Jet AI Agents - Build business AI agents in minutes

Hey everyone 👋

We built Jet AI Agents because most teams still:
- jump between tools
- manually run workflows
- rely on engineering for simple internal tools

So we asked:
👉 what if business teams could build their own AI agents — on top of real data — and actually automate work?

With Jet, you can:

- build ai agents without code

- enrich AI with your data & 200+ integrations

- let them answer questions and take action in Slack, Telegram, WhatsApp and more

- instantly generate visual reports

- self-host in your own environment

- use open-source AI models

- bring your knowledge into AI — from files, drives, websites, and multiple formats like DOCX, PDF, JSON, MP3.

We’ve also created a few templates to help you get started (the real magic is in customizing them ✨):
📊 Data Analysis Agent 👉 https://www.jetadmin.io/agent-templates/bigquery-data-analyst

🗓️ Meeting Preparation Agent 👉 https://www.jetadmin.io/agent-templates/meeting-prep-agent

🎧 Support Agent 👉 https://www.jetadmin.io/agent-templates/customer-support-agent

📝 Meeting Analysts Agent 👉 https://www.jetadmin.io/agent-templates/meeting-notes-agent

Would love your feedback:
👉 What’s the first workflow you’d automate?

The 'relying on engineering for simple internal tools' line is the story of my life. My backlog is 6 months deep with 'just one more internal dashboard' requests. If Jet lets my business ops team build their own BigQuery analyst without touching a line of code, you’ve just saved me 20 hours a week. Does it handle write-back permissions safely?
@anton_svetlov

interesting take on bundling workflows + agents under one platform. one thing i'd be curious about — when you have agents calling into the same data layer simultaneously, how do you handle scope conflicts? in my own setup (claude code on a nuxt 3 + go-zero stack) i ended up writing per-folder AGENTS.md files just so concurrent sub-agents wouldn't step on each other.

does jet handle that internally or is it more about the orchestration layer? curious about the design choice.

"AI agents your team will trust - because they built them themselves" hits different. we've had mixed results with off-the-shelf AI tools in healthcare workflows, but letting domain experts build their own makes sense. what's the learning curve like for non-technical users?

the Slack integration is smart - we've been looking for something that lets our team build agents without pulling devs away from core product work. curious how the 200+ tool connections handle auth and permissions? does each team member need to connect their own accounts or can you set up shared service accounts?

The knowledge ingestion piece is interesting. When you bring in docs from drives and websites, how does the system handle freshness? Is it a one-time import or does it stay in sync with the source? That feels like the part that makes or breaks trust in agent answers over time.

congrats on the launch. the "answer questions AND take action" piece is the part i'd want to grill when an agent has write access through 200+ integrations, how does jet handle action scoping ? does the business user who builds the agent have to whitelist specific actions per integration, or is it all-or-nothing per connection ? a poorly scoped agent firing into the wrong slack channel or messaging the wrong whatsapp contact is the failure mode that scares me as a builder.

As a founder, I’m tired of asking my engineers to build simple internal automations. If I can connect our BigQuery data to a Slack agent without a ticket, it changes our speed entirely. Grabbing the PH exclusive now to test this out. @f1nal @Jet Admin

Love this. Building agents that actually live inside Slack and WhatsApp instead of yet another dashboard tab is the right call. Curious which integrations teams reach for first when getting started?

genuinely curious where the 200 integrations thing breaks down in practice. Every tool says that and then you find out your specific tool is the one that doesn't work. what are the gaps rn?