The AI layer
Don’t just query your operations. Run them.
Moonschat is the autonomous AI agent embedded directly inside Moonfoos ERP.
It doesn’t just answer questions about your sites, cleaners, and contracts—it plans shifts, notifies workers, reconciles timesheets against invoices, drafts client dispute responses, and enforces site SOPs.
Spend less time jumping between 10 ERP tabs and let an intelligent assistant do the legwork.
Worker Dispatch & Smart Scheduling
Moonschat cross-references live room occupancy, historical clean times, and worker availability to recommend the optimal schedule—then executes the changes.
- Calculates the exact staffing requirement based on stayovers vs. heavy checkout cleans.
- Checks worker regular shifts, travel distance, and overtime limits.
- Sends dispatch notifications and shift updates via email or SMS directly to cleaners once approved.
Timesheet & Invoicing Reconciliation
Catch billing leakage and payroll discrepancies before the month closes.
- Flags unsubmitted timesheets, phantom breaks, and clock-out mismatches.
- Audits missing quality-control (QC) sign-off photos before the client demands a credit note.
- Prepares clean invoice drafts matching contracted billing rates and approved hours.
Customer Communication & Dispute Resolution
Never guess what happened on site. Moonschat retrieves the timestamped visit log, the assigned cleaner, and the supervisor inspection report, then drafts a tactful, factual customer email.
- Surfaces the proof: Check-in time, duration, and uploaded completion photos.
- Drafts professional, client-ready responses balancing customer retention with operational truth.
- Proposes corrective actions (e.g., auto-scheduling a supervisor follow-up inspection).
Instant Operational SOPs & Internal Procedures
Save your operations team from endless phone calls with supervisors. Moonschat delivers concise, site-specific instructions grounded in your documented protocols.
- Delivers instant, bulletproof instructions on floorcare, hazardous material handling, and security clearances.
- Site-specific nuances: Gate codes, alarm disarmament sequences, and key handover rules.
- Multi-lingual adaptation: Explains complex SOPs in plain, concise language suitable for multicultural teams.
Moonschat
Right-sized models
Flash for instant lookups, Plus for everyday work, DeepThink for multi-step analysis. Chosen per message, not per contract.
Privacy by default
Names, emails and personal data are masked before anything reaches the model, and restored only inside your own session.
Scoped to the team
Every site, location and department gets its own workspace and its own instructions for the bot.
Moonschat
Why it is different
Grounded in your records
The answer is built from your schedules, sites and invoices, not from the open web. If it is not in your data, Moonschat says so.
Scoped to the team
Every site, location and department gets its own workspace and its own instructions for the bot. The answer respects the boundary.
Privacy by default
Names, emails and personal data are masked before anything reaches the model, and restored only inside your own session.
Skills for real workflows, and the model behind them
A job, packaged once
A multi-step job such as a dunning run or a reconciliation review is packaged once as a Skill: versioned, permission-aware and auditable, so the agent can carry it out without a developer.
The web, on your terms
Turn web search and page fetching on when the answer sits outside your system. Leave them off and Moonschat answers only from what you already know.
Chosen per message
Flash for instant lookups, Plus for everyday work, DeepThink for multi-step analysis. Chosen per message, not per contract.
Security & trust
Masked before it reaches the model, scoped to the team, auditable by default. Moonschat answers from your records and nothing else.
FAQ
Questions we get asked
Where does Moonschat get its answers?
From your own records — schedules, sites, timesheets and invoices. It does not answer from the open internet, and where the answer is not in your data it says so rather than filling the gap. Each answer arrives with the record behind it, so you can check the working rather than take the summary on trust.
Can it see data from other clients?
No. Every site, location and department gets its own workspace, with its own instructions for the bot, and the answer stays inside that boundary. Your records are the only ones it reads.
What is a Skill?
A multi-step job — a dunning run, a reconciliation review — packaged once so the agent can carry it out on its own. A Skill is versioned, permission-aware and auditable, which is what makes it something you can leave running rather than something you have to watch.
Can I restrict what it can access?
Yes. Access follows the same role-based permissions as the rest of the platform, and the two routes out of your system are switches you hold: turn web search and page fetching on when the answer sits outside it, leave them off and Moonschat answers only from what you already know.
How is personal data handled?
Names, emails and personal data are masked before anything reaches the model, and restored only inside your own session — the model works on the shape of the record rather than on the people in it. The wider picture, from hosting to retention, is on security and data.
Goes well with
Put the AI inside your operation.
Thirty minutes, live product, no slides. Bring the question you currently answer by opening three screens, and type it once instead.