Trabalho XXI failed. What it means for SMEs (and why investing in AI just got more obvious)

Trabalho XXI failed: three barriers and a document

Portugal’s proposed labour reform, “Trabalho XXI”, collapsed without an agreement at the Social Concertation table. For an SME, this isn’t just politics. It’s an operational signal: the rules stay as they are (for now), but uncertainty increases.

And uncertainty has an immediate effect on the ground: it raises the cost of a hiring mistake.

Note: this is not legal advice. It’s a practical management read.

What was on the table (what actually changes life for operators)

The proposal touched 100+ articles of the Labour Code. But 4 areas matter most for SMEs managing capacity, risk, and cost:

1) More room for fixed-term contracts

Longer maximum durations and broader grounds. For many SMEs, the translation is simple: more time to reduce the risk of “hiring wrong”.

2) The return of individual time banks

A mechanism to absorb seasonal peaks without immediately hiring. If your business has campaigns, seasonality, or uneven demand, this directly affects planning.

3) Termination and reinstatement (the radioactive part)

The proposal expanded the possibility for employers to ask courts to avoid reinstatement in cases of unlawful dismissal. This was singled out as one of the most controversial changes [2].

Whatever your political stance, this is the behavioural driver: when legal risk goes up, the hiring threshold goes up with it.

4) Outsourcing after redundancies

There was an intention to roll back restrictions on outsourcing for 12 months after collective redundancies/job extinction for the same needs. Supporters saw restructuring flexibility; unions saw perverse incentives.

Why it failed: 3 boards, 3 potential vetoes

This didn’t fail purely on a “business vs unions” axis. Any serious reform has to pass 3 tests at the same time:

Social Concertation: without a minimum agreement (or at least UGT buy-in), the proposal is politically fragile [1][4].
Parliament: a minority government needs shifting coalitions. Chega is now explicitly part of the labour-law equation, with its own conditions [3].
The Presidency: the veto signal is clear when there’s no concertation agreement [5].

Practical translation: even if it goes to Parliament, expect back-and-forth, concessions, diluted versions, and more months of uncertainty.

What changes now for SMEs: almost nothing. The problem is what comes with it.

In the short term, the most likely outcome is:

rules stay as they are
the debate continues

The real risk for an SME isn’t “rigid law” vs “flexible law”. It’s not knowing what the rules will be in 6 to 12 months.

That uncertainty mostly penalises one thing: growing headcount by reflex.

Because when the cost of a mistake rises, the rational response is:

hire fewer
hire better
build operations that don’t need “more people” to absorb more volume

The operational conclusion: hire fewer and better (and stop hiring to fight fires)

A common SME failure mode is:

the process breaks
everything becomes urgent
you hire to patch the hole
the hole moves

In a high-uncertainty context, that loop gets more expensive.

A simple rule that works:

Before opening a role, list the 10 weekly tasks that person would do.

Split them into:

repetitive, predictable tasks (admin, follow-ups, triage, reporting, documentation)
context-and-judgment tasks (negotiation, exceptions, quality, client relationships)

In 2026, the first group shouldn’t be your reason to hire. It should be your reason to automate.

AI as elastic capacity (without precariousness)

The useful promise of AI for SMEs isn’t “replacing people”. It’s removing repetitive work so your team can focus on what requires thinking.

When done properly, the result is concrete: fewer fires, more control, more predictable throughput, with the same team.

Where I see fast impact in SMEs:

workflow automation across tools (less copy-paste, fewer things falling through gaps)
internal assistants (company knowledge, standard answers, onboarding, proposals)
lead handling and follow-up (fast response + clean human handoff)
backoffice (email triage, requests, invoices, reporting)
practical training so the team uses AI without hype and without unnecessary risk

The goal is capacity where the business bleeds time, not flashy demos.

If you want to act instead of comment: what I’d do in 2 weeks

If you’re at a point where “hiring more” feels inevitable, I’d start with a short operational diagnosis:

map the repetitive work that’s draining time
pick 2 to 3 automations with clear impact
design the process so work is visible and under control
only then decide whether you still need to hire, and for what exactly

Start with one working session

In a single focused session, we map where time is being lost, identify high-leverage automations, and decide what to build first.

Book a session →