What if the schedule just happened? For example: A recruiter sends a message, and each interview with five candidates, three hiring managers, and two time zones gets booked, confirmed, and updated automatically. No links, no back-and-forth, no spending hours with 20 emails. Everyone gets the right invitation at the right time, no matter what channel they actually use. That’s what we created Vela to do.
You loop Vela into your email, SMS, WhatsApp, Slack, phone or integrate into an ATS, etc. and it takes over: reads references, checks calendars, proposes times, follows up when people ghost, and rebooks when things change.
One of our first customers is a staffing firm that searched for a scheduling solution for about eight years. Their coordinators manage hundreds of candidate-client interviews where each party requires separate email threads, separate Zoom accounts to avoid double-booking links, and calendar invites connecting parties who never directly communicate. A client reschedules an interview and it gets added to four other interviews. A candidate replies on SMS to a thread started on email. Vela solved this in just 10 minutes of onboarding.
The hardest part has been the data problem. Scheduling behavior varies greatly across different populations. C-suite people respond to emails within hours and expect formal 3-option offers. Truck drivers applying for logistics roles reply to SMS from shared device at odd hours with “y tm wrks”. The failure mode isn’t parsing – it’s applying the wrong interaction pattern to the wrong segment and watching the conversation collapse. We’re building behavioral datasets from thousands of real interactions: response latency by role, channel preference by demographic, follow-up time curves, how many options to offer you before decision paralysis occurs. This data does not exist anywhere.
The main agent challenge is the status of all channels. When someone responds to an SMS on a thread started by an email, Vela needs to unify identities, merge context, and continue without losing information. Phone numbers don’t map clearly to email, people use aliases over text, shared devices mean the responder may not be the one you contacted. Temporal NLU has its own problem – “next Friday” means different things on Monday vs Thursday. We extract structured constraints from natural language and solve against the calendar position. When ambiguity cannot be resolved, Vela asks – but deciding when to ask versus guess depends on the risk of being wrong.
We live with paying enterprise customers and every customer still has issues that surprise us. Case studies on our site (https://tryvela.ai/case-studies/). You can watch the demo here: https://www.youtube.com/watch?v=MzUOjSG5Uvw.
We would love feedback from anyone who has worked on multi-agent coordination, conversational AI across channels, or constraint satisfaction in messy real-world domains. Looking forward to your comments!
<a href