A specialist team that never clocks off
Most rental software waits to be asked. The overdue off-hire is in the system — if someone thinks to run the report. The quote going cold, the machine sitting idle, the certificate about to lapse: all discoverable, rarely discovered in time.
Renttix puts a team of seven specialist AI agents on the watch: Hire Desk, Ops Manager, Revenue Recovery, Customer Intelligence, Compliance, Fleet Intelligence, and Business Analyst. Each is opt-in and works a narrow beat with thresholds you tune — the Hire Desk chases quotes before they go stale, Revenue Recovery hunts unbilled hire days and aged debt, Customer Intelligence flags accounts whose ordering pattern has broken, Compliance watches inspections and expiring certificates, Fleet Intelligence weighs the economics of every asset.
What they produce is not commentary. It is concrete, ready-to-approve work: a catch-up invoice for hire days that were never billed, a dunning nudge for the account slipping past terms, a retention follow-up for the customer going quiet, an inspection booking for the machine running out of certificate.
Underneath them run thirteen deterministic morning digests — overdue off-hires, unpaid invoices, today's dispatch, idle assets, delivery conflicts, stale quotes, revenue leakage, at-risk customers, compliance obligations, and fleet economics among them. There is no language model in that pipeline, so there is nothing to hallucinate: every line is computed from your live data and quantified in your own currency, whichever currency that is. Digests arrive where your team already looks — the dashboard, email, Slack, WhatsApp, or any Slack-compatible webhook you point them at.
Proposals you approve, not actions you discover
An AI that quietly moves money is a liability, so Renttix draws a hard line: the proactive layer never moves money on its own. It raises proposals — a concrete action, validated inputs, and the rationale behind it — triggered by platform events, digest findings, or agent sweeps.
Proposals land in the AI Command Centre, ranked into three lanes: actions awaiting approval first, warnings next, information last. Your team approves or rejects with one click. An approved proposal executes under the approver's own credentials and permissions, fully audited — so a proposal can never do anything its approver could not have done themselves.
Guardrails sit behind the button. Duplicate-proposal protection stops the same catch-up invoice being raised twice. A money-rate circuit breaker pauses any agent that starts proposing money moves past a safe rate — automatically, until a human steps in. And for one genuinely safe piece of non-money housekeeping — logging a follow-up task — you can opt an agent into direct execution, fully audited, while anything touching money always waits for a person.
The same discipline covers AI-generated reports. Ask the assistant for a CSV — unpaid invoices, overdue off-hires, today's dispatch, or a straightforward export of customers, products, orders, or depots — and it arrives as a download link. The link re-checks the downloader's role on every request: forwarded to a colleague without the permission, it politely refuses.
Governance you can verify, billing you can predict
AI in a business system earns trust through controls, not promises. Renttix permissions are deny-by-default: the model only ever sees the tools the signed-in user's role allows, so a forbidden action is not refused — it simply does not exist. Every conversation is scoped to a single tenant. Personal data can be masked into opaque placeholders before text reaches the model and restored on the way back, card numbers are masked unconditionally and never passed through, and content from documents and messages is fenced as data, never instructions — a standing defence against prompt injection.
Everything the AI does lands in a hash-chained audit trail, verified with one click — if a record has been altered, verification says so. Users can rate any answer, thumbs-up or down with a comment, and every rating lands in that same tamper-evident trail. Releases are gated separately: a behavioural evaluation suite must pass before a build ships.
Control includes the off switch — several of them. Every agent's daily sweep is opt-in and can be switched off in your own settings at any time, the assistant can be disabled for an entire tenant, and a platform-wide freeze can halt all AI writes at once. Even the assistant's credential is out in the open: it appears in your Developer settings as 'Renttix Assistant (auto-provisioned)', visible and revocable like any other API key — and every re-issue goes through the audited provisioning flow.
AI usage is billed in credits you can see. Every AI turn is metered, the monthly allowance scales with the seats you already pay for, and top-up packs never expire — they are the rollover. Owners and admins can cap any individual user's monthly spend, the usage dashboard shows where credits go, and a monthly report email summarises usage against allowance with the top features and heaviest users. When credits run out, new AI work pauses until you top up.
Try this feature in the live demo
Explore the full Renttix platform. No signup required.
AI operations FAQ
Seven specialists — Hire Desk, Ops Manager, Revenue Recovery, Customer Intelligence, Compliance, Fleet Intelligence, and Business Analyst — alongside the general-purpose assistant. Each agent's daily sweep is opt-in per tenant and tuned with your own thresholds, and every agent works only with the tools your roles allow.
Anything that moves money always waits for human approval. Proposals execute under the approver's own credentials, with duplicate protection and a money-rate circuit breaker behind them. The one exception you can opt into is narrow and money-free: agents may log follow-up tasks directly, and every action is audited either way.
No. All thirteen digest kinds are deterministic — computed from your live data with no language model in the loop, so there is nothing to hallucinate. Figures are reported in your own currency and delivered to the dashboard, email, Slack, WhatsApp, or any Slack-compatible webhook you choose.
Per-turn credit metering against a monthly allowance that scales with your seats. Top-up packs never expire, per-user limits cap individual spend, and a usage dashboard plus a monthly report email show exactly where credits went. When credits run out, new AI work pauses until you top up.
Ready to modernize your rental operations?
Payments + deposits enabled • Quick setup • No credit card trial