PRICING

Pay for staff, not seats.

Your whole team gets access. You pay for the AI employees you keep on shift.

STARTER
$290/ employee / mo
For small teams hiring their first one or two.
  • Up to 3 AI employees
  • Unlimited humans
  • 10 live workflows
  • Company chat, 5 connected systems
  • Email support
Start free
MOST HIRED
COMPANY
$240/ employee / mo
For companies staffing whole departments.
  • Unlimited AI employees
  • Unlimited workflows
  • Unlimited connected systems
  • Custom roles & job descriptions
  • SSO, SCIM and audit log export
  • Shared Slack channel
Start free
ENTERPRISE
Talk to us
For regulated industries and global rollouts.
  • Everything in Company
  • Private deployment or VPC
  • Data residency controls
  • Custom retention & DPA
  • Named implementation lead
  • 99.9% uptime SLA
Book a call
NO CREDIT CARD · 14-DAY TRIAL · CANCEL ANY TIME

Questions we get asked

What counts as an AI employee?

One named role on shift — like Finance or Support. It can run any number of workflows and conversations; you pay per role, not per task or token.

Do our own team members cost anything?

No. Humans are unlimited on every plan. You only pay for AI employees you keep on staff, and you can retire one at any time.

Can we use Dutta Talk on WhatsApp or Telegram?

Yes. Link a WhatsApp or Telegram account to your Dutta account with a one-time code generated in the web app, and the assistant answers there — including voice notes, which are transcribed so it can respond. It only sees messages sent to the Dutta bot, anything it wants to send still stops at your approval, and you can unlink at any time from Profile → Chat Channels.

What can an AI employee actually touch?

Only what you grant it. Credentials are scoped per system — read, write, or send — and nothing beyond that mandate is reachable. Permission is granted per system, not per employee. You can revoke access at any time, effective immediately, and you name the decisions that must wait for a person.

Where does our data go?

Into your Dutta workspace only. It is never used to train shared models. Enterprise plans support private deployment and data residency controls.

What if Dutta gets something wrong?

Every action is logged with its sources, and you set what must be approved before it happens. Replay any run to see the full reasoning trail.