AI in Janitorial Services: Scheduling, Quality and Contract Margin

Cleaning contracts are won on price and lost on complaints, with margin determined by labour hours that are frequently not measured per site. The work happens overnight in buildings nobody supervises, which makes verification the central operational problem. This guide covers where AI helps, drawing on FISTA Solutions' AI agents work in field operations. It complements how to build a dispatch optimization agent and the AI for field operations whitepaper. This article is general guidance, not legal advice.

Why do labour hours decide margin?

Because labour is almost the entire cost. A contract priced on assumed hours and delivered in more is loss-making, and in many contractors the actual hours per site are unknown because time is recorded against the company rather than the contract.

That means unprofitable contracts persist undetected, cross-subsidised by profitable ones, until an aggregate margin decline prompts an investigation that finds them.

Element	Measurable	Typically measured
Hours per site	Yes	Rarely
Margin per contract	Yes	Rarely
Complaints by site	Yes	Sometimes
Consumable usage by site	Yes	Rarely
Task completion	With verification	Rarely
Staff turnover by site	Yes	Sometimes

What determines renewal?

Complaints. Clients rarely terminate on price alone; they act on whether the building is consistently clean and whether issues are fixed when raised.

Complaint patterns by site predict renewal risk well before the renewal conversation, and they are usually recorded somewhere unstructured. Structuring them and watching the trend converts a surprise non-renewal into a manageable account problem.

What is the verification problem?

Work happens overnight in buildings nobody supervises. Verifying that it was done, and done to standard, without stationing a supervisor at every site is the operational challenge the whole sector faces.

The available approaches — task check-ins, photographic evidence at defined points, sensor data where it exists, and periodic inspection — each give partial coverage. Combining them into a picture of what was actually done, rather than what was scheduled, is what makes quality manageable.

Why do consumables matter?

Because they are a controllable cost typically ordered on habit rather than consumption. Usage varies with building occupancy, and a standing order that suited a fully occupied building does not suit a hybrid one.

Ordering against actual consumption reduces cost and reduces the run-outs that generate complaints, which are among the most common and most avoidable.

What about scheduling?

Constrained by site access windows, travel between sites for mobile teams, staff availability, and the contracted hours each site requires. It is a routing and rostering problem where the hard constraint is contracted delivery rather than efficiency.

Continuity matters too: staff who know a building work faster and better, and rotating them constantly costs hours that do not appear in any plan.

What about staff turnover?

The sector's largest hidden cost, and it varies by site. High turnover at a particular contract usually indicates something about that site — access difficulties, unreasonable client expectations, or hours that do not match the work — and identifying it is cheaper than continuously recruiting.

Who should own it?

Operations, with account management owning the client relationship and complaint response. Contract profitability should be visible to both, because the commercial and the operational levers are different and both are needed.

How is it evaluated?

Margin by contract and site, complaints and resolution time, hours delivered against contracted, consumable cost per occupied area, and staff turnover by site. Hours delivered alone says nothing about whether the contract makes money or the client is satisfied.

What goes wrong?

Hours recorded at company level, which hides unprofitable contracts. Verification by periodic inspection only. Consumables on standing orders. Scheduling that rotates staff without regard to site familiarity. And complaints handled individually without watching the pattern.

What does it cost to run?

Low; the data volumes are small and the scheduling problems are modest. The investment is in capturing site-level hours and completion data, which requires a change to how work is recorded rather than a system purchase.

What should you do first?

Calculate margin for your ten largest contracts using actual hours rather than contracted hours. The distribution is usually wider than expected, and it identifies which contracts need renegotiation or operational change before renewal comes round.

How does occupancy data change cleaning?

Fundamentally, where it is available. Hybrid working has broken the assumption that a building is used the same way every day, and cleaning to a fixed specification in a building that is half empty on Mondays and Fridays wastes hours on the quiet days and under-delivers on the busy ones.

Occupancy-led cleaning — adjusting frequency and focus to actual use — reduces cost and improves perceived quality simultaneously, which is rare. It requires occupancy data the client usually holds and rarely shares, which makes it a commercial conversation as much as an operational one.

What about specification and pricing?

Contracts are frequently priced against a specification written years ago for a building that has since changed. Comparing the specification against actual occupancy and observed need gives both parties a basis for renegotiation that is evidential rather than positional, and contractors who bring that conversation tend to keep contracts rather than lose them on price.

How FISTA Solutions helps

FISTA Solutions builds cleaning operations systems with site-level hours and margin visibility, task verification combining check-ins and evidence, consumption-based consumables ordering, scheduling that preserves site familiarity within contracted constraints, and complaint pattern tracking by site, through AI agents, AI enablement, and forward deployed engineers. The record behind the approach is 150+ projects for 50+ companies with 47% efficiency gains.

To find out which contracts actually make money, message FISTA on WhatsApp, or read the AI for field operations whitepaper.