Jev is democratizing high-performance zero-shot classification.
Just like ChatGPT pushed people to solve problems with generative text, Jev is pushing developers to re-model their problem space as classification tasks to fit its interface.
The winning systems will be Jev + LLMs
Introducing ax - the Agentic Experience suite.
Ora is already the best way for humans to understand agents.
And now it's also built directly for agents.
One tool for seeing exactly how an agent experiences any product, built to fit into any agentic loop.
Simply `npx ax` .
We observed over a thousand agent runs with @oradotai to understand what makes webpages readable to agents.
We found that the way servers respond to requests matters more than having 𝚕𝚕𝚖𝚜.𝚝𝚡𝚝 ↓
https://t.co/Q8l0tHdOIQ
Being in the answer isn’t the same as winning it. Agents keep exploring after the first result, and every step after that tests whether they can read you and complete the task. We see it in the data.
AEO will get you surfaced. It doesn't mean you get chosen.
Being in the answer is Discovery - necessary, and completely insufficient. Nobody stops at the first answer. The human asks a follow-up: comparing pricing, features and alternatives.
Every follow-up tests whether agents can actually read you. When they can't, the answers come from third-party articles and old reviews. Someone else's framing of your product, your pricing and more - in the middle of your deal.
Then the last test: can the agent act? Sign up, book, buy, complete the task. If not, the deal closes somewhere else - usually with whoever passed all three.
Discovery, Access, Use. AEO covers the first.
Surfaced is step one. Being accessible to agents is what gets you chosen.
Introducing https://t.co/8MeR16MRQn, a tool to measure how well agents can read your site. Backed by @oradotai's research, you can run:
▪︎ Audits with 100+ checks
▪︎ Visualizations of agents using your site
▪︎ One-click prompts to fix problems
▪︎ A CLI for agents
@hostmyblogco Our research at https://t.co/V2QmNFFFxS shows that agents almost never prob the llms.txt by themselves, BUT if you create the right linkage to it from other pages (e.g homepage) or resources (e.g sitemap) they do use it + it improves their success rate
@rrmdp@jobboardsrch Make sure your site as a proper linkage to your llms.txt on various pages and it will increase the chances agents will find it and use it
You can use https://t.co/KwQuj3bVML to see if agents can actually find your llms.txt and other resources
@zenorocha Solid start, but we’re just scratching the surface. Check out the full audit at https://t.co/1V6Fu4LY6I, and spin up live agents to see them in action at https://t.co/pGiNWcBwem
@0x_Kovan Nice! But pay-per-request only works if the agent gets that far. it has to discover you and access your product first. most products fail silently before the payment step.
Ora scan and rank exactly that https://t.co/1V6Fu4LY6I
Agents aren’t coming. They’re already at the door.
Ora lets them in.
@telnyx, the leading voice platform for real-time AI agents, is already acting on it with Ora: 30%+ of their agent signups now convert to paid.
Read the story: https://t.co/Q6GAmxoNUI