Restaurant Staff Training Requirements in Colombia: what the law mandates and what you choose to train

What training does the law require of a restaurant in Colombia?

The law requires two fronts: food handling for everyone who touches product and the training tied to the SG-SST occupational safety system, with an emergency brigade and first aid, and neither one tolerates improvisation.

Everything else, from suggestive selling to table protocol, is the owner's call, and that is where the trap sits, because many restaurants mix both boxes into a single yearly plan and end up paying for classroom hours to teach how to uncork a wine while the food-handling record of a new prep cook is still unsigned. My position is firm: the mandatory part is met with the sufficient minimum, per person and per location, with attendance list, content, assessment and signature on file, and it is confirmed with your municipal health office, because requirements can vary by territory; the voluntary part is measured against a shift number or it simply does not happen.

When the usual training plan falls short?

The usual plan, one course a year for everyone and a binder of certificates, falls short the day turnover outruns the calendar.

If a server starts in March and the group course is in October, that server spends seven months on the floor untrained, and the signal shows up in the register: average checks that drop on shifts with new people, more plates sent back and complaints that repeat the same mistake. The number of people at stake is not small either, since bars, gastropubs and nightclubs generate close to 96 thousand jobs a month according to Portafolio (2024), and in a sector that moves that much employment a yearly plan arrives late by design. When a new hire's food-handling record does not exist on inspection day, the problem stops being about service and becomes a SANITARY one, and that is the real limit of the original model.

Alternative 1: short training inside the shift

In-shift training suits the small or mid-size restaurant with high turnover and a tight register, where pulling four servers off the floor for a full afternoon costs more than the course itself. It works in ten- or fifteen-minute blocks before opening, with one topic a day (a declared allergy, the dessert sale, a split check, closing a table) and someone who checks on the floor rather than in a test. Its switching cost is low in money and high in discipline, because the floor manager has to prepare the block, repeat it and log who received it. The downside deserves plain words: this format does NOT replace the food-handling course or the brigade training, which demand hours and records that fifteen minutes cannot produce. For example, a daily fifteen-minute block held six days adds up to an hour and a half of practice per week without closing a single table.

Alternative 2: outside training only for what the law requires

Hiring outside training makes sense when it is limited to what the law requires and bought with records that can withstand an inspection. It is the alternative for the owner with two or more locations, or with a production kitchen, who needs to certify many people at once and has no one inside with the technical authority to sign off on sanitary training. That is how the yearly minimum set by Resolution 2674 of 2013 gets covered, and the switching effort lies in logistics: shifts covered location by location and a file with each handler's name, date, hours and assessment. The most repeated mistake is stretching that contract to service topics, because an outsider teaches generic protocol and does not know your menu, your check or your peak hour. Pay outside to comply. And to sell more, train in-house, where the indicator shows up every night.

Alternative 3: a buddy per position for new staff

Assigning a buddy per position is the way out for the restaurant that already has a stable core and loses money in the first two weeks of every new hire. A server with judgment walks the newcomer through the first shifts, teaches the menu by what actually sells and corrects at the table, in front of the guest. The switching cost is moderate, since the buddy has to be rewarded somehow, in schedule or in a bonus, and someone has to write a short list of what the newcomer must master before working alone. But a real tension of the trade appears here: the server who sells best is often the worst teacher, and the owner almost always picks by check average. The fix is to choose for patience and consistency, and to judge the buddy by how the trainee performs at the end of the first month, never by the buddy's own sales.

What happens if you cut all training to save cash?

If you cut all training, you save the first month and pay for it later with lost sales and with sanitary risk. Follow the full chain:

the new server cannot suggest and the check per table drops; without that revenue the owner squeezes schedules, the team gets tired and leaves; someone else comes in untrained and the cycle restarts, now with a handler without records on the day the inspector arrives. In the Masterestaurant method, Diego F. Parra treats training as protection of the check and not as an expense, which is why the voluntary part is cut only when it moves no indicator. Scale matters too, because the restaurants in the ANDI Gastronomic Sector Chamber generate more than 34,822 jobs (ANDI, 2024), and each of those positions depends on the business not improvising with its team when cost pressure tightens.

When NOT to change how you train?

Do not change how you train if the mandatory part is already up to date with records per person and per location, and if shifts with new people sell the same as shifts with veterans.

In that case the system works, even if it looks homemade, and moving it for fashion costs floor-manager hours that pay off more on the floor today. Nor is it wise to change in the middle of high season, because a new format needs weeks of adjustment and December forgives no rehearsals. I got this wrong for years: I recommended redesigning every plan I found, and some only needed filing and order. The honest test fits on one page. Compare the average check and the returns of shifts staffed by people with less than a month against the rest of the calendar; if the gap does not show up, leave the plan alone and spend that energy on the menu or on costing.

What changes when you separate mandatory from voluntary?

The most common mistake is training everything in one format, so the first thing I fix in an owner's plan is the ORDER:

close the mandatory part with proof per person and per location, then design voluntary training against a floor metric, because a service course with no metric is one more certificate for the binder. Every hour off the floor is paid twice, the server's hour and the table nobody serves, and that math is heavier now that Acodrés puts the rise in sector production costs at 26.3% (Infobae, 2025). That is why front of house training that works lives inside the shift, in ten or fifteen minute blocks. What happens if a restaurant trains nothing beyond what is mandatory?

What changes when you separate mandatory from voluntary — in practice

Month one looks fine, by month three the captain spends the night fixing tickets, around month six the best servers leave for places that train them, and new hires learn from whoever does it worst, so the early saving turns into turnover and a falling average check. Here is the paradox: the mandatory part moves cash the least and the voluntary part moves it the most, yet most owners invest the other way around. The fix is to meet the law with the sufficient, well-documented minimum and shift the freed budget to cheap in-house formats. Retention matters too. Bars, gastropubs and nightclubs generate close to 96 thousand jobs a month (Portafolio, 2024), and the ANDI chamber's restaurants add more than 34,822 jobs (2024), so your good server has options two blocks away. In Diego F. Parra's method, the internal micro-credential per role is the cheapest retention tool a restaurant has.