Sooner or later the admin overflows and the question arrives: should you hire someone, or automate part of the work? If you are self-employed (an autónomo) or run a small company in Spain, it comes back each time your business grows. It is often framed the wrong way: a salary gets compared with a software subscription, and nobody looks at what each option actually does. An employee costs more than the payslip shows; an automation can cost less than people fear, but only covers part of the work. This guide helps you decide whether to hire or automate, block by block, with your own numbers.
When to hire and when to automate: the basic rule
Automate work that is repetitive and follows the same steps every time, and hire people for work that needs judgment, human contact or decisions. In many companies, the right question is not which option to pick but how to split the work: a tool runs the routine workflows and a person takes care of the rest.
To get that split right, look at three things for each part of the work: what it involves, how often it comes up and the real cost of each route.
Hire vs automate: which work suits a tool and which a person
The first filter, before any cost comparison, is the nature of the work your business needs done. Repetitive, predictable processes can be automated; work with exceptions, judgment calls or customer contact needs experienced people. This table shows where to draw the line:
| Type of work | Examples | Best fit |
|---|---|---|
| Repetitive and predictable | Copying data between systems, sending reminders, filing invoices, updating a spreadsheet | Automate |
| Repetitive, with exceptions | Chasing late payments, checking paperwork before it goes to your gestoría | Automate the fixed part, keep a person for the exceptions |
| Reading and sorting text | Reading supplier invoices, sorting incoming emails, drafting a first reply | An AI tool, with a person checking the result |
| Judgment or customer contact | Negotiating a price, handling a complaint, making a decision on an exception | A person, supported by technology |
Picture a family-run travel agency in Valencia that receives requests by email, by WhatsApp and through its website form. Copying each request into the bookings spreadsheet and sending payment reminders is repetitive work for a tool; putting together a tailor-made trip takes a person's judgment. The two are different kinds of work: an automation does one thing the same way every time, without tiring, while a person can answer the phone, notice when things look odd and adapt to an unusual request.
How often does the work actually repeat?
Part of the answer depends on volume: the more often a piece of work repeats, the more it makes sense to automate it. If it only comes up a few times a month, doing it by hand may cost less than having it automated, so measure the real volume instead of guessing. A job with low volume, or low value, may not be worth the build.
Volume matters when you hire, too. A team has a limited capacity, so a growing business needs more people or has to free up the time of the ones it has. A well-built system absorbs extra volume without overtime, even in summer for businesses on the Costa Blanca, provided the process stays the same. To choose where to start, see which tasks to automate first.
The real cost of hiring someone in Spain
In Spain, a new hire costs the business more than the salary on the payslip: social security paid by the company, holidays, sick leave, a workstation, recruitment, training and turnover all come on top. Ask your gestoría for the total cost of the role over a full year before you compare it with anything else.
Base salary vs total cost of the role
The base salary is just the starting point of the comparison. Several of these expenses only show up during the first year:
- Social security: the company pays its own contributions on top of the gross salary.
- Holidays and sick leave: every year, the work stops for a while or someone has to cover it.
- The workstation: a desk, a computer, a phone and software licences.
- Recruitment: the job advert, the interviews and the choice take time away from your customers.
- Training and management: until the new hire knows your processes, their productivity is lower than their cost, and someone in your team has to explain and check.
- Turnover: if the person leaves, you start again, and what they had learnt goes with them.
In return, people bring what no software offers: judgment, initiative, human contact and the ability to deal with the unexpected. The relationships someone builds with your customers and suppliers over the years are hard to hand over to a tool.
The paperwork of a first hire
If you have no employees yet, your first hire also creates admin: you must register as an employer with the Social Security system, register the employee before their first day and issue monthly payslips. Many companies leave the contract and the payroll processing to a gestoría, which can also tell you which collective agreement (convenio colectivo) applies to your sector: the agreement can set minimum wages and working hours.
What it costs to automate a process
Automation costs money too, but in a different way: much of the spend comes at the start, with smaller ongoing expenses afterwards. Take into account the one-time build (studying the process, setting up the automation and testing it with real documents), the subscription to the software it runs on, maintenance when a supplier changes its invoice layout or your process changes, and your own time to explain the work and check the first results.
What gets expensive is often not the tools themselves but how you build them: on top of a disorganised process, you end up paying for something that does not work. Before you accept a proposal, ask for a breakdown of maintenance, line by line. For the full method, read how to work out the cost to automate a process and why maintenance is the hidden cost of automation.
Existing tools vs a custom build
Some of this work does not need enterprise systems: it can be covered by software you already pay for, such as email filters, your invoicing program or the integrations between apps your company uses. A custom build makes sense when the work runs across several systems that do not talk to each other, or when no existing tool is the right fit for your processes. Before paying for new automation tools, see how far no-code automation can take you.
Where does AI fit in the decision?
AI widens what you can automate, because it can read and write ordinary text instead of only following fixed rules. It still makes mistakes, so people need to check its work before it reaches a customer or the tax office.
AI and automation are not quite the same thing. Classic automation follows instructions set in advance: when an email arrives with an attachment, save it in a folder. A tool built on an AI model interprets content: it can tell an invoice from a quote or work out what a customer is asking for.
What an AI tool can take on
- Reading a receipt or a supplier's bill and extracting the date and the amount for your bookkeeping.
- Sorting incoming emails: quote requests, complaints, invoices, spam.
- Drafting a reply to a frequent question, for you to check and send.
- Summarising a long email thread or a document.
What AI cannot do on its own
AI can sound sure of itself and still be wrong, for example by misreading a figure or inventing a detail. It does not know your customers, takes no responsibility and does not notice when a case is an exception.
If you give it client data, data protection rules (the GDPR) apply to that processing, so read the privacy terms and check which companies can access the information. Let AI prepare the work and leave the decision to a human.
Do you need an AI agent or a custom AI?
An AI agent chains several actions on its own: it reads a request, looks up information and prepares an answer or the next step. That can save time, but the more it does unsupervised, the further a single mistake can travel. For a small business, a tool that does one clear thing is easier to check than a custom AI agent that tries to run all your workflows. For examples, see our selection of AI tools for freelancers.
What do British studies say about hiring and automation?
In short: they describe more firms adopting the technology, some roles replaced by automation and many employers choosing to train their current staff rather than recruit. These are UK studies, not Spanish ones, but the choice they describe is the same one a business in Valencia or Alicante faces.
AI adoption is rising, but the evidence is still incomplete
The Office for National Statistics follows business use of the technology in its Business Insights and Conditions Survey. In a freedom of information reply on how artificial intelligence is affecting employment, it reports that the share of UK businesses using some form of it has been rising steadily since the question was introduced. The same reply notes that, in its Opinions and Lifestyle Survey, workers in administrative, secretarial, sales and customer service roles were more likely than others to think their job could be at risk.
The Department for Science, Innovation and Technology commissioned research on AI adoption to understand how firms use the technology and the barriers. The factors it reports as most often holding firms back are a lack of identified need and limited skills and expertise; ethical concerns, high cost and unclear regulation are also rated as significant. A lack of identified need is a useful reminder: automation only pays when there is a real, repeated job to take over.
The government's assessment of capabilities and the impact on the UK labour market is candid: the available evidence does not yet give clear answers to many of the policy issues that matter. It also observes that autonomous agents are used for automation more often than for assisting human work.
Employers are retraining before recruiting
The CIPD, the professional body for HR, publishes a resourcing and talent planning survey. Its latest report found that some organisations replaced jobs with technology and automation over the last year, while upskilling existing employees was the most common response to recruitment difficulties. Employers also reported that higher application volumes did not translate into good-quality shortlists.
Skills England's employer guide on upskilling sets out practical training principles rather than recommending specific products, notes that workforce capability is often developed informally and is published with case studies from several sectors. The adoption plan for the digital and technologies sector, written by the government's champion for that sector, describes firms there facing the same practical issues as everyone else: skills, the effect on jobs, data and cyber security.
How a micro-business can build the skills
You do not need a training department to follow the spirit of the UK guidance. Learning can happen on the job: one member of staff tries a tool on a single routine chore, writes down what worked and shows the others. The CIPD report and the Skills England guide both favour developing talent in-house, and that suits a modest workforce well.
- Give each person a clear role in the new process: who prepares, who checks, who decides.
- Set aside a few hours at regular intervals for learning, as you would for any other professional skill.
- Ask the supplier of the services you use for help and support material before you pay for outside training.
- Note what goes wrong: that record is your own evidence of whether the change is worth it.
What this means for your decision
- Before you recruit, check whether someone already on your staff could take on the new work with some training and support.
- Count learning as part of the cost: whoever runs or checks an automation needs room to develop the skills.
- Look closely at the routine parts of admin and customer service roles, where UK workers were more likely to see a risk to their job.
- Take this research as context, not a forecast: the government's own assessment says the evidence is still developing.
Timescales and risks of each option
Neither option is instant. Hiring takes recruitment and onboarding before the new hire is up to speed; automating takes time to study the process, build, test and adjust. Whichever you choose, set a date to check whether the decision is actually working for your business.
The risks are different, too. With people, turnover and sick leave can leave a gap, and the know-how can end up in a single head. With a tool, a change in regulations, an error nobody notices, a free plan whose conditions change or a breach involving your customers' data can leave the system useless. Document your workflows to reduce both risks: a documented automation can run for years if the process does not change.
The middle ground: part-time work and a hybrid model
Often the answer is a mix: a tool takes the repetitive part, with a person handling the rest. That person can be you as the business owner, a member of your team or someone you hire part-time, now free to look after customers, deal with exceptions and check what the automation does.
In a hybrid model, decide from the start who owns the automation: who notices when it stops, reads the error message and updates it when something changes.
Part-time vs full-time: is half a role enough?
A part-time contract (contrato a tiempo parcial) for half the hours costs roughly half as much in salary and social security and keeps the human flexibility, although recruitment and training cost about the same. It becomes questionable when a good share of those hours goes into work a machine would do without errors.
Signs that point to automating or to hiring
Signs that automation wins
- You do the same work several times a week, in the same order.
- You copy information by hand from one system to another.
- The volume keeps growing and the backlog builds up.
- Errors come from tiredness, not from a lack of knowledge.
Signs that hiring wins
- The work changes every time and needs judgment.
- Personal contact is what sets your company apart, as it does for many service businesses.
- The procedure changes often, so a tool would soon be out of date.
- Customer relationships suffer because nobody has time to look after them.
Common mistakes when deciding
A frequent mistake is to compare prices only: a salary against a subscription. That comparison falls apart as soon as you forget what it costs to maintain the tool or to supervise the new hire. The question to ask is what each option does, not only its cost.
Other mistakes are automating a process that is not yet in order, taking on people to cover up an organisation or management problem, leaving the automation without an owner, and automating a task whose rules change every quarter: it is a poor candidate, because the build would need redoing each time. For more on these traps, read about the costliest automation mistakes and what you should never automate.
How to make the decision in six steps
Before you hire, work through this list:
- Write down exactly what the person you hire would actually do, and split it into blocks.
- For each block, note the number of times it comes up per day or per week and how long it takes.
- Mark whether it needs judgment or only execution.
- Set aside the blocks that repeat without variation: these are the ones you could automate.
- Ask your gestoría for the total cost of the role, full-time and part-time, not just the salary.
- Price the automation of those blocks, taking maintenance into account, and compare what is left with the cost of a hire.
You will often find that half of the work is repetitive. Automate it, and with the hours you free up, the other half may no longer justify a full-time role. To see how a project unfolds, read how automating a process works, step by step.
We have put together a free diagnostic that prices a task in about two minutes: four questions, with the result on screen. Run it on your heaviest blocks before you publish a job advert. If it concludes that none of them is worth automating, hiring may well be the right answer, and it will tell you that too.