Backdrop - AI Coworkers that run your projects and operations
by•
AI made execution faster. The bottleneck has shifted to deciding what to build while the knowledge behind those decisions is scattered across people, tools, and AI chats.
Backdrop provides AI coworkers for projects and operations that understand your company, work with your team,and build shared company context. Across Slack, GitHub, Linear, Notion, Asana, Google Workspace and more, they synthesize customer feedback, create plans and specs, manage tickets, draft documents, and keep work moving.
Replies
Everyone here is reacting from the "team with scattered Slack/Notion knowledge" angle, which makes sense since that's who you built for, but I'm curious where the line is for a one-person team. If it's just me, GitHub, and a task tracker, is there still enough fragmented context for Alex to be worth wiring up, or is the value fundamentally about reconciling multiple people's half-knowledge? Not knocking it, genuinely trying to figure out if this category applies below a certain team size.
Backdrop
@cannetjam Really fair question. The reconciling everyone's half knowledge part won't do much for you solo, that's honestly a team thing. But a decent chunk of the value doesn't really care about team size. You still don't have to build the agent or set up the integrations (in your case Alex just shows up already working across GitHub and your tracker). And your context is still scattered, just across time and tools instead of people. Why you scoped something the way you did, what you decided and why, what's sitting half done where, etc. Having a memory of all that is valuable. And honestly the delegation part might matter more solo, stuff keeps moving while you're heads down and there's no one else to hand things to.
So yeah it applies, the team stuff is just another layer on top when you grow
that's a great breakdown, especially the point about reading data being maybe 20% of the work. the Google Workspace / Figma example makes sense too, since a PM and an engineer looking at the same file are basically doing 2 different jobs even though the underlying API call is identical. are you building the role-specific behavior as hardcoded rules per app, or is there a more general framework so a new integration doesn't mean starting from scratch each time?
Backdrop
@omri_ben_shoham1 Thanks! Honestly a bit of both. There's a general framework underneath, the integration layer deals with auth and tool access, and the role specific stuff lives in skills (eg how a PM works). So new integrations aren't starting from scratch, a lot of it carries over. But some tools are just weird enough that parts end up custom. There's always some last stretch per app you end up custom building.
The shared company context across Slack, GitHub, Linear and Notion is the piece I would test first — for a community or product team the risk is not capability, it is a coworker surfacing something from a private channel into the wrong doc. On setup, can I scope what each coworker reads — connect Slack but exclude specific private channels, or restrict a coworker to one Linear team — or is it all-or-nothing per integration? And when it synthesizes customer feedback into a spec, does it cite the source threads so I can trace a claim back before acting on it?
Backdrop
@hazy0 Great question! On scoping, it’s not all or nothing. In Slack, agents only see the channels you invite them into. In Linear, you can restrict a coworker to specific teams. Private channels and other teams stay out of reach unless you bring them in. On synthesis and traceability, when an agent turns feedback into a plan or spec, the conversations, links, and comments also show up in one place on the dashboard. So you can see what happened and trace claims before you act on them.
That per-channel invite model is the right primitive, thanks. The follow-up that matters for me: once it synthesizes across tools into that shared dashboard, does the output inherit source-level permissions? Concretely — if I invite a coworker into a sensitive Slack channel and it rolls that into a spec, can a teammate who is not in that channel read the synthesized claim on the dashboard, or does the traceability view stay gated to people who had access to the underlying source?
Backdrop
@hazy0 You get to control who has access to our dashboard and any output Alex creates.
Backdrop
@hazy0 Hi Hazy, you can absolutely scope the coworker, so it only has access to specific things (for eg, on slack it only has access to channels you invite it to). For every task it does, you can see a full history of how it came up with the answer with the sources but you can also instruct it to write the sources in the doc and it will. The agent has memory and remembers your preferences for how it should write responses, docs, etc
Interesting direction.
As AI coworkers become more common, what do you think users will ultimately buy—the individual agents, or confidence that the whole operational workflow actually gets completed?
Backdrop
@aryan787544 I think people will still buy agents, because the agent is the unit through which confidence becomes tangible. For eg: Alex owns PM and ops, so you know what she is responsible for, can inspect what she did, and can evaluate whether she did it well.
A blanket promise that “the whole workflow gets completed” is much harder to trust or verify. So the outcome is ultimately what matters, but it needs to be delivered in clear pieces with visible ownership and accountability.
People buy the agent, but what they’re really buying is the workflow getting completed in a way they can verify.