How to optimize the 40-hour Law with a Workforce Management Agent

Law 21,561 —the so-called 40-hour Law— was passed in May 2023 with two headlines applauded by the press: "work-life balance" and "quality of life". Both are true. What almost no one said in the photo op is that every hour removed from the weekly calendar forces a redesign of a chain of operational decisions that affect everything: shifts, headcount, buses, food service, compliance, overtime, peak coverage, turnover and even collective bargaining.

Nine days before the second step takes effect (from 44 to 42 hours), most of the companies we know still think of this problem as a "payroll adjustment". That is a mistake. Whoever plans well gains productivity. Whoever does it with improvised spreadsheets loses money and, eventually, people.

1. Law 21,561 in operational terms

The full text defines the gradual reduction of the maximum ordinary weekly workweek. If you prefer the executive version, this is what matters:

Beyond the ordinary workweek, the law also updates:

• Hours bank: the option to agree on bands (up to 52 hours one week and fewer the next), as long as the average over the cycle respects the cap.
• Overtime: it stays capped at 2 hours per day, but is now computed on the new base. A company that used to pay 3 overtime hours per shift ends up seeing an 8-12% rise in the overtime line.
• Early 4x3 distribution: the option of 4 workdays and 3 days of rest, available to those who sign an agreement. Attractive but operationally aggressive if it is not modeled well.
• Article 38 special shift schedules (mining, aviation, healthcare, ports) keep their rules, but the average cap also drops.
• Parental care: the right to a time band for workers with children under 14. A tough new constraint for whoever assigns shifts.

2. The four real pains of cutting one hour

When we accompany schedule planning for the reduction from 44 to 42 hours, the same four pains come up again and again; only the numbers change:

2.1 Headcount

If your operation runs 24/7 with four rotating shifts, going from 44 to 42 hours means the same workers each cover 2/44 ≈ 4.5% less time. To maintain the same output you need to compensate with more heads or more productivity. The temptation is to hire; the cost can be +3% to +5% of the monthly CLP payroll. Before hiring, you have to see whether a schedule redesign covers the gap for free.

2.2 Peak coverage

Retailers, logistics and customer service do not spread demand evenly. Cutting one hour from the entire workforce simultaneously creates specific gaps during peak hours. Companies that cut linearly end up with empty registers on Friday at 7:00 p.m. and surplus staff on Tuesday at 11:00 a.m.

2.3 Silent compliance

The new cap is often breached unintentionally: a transfer that crosses the clock, a shift swap accepted over WhatsApp, a mandatory training session that counts as work time. The Labor Directorate increased enforcement since 2024 and fines range from 1 to 60 UTM per affected worker. Multiplied over a monthly cycle, it is expensive.

2.4 Collective bargaining

When the law dropped to 44, many unions signed bonuses in exchange for the cut. Now the union shows up with new demands and bonuses on the table. If HR does not have a precise simulation of the operational impact, it negotiates blind.

3. The math that almost no one gets right

The real operational question is not "how many more workers do I need?" but "what is the schedule that covers my demand with the smallest legally possible headcount?". The two look alike but are not the same thing.

A simplified example. A plant with constant demand for 100 simultaneous people for 24 hours, 7 days a week, before the law:

• Weekly person-hours needed: 100 × 24 × 7 = 16,800
• Maximum workweek per worker: 45 h
• Minimum headcount: 16,800 / 45 = 374 workers

With 42 hours, assuming the same demand:

• Minimum headcount: 16,800 / 42 = 400 workers (+26, +7%)

But this calculation assumes constant demand, zero absenteeism, zero conflicting shifts and zero peaks. In reality demand varies, there is 4-8% absenteeism, night shifts are paid with a surcharge, transfers consume work time, and collective agreements prevent certain rotations. When you model all of this, the real headcount can range from +4% to +12% depending on the industry, or it can drop if the current schedule has slack.

That "slack" is the opportunity. Companies with fixed schedules (the same shifts from Monday to Sunday, regardless of demand) pay all year to cover a peak of a couple of hours. The agent attacks precisely that inefficiency.

4. What exactly a Workforce Management Agent does

ValueData’s Workforce Management Agent is not an automatic spreadsheet nor a model that spits out an Excel file. It is an autonomous agent that combines mathematical optimization (MILP + heuristics) with AI agents to read collective agreements, understand the operation and explain every decision. Specifically:

Inputs it consumes

• Historical demand by time slot (14 days minimum, ideally 12 months).
• Current workforce: skills, contracts, agreed workweeks, time bands.
• Collective agreements in force (a specific agent reads them and extracts rules).
• Legal constraints: article 38, compensatory rest, parental care.
• Costs: base pay, night surcharges, overtime, shift bonuses.
• Event calendar: month-end closings, campaigns, scheduled maintenance.

What it delivers each week

• Optimal schedule of shifts for the next 4 weeks, with the minimum cost compatible with projected demand and all legal constraints.
• Proactive alerts: "worker X projects 172 hours in the cycle, 3 over the cap" or "the Saturday 3:00 p.m.–11:00 p.m. shift has 20% less staffing than the historical peak".
• What-if scenarios: it simulates 4x3 vs. 5x2 vs. 7x7 for an area and shows the impact on cost, compliance risk and estimated satisfaction.
• Natural-language explanation of each recommendation. HR doesn’t see a number alone; it sees why the agent moved María from shift A to B, which rule of the agreement applies and how much it saves.

Integrations

The agent connects with the tools you already have: SAP SuccessFactors, Buk, Rankmi, a Kronos biometric clock, or Excel spreadsheets while they migrate. It does not need to replace the HR layer, just read it and write back the approved schedule.

5. Case: 820 workers at a manufacturing plant

A recurring base case in the industrial sector. A company with three plants in the Santiago Metropolitan Region, 820 workers, a regime of 5 rotating shifts (including nights), demand with strong bimonthly peaks (clients’ accounting close) and a collective agreement that limits shift changes to 72 hours’ notice.

Summary: the company hired 17 more people (not 50 as its first manual calculation projected), nearly halved its overtime pay by redistributing the load, eliminated the risk of sanction and improved coverage of the critical peak. The net monthly savings (CLP 23 M) paid for the implementation in less than 6 months.

6. A realistic implementation (6-10 weeks)

Order matters. Skipping phases produces schedules that look valid but break the collective agreement by the second week.

1. Diagnosis (2 wks). We review payroll, agreements, historical demand and the attendance system. Out come the base model and the hard constraints.
2. Technical integration (3-4 wks). We connect SAP/Buk/Rankmi or whichever system you use. A reading agent processes the collective agreements and translates them into rules.
3. Controlled pilot (2 wks). We run the agent in shadow mode over one plant or business unit. HR compares the agent’s schedule with the manual one and the ValueData team tunes the parameters.
4. Rollout (1-2 wks). The agent becomes the official source of the weekly schedule. A human stays in place approving the major changes.

7. Frequently asked questions