How it works · Flow F16

Discover, trial, rent, supervise, review — or return

Find the right agent or squad, trial it on synthetic data, sign a rental, onboard it into your policies, supervise it in shadow and draft modes, and review it monthly with an Investment Statement.

What the Rent a Digital Employee flow does

Renting a Digital Employee is One3's headline business model. The rented agent works inside One3's apps, for your team, under your supervision.

You describe your goal, industry, language, systems and budget. One3 recommends an agent or a squad. You try it in a sandbox on synthetic or redacted data, at no rent. You see evidence: a quality score, a cost estimate range and the risk tier. If you want it, you sign a rental contract.

Onboarding loads approved knowledge, permissions, brand voice, playbooks and escalation contacts, and you issue a mandate that limits what it may do and spend. It starts in shadow mode (A1), then drafts for approval (A2), then works in supervised production within budget and policy. Each month you get a scorecard and an Agent Investment Statement. Then you decide: keep it, improve it through a versioned release, replace it while keeping your context, or return it — with your data exported or deleted and a deletion certificate.

The first-party Digital Employee Store with 20 templates Listed and 6 Verified is planned for W2. Training Gym and the Replacement Engine as full apps are W3.

Every step on this page is an Action: a typed verb with a risk tier (R0 read-only to R4 irreversible or financial), a permission check and a reversibility class. Agents and humans call the same Actions, so the same approval rules apply whether a person clicks a button or a Digital Employee acts.

Rent a Digital Employee, step by step

Thirteen steps. Your data, memory and playbooks never leave your tenant.

1
1. Discover
**AI Workforce · Agent Rental Marketplace.** State the goal, industry, language, systems and budget.
2
2. Recommended agent or squad
**AI Workforce · Squad Builder.** A shortlist with a cost range and stated assumptions — not a promise.
3
3. Sandbox trial
**AI Workforce · Synthetic Customer Simulator.** Synthetic or redacted data. No rent and no pass-through costs, because no real people are contacted.
4
4. Evidence
**Performance Arena.** Quality score, cost estimate range and risk tier.
5
5. Sign the rental contract
**Human gate.** Rent, credits, caps, SLA, outcome terms and data terms.
6
6. Onboard and issue the mandate
**Platform · Identity & Access · Admin & Policy.** Approved knowledge, permissions, brand voice, playbooks, escalation contacts and a client mandate with purpose, Actions, channels, spend limit and expiry.
7
7. Shadow (A1), then draft (A2)
**Agents Studio.** It proposes, then drafts for one-click approval, until per-skill thresholds are met.
8
8. Supervised production
**AI Workforce · Investment & ROI Ledger.** Works within budget and policy with a live spend meter.
9
9. Monthly review
**Human gate.** Scorecard plus the Agent Investment Statement.
10
10. Improve
**Training Gym.** A versioned release, tested and canaried.
11
11. Replace
**Agent Replacement Engine.** Swap version or template; your context bundle is preserved and both run in shadow before cutover.
12
12. Return
**Offboarding.** Mandate revoked, identity revoked, data exported, memory and adapters deleted, deletion certificate issued.
13
13. Final statement
**Investment & ROI Ledger.** A final statement closes the rental.

Guardrails and KPIs

Item	Detail
Guardrails	Tenant data, memory and playbooks never leave the tenant instance; promotion needs an Eval Coverage Score; avatar or voice use only through the Rights Registry
KPIs	Trial to rental conversion, time to supervised production, % agents with positive return after the minimum sample, replacement rate, rental retention