Overview
A company dashboard + your company profile — the state of the whole org at a glance.
Company Admin
Run the whole company from one console.
People, AI agents, communications, workspace, finance and governance — administered from a single screen, with permissioned access for every role. The HRIS, the identity provider, the device console and the stack of IT tickets, collapsed into one place that’s easier than any company you’ve run.
hq.plantel.ai/#company-admin
Live product Forty sections. One console. Everything your company runs on.
Reached in one click from your own menu, Company Admin is a tenant-scoped console gated to owners and admins, with auto-save on every field, an in-app back stack, and a per-company operations log that exports to CSV. It auto-locks after ten minutes idle.
People & Access
Team & people, departments, roles & permissions, security, SSO & auth, vault, OTP, HQ cards.
Agents & AI
Your AI workforce, agent infrastructure, Plantel AI settings, and API & keys.
Communications
Phone, chat, Huddle, calendar, notifications and shared inboxes — administered in one place.
Workspace
CRM, Drive & docs, Web Intelligence, Signature, business cards, tasks, notes, Brain, the dashboard layout, the 3D office, branding, integrations and feature flags.
Finance
Wallet & billing — deposits, spend, per-seat costs and limits.
Governance
Storage, data & retention, compliance, audit & logs, Workforce Intelligence, and maintenance.
Every teammate — one profile, total control.
Team & people
One surface for every human seat — and a 12-tab drill-in on anyone.
See every person as a seat with role, department, status and onboarding progress, with the live human-seat cost in the header. Click anyone for a deep profile: editable identity, access & seat (suspend, remove, transfer ownership), security (2FA, active sessions, last sign-in, force-sign-out everywhere, the tenant password policy), per-user notifications and settings, the full permission matrix, and an activity & audit feed — twelve tabs where you’d otherwise open an HRIS, an IdP and a settings panel.
- Inline role changes + deliberate, owner-only ownership transfer
- Per-person 2FA, session list, force sign-out, password reset
- 12-tab drill-in: sessions, activity, files, comms, permissions, productivity, billing, audit
- Onboarding progress tracked per person
hq.plantel.ai
Live product Roles & permissions
120+ permissions across 18 areas — granular access without an IdP.
Five built-in roles — Owner, Admin, Manager, Employee, View Only — plus custom roles you define. Each role sets smart defaults across a permission matrix you can override per person, and it’s enforced at runtime: a Manager can only ever grant access they hold themselves. No separate identity platform, no access-review spreadsheet.
- 18 permission categories, 120+ individual flags
- Built-in roles + custom roles per company
- Per-person overrides on top of the role default
- Enforced live — nobody grants what they don’t hold
hq.plantel.ai
Live product GeneralUsersAgentsTasksBillingEmailPhoneCalendarHuddleChatVaultDriveDepartmentsTeamPlantel AIPlatform adminReportsCRM builder
Hire someone in one guided pass.
Add User is a 10-step flow that fully provisions a seat and creates the account atomically — identity, role, every permission, an AI budget, a phone extension, an email mailbox, vault credentials, and a welcome email with a secure set-password link. No five-tool checklist, no IT ticket.
- 1Identity
- 2Personal details
- 3Role & seat
- 4Permissions
- 5AI budget
- 6Phone
- 7Email
- 8Monitoring
- 9Credentials
- 10Review & create
Departments, reporting lines and a drag-to-reorganize org chart keep the structure honest — change a department or a manager and the sidebar, Live Office, org chart and agent wizard all update at once.
Oversee your whole workforce — humans and AI agents — in one console.
A single pane for live status, live screen, productivity and risk across the people and the AI agents working side by side. Built for a mixed workforce — and built governed: monitoring is OFF by default, enabled per person by an admin, behind a disclosure attestation and organization-level disclosure verification, jurisdiction-aware for all-party-consent states, retention-bound, and every access is itself written to a tamper-evident audit. Oversight you disclose — never covert surveillance.
One roster, humans + agents
Every monitored person and every AI agent in one filterable roster with a real live-status dot. People with monitoring off never appear — off by default, always.
Live status, six work states
Working · Idle · In a Huddle · On a Call · In a Meeting · Away — fused from real presence, calls, meetings and calendar, with a fleet-pulse ring and a status river.
Live view + agent controls
Watch any agent in real time with a control bar — Watch, Pause, Take Over, Terminate — each gated to its own server-side authority and fail-closed. Every agent frame carries a disclosure badge that can’t be turned off.
Captured data + session replay
Per-person capture coverage, a session-replay scrubber, and a fused timeline. A DLP red-thread stitches clipboard → file → upload into one connected exfiltration trail.
Productivity & focus
A transparent focus score (the math is shown), productive/neutral/unproductive bands, active hours, context-switches and a collaboration graph — for humans and agents.
Board-ready reports
Five report families — productivity & focus, throughput & workload, risk & anomaly, agent performance, training & SOP coverage — scoped by team and saved to Drive.
Insider-Risk Radar — a score you can explain
A transparent, weighted risk score per subject built from eight named behavior signals, each with severity points and a plain-language “why”. No black box — a risk fingerprint you can actually defend.
Off-hours activityIdle anomalyAuthority denialsRender instabilityTask anomalyConcurrency conflictDLP exfiltrationAudit tamper
Toggle each signal into Alert, Flag, Record or Block in the behavior-rules engine; rules persist per company and re-score live, with an instant in-app alert when anyone crosses the critical line.
Eight capture categories — exactly what you need
Per-person capture, not all-or-nothing. Sensitive categories are clearly flagged, keystroke capture excludes password fields, and a Light intensity tier forbids the heavy ones entirely.
- Website & app usage
- Screen recording
- Keystroke loggingsensitive
- Mouse & interaction
- File operations
- Communicationssensitive
- Network & DLP
- Productivity metrics
Retention is set per policy (60 days by default, auto-purged after). Full-desktop capture and a live agent-VM video stream arrive with Plantel OS — today’s capture is browser-scoped.
Watch how a role is really done — then train an agent to do it.
Turn a job into a deployable AI agent — with consent, and a human in the loop.
With monitoring on for a role and an explicit secondary-use disclosure on the record, Plantel builds a learnable “trained role” from what that person actually does — their tasks, the access they need, who they escalate to, the rules and the edge cases. A real coverage meter climbs as the work is observed (with nothing captured, it honestly reads “0% observed, scaffold ready” — never a faked number). At 100% learned, you’re notified, and the Add-Agent wizard can spin up the agent “from a trained role”, pre-loaded with the role’s skills and SOPs. The role files are shared, not copied — improve the procedure once and every agent built from it learns it instantly.
- A real coverage-based learning meter — never random, never seeded
- Human-in-the-loop validation: the agent runs in parallel with the person before any handoff
- One click from a 100%-learned role to a fully-built agent (face, voice, persona, comms)
- Shared role files: update once, every cloned agent updates too
hq.plantel.ai
Live product - 1 ObserveWith consent, monitor the role across the chosen categories.
- 2 SegmentSplit the captured work into distinct tasks.
- 3 AbstractCluster tasks into invariant workflows + decision points.
- 4 Generate skillSave parameterized, reusable skills to Drive.
- 5 TrainProvision an agent with the skills, in training mode.
- 6 ValidateRun the agent in parallel with the person; compare outputs.
- 7 HandoffOnly after it passes does the role transition.
Enabling role-learning to replace a role requires a separate, on-the-record disclosure on top of monitoring consent. It’s positioned as transparent, consented role-learning with shared improvement — and there’s always a human-in-the-loop validation gate before anything hands off.
Capture your likeness and voice — and send your double to the meeting.
In a few steps you add 1–5 photos of your face (or use your profile photo) and a 30-second voice sample, and Plantel builds your personal clone identity for use across Huddle, Phone and the Video Studio. Likeness and voice capture work today; photoreal video and voice synthesis activate on the Plantel clone service. Trained inside Plantel — nothing sent out.
hq.plantel.ai
Live product Administer humans and AI from one console.
People, permissions, onboarding, governed oversight and role-learning — in one place.