate. getgloss.dev

Not to brag, but I think we've got the slickest booth at #DPW Stop by today for your Superagent trading cards. Yes, really.

Some tools require guardrails stronger than *just vibes*

Check their work out here improving capability of long running sessions : github.com/Vincen-Le/auto…

This weekend our engineers went to an hackathon hosted by @modal @AntlerGlobal @OpenAIDevs @raindrop_ai focused on improving Autoresearch. Vibhav, Galen, and Vincent won the Raindrop track by developing an autonomous LLM as a judge to live eval agent traces and steer subagents

keyboard_arrow_left Previous keyboard_arrow_right Next

we are now THE zip

Zip @theziphq

5 months ago

hello zip.com

0 0 4 330 0

Your current coding tool only knows what's on the invoice. Pretty unimpressive if you ask us. Zip AI knows what was purchased, from whom, and why. Because, unlike some people online, we believe *context matters* ✨ Zip's AI Invoice Coding Agent uses your full purchasing context to code accurately, combined with historical coding inferences, to truly automate the invoice coding process. Read more: ziphq.com/blog/ai-invoic…