AI for SMEs: where to start without spending thousands of euros

Illustration: scattered documents (email, spreadsheet and chat) organized into folders with the help of AI

What if your company were falling behind while you wait for AI to become simpler, cheaper and less risky?

You don’t need to invest thousands of euros in artificial intelligence. You need to start by improving one small part of your business.

Are you worried about investing in artificial intelligence and not getting your money back?

Have you heard about ChatGPT, AI agents and automation, but you don’t know how to apply them to your company?

Would you like to improve your team’s productivity, but you don’t have the budget to hire AI specialists or the time to learn dozens of new tools?

And what if your competitors started working faster, replying to their customers sooner and cutting administrative tasks while you carry on doing things the way you always have?

If any of these questions sounds familiar, you are not alone.

Thousands of small and medium-sized businesses are in a similar situation. They know artificial intelligence is changing the way we work, but they are not sure where to start, how much to invest or what results to expect.

And that is understandable.

As a business owner, you have to keep costs under control, look after your customers, pay your employees and make sure the business keeps running. You can’t afford to experiment aimlessly or spend money on solutions that don’t solve any real problem.

But there is something important you should bear in mind:

The biggest risk is not starting with a small investment. The risk is to keep spending time and money on tasks that could be done far more efficiently.

1. Before talking about artificial intelligence, let’s talk about your company

Imagine that tomorrow one of your best employees leaves the company.

Could a colleague take over their tasks without much difficulty?

Where are their contacts, their procedures, their templates and their knowledge about your customers?

Are they documented in a shared system, or scattered across emails, WhatsApp conversations, Excel files and personal folders?

Now imagine a customer asks you about a quote you prepared six months ago.

Can you find it in thirty seconds, or do you need to ask several people for help?

And how many hours does your team spend every week copying information from one document to another, updating spreadsheets, preparing reports or searching for files that should be available to everyone?

We often think we need artificial intelligence when, in reality, the first step is much simpler: organising the information and the processes we already have.

Technology can’t automatically fix a company that doesn’t know where its data is, how its tasks are carried out or who is responsible for each process.

That is why, before adopting sophisticated tools, it is worth building a solid foundation.

And you can do it gradually.

2. First step: organise your company’s knowledge

You don’t need to start by buying expensive business software.

Start by identifying what information your team uses to do their work.

For example:

  • Information about products and services.
  • Prices, quotes and commercial terms.
  • Internal procedures and working rules.
  • Onboarding manuals for new employees.
  • Frequently asked questions from customers.
  • Email and document templates.
  • Sales, customer service and administration processes.

Then organise this information in a simple structure, divided by department, function and type of document.

You can begin with the tools you already have, or use accessible solutions such as shared folders, Google Drive, Microsoft 365 or a documentation tool like Obsidian, depending on your business needs.

The goal is not to pile up documents. It is to make sure the important information is organised, easy to find and possible to keep up to date.

Think about what this means for a company with ten employees.

If each person spends ten minutes a day looking for information or asking colleagues, we are talking about more than three hundred hours a year, based on ten employees and around 220 working days.

How much is that time worth to your company?

Organising knowledge does not automatically eliminate all those losses, but it lets you see where they happen and start reducing them.

It also creates a foundation so that, in the future, you can use AI assistants able to consult authorised company information and help your team find answers.

Your first investment doesn’t have to be in artificial intelligence. It can be in putting things in order.

3. Second step: connect information to day-to-day work

Once the information is organised, the next question comes up:

What is the point of having perfect documentation if your employees keep copying the same data by hand every day?

Think about a typical customer journey.

A salesperson receives an enquiry by email or WhatsApp. Then they copy the details into a spreadsheet, prepare a quote, check the prices in another document and send an email. Later, someone updates the status of the opportunity and prepares a report for the sales manager.

How many times is this process repeated during the week?

At how many points can information get lost?

And what would happen if part of that work could be automated?

You can start by connecting a few tools through their native integrations, their APIs or automation platforms. When needed, you can also bring in AI agents or protocols such as MCP to provide controlled access to certain tools and information sources.

You don’t need to connect every system in the company from day one.

Choose a single repetitive process, for example:

  • Automatically recording the data from a contact form.
  • Generating a draft reply to frequently asked questions.
  • Creating a summary of a sales meeting.
  • Preparing a weekly report from existing data.
  • Alerting the manager when a sales opportunity has gone too long without follow-up.

Start with one task, check the result and improve the process before expanding it.

But do it carefully: protect your customers’ data, control access permissions and ask for human review whenever a decision could have significant consequences.

Automating doesn’t mean adding more technology. It means removing unnecessary work.

4. Third step: analyse what each job really involves

This is one of the steps with the most potential, and also one of the most overlooked.

Before deciding which tasks artificial intelligence should do, you need to understand how your team works.

Talk to your employees.

Ask them:

What tasks do you repeat every day?

What information do you have to copy from one system to another?

What work takes you longer than it should?

What problems keep coming up again and again?

Which tasks depend exclusively on one person’s experience?

Then sort the activities into three groups.

Group A: repetitive tasks that could be automated.

For example, classifying requests, extracting data from documents, preparing drafts, updating records or generating routine reports.

Group B: tasks where AI can help people.

For example, preparing sales proposals, researching a market, summarising meetings, writing content or analysing information before making a decision.

Group C: tasks that require judgement, experience and human responsibility.

For example, negotiating a complex contract, resolving a serious conflict with a customer, making strategic decisions or taking responsibility for a business decision.

The idea is not to replace people indiscriminately.

The idea is for your team to spend less time on mechanical tasks and more time on the activities where they truly add value.

And there is something you shouldn’t forget: employees know many of the everyday problems that management doesn’t always see.

Involving them from the start helps uncover better opportunities for improvement and reduces resistance to change.

5. Fourth step: turn your best employees’ experience into a business asset

Imagine you have a salesperson who closes more deals than the rest of the team.

Do you know exactly what they do differently?

How do they identify a customer’s needs? What questions do they ask? How do they handle objections? When do they follow up? What information do they need before preparing a quote?

If all that knowledge exists only in their personal experience, the company depends too much on one person.

You can start documenting best practices through interviews, examples of real deals, sales procedures and training materials.

With the appropriate legal safeguards and permissions, you can also use call summaries, transcripts and sales records to identify useful patterns.

Then turn that knowledge into resources the whole team can use:

  • Sales conversation guides.
  • Lists of questions to uncover customer needs.
  • Follow-up procedures.
  • Proposal templates.
  • Internal assistants that help prepare replies.
  • Training materials for new employees.

In this way, individual experience can gradually become shared knowledge.

This doesn’t mean every salesperson will achieve the same results. Every customer is different and human experience remains essential.

But it does mean your company can learn from what works, share it and keep improving.

The knowledge that helps you sell more, serve customers better and work with higher quality shouldn’t disappear when someone leaves.

6. How much do you need to invest to get started?

This is where many business owners get stuck.

They think they need to hire a team of specialists, buy a complex platform and redesign all their systems.

But not every company needs a digital transformation project costing tens of thousands of euros.

In many cases, the most sensible approach is to start with a small trial, using tools the company already has and adding new solutions only when there is a clear need.

A possible roadmap would be:

Phase What you can do Goal
Week 1 Identify a specific problem and measure the time it takes Know where it is worth acting
Week 2 Organise the documentation needed to solve it Build a reliable information base
Week 3 Try an automation or an AI assistant Check whether it works in practice
Week 4 Compare results, fix errors and decide the next steps Make decisions based on results

The timeline is indicative. Some projects can be completed sooner; others will need more time, especially if there are legacy systems, disorganised data or specific security requirements.

To begin with, you can choose a task where the cost of a mistake is low and the results are easy to measure.

For example, if preparing a weekly report takes four hours and you manage to cut that to two, you will have recovered two hours a week.

But don’t confuse time saved with money automatically earned. To know the real benefit, you need to consider the cost of the tools, maintenance, human review and how the freed-up time is actually used.

The right question is not how much AI costs.

The question is:

What specific problem are we going to solve, how much does it cost us today, and how will we know whether the solution is worth it?

7. You don’t need to do everything. You need to start with the right thing.

If you run an SME, you probably have many priorities and few resources.

That is why I wouldn’t recommend starting with ten new tools or trying to automate every department at once.

I would recommend choosing a single problem that meets three conditions:

  1. It happens often.
  2. It consumes time or causes errors.
  3. Its improvement can be measured simply.

It could be handling sales enquiries, preparing quotes, searching for documentation or producing reports.

Document how the work is done today. Set a starting baseline. Try a small solution and compare the results.

If it works, expand it.

If it doesn’t, learn, correct or change approach.

That is the real value of moving forward in stages: you don’t have to bet your whole budget on an idea you haven’t yet validated.

And remember that a simple solution your team uses every day is usually worth more than an advanced platform nobody knows how to make the most of.

8. Digital transformation doesn’t start with a big investment. It starts with a decision.

Maybe today you don’t need an AI agent running your whole company.

Maybe you need your employees to find the right documents without wasting time.

Maybe you need to stop copying the same data between three tools.

Maybe you need to know why some sales opportunities are lost.

Or maybe you need the knowledge of your best employees to stop being scattered and become an advantage for the whole business.

Any of these improvements can be a good starting point.

You don’t have to become an expert in artificial intelligence. Nor do you need to hire an in-house technical team straight away.

What you need is to understand your processes, identify the right opportunities and move forward with a method that makes sense for your company.

Artificial intelligence can help you, but the results depend on the quality of the information, how the processes are designed, how well the team adopts the changes and responsible implementation.

Don’t wait for the perfect budget, the perfect team or the perfect plan.

Start with one task. Measure the result. Learn from the experience. Improve the next process.

A small change that works today can become the foundation of a much bigger improvement tomorrow.

The question is not whether your SME needs to adopt all the artificial intelligence available. The question is what you can improve this week so that your company works a little better than it did last week.

Where to start in your company

If you want to explore how to apply AI and automation to your business, start by identifying one task that takes up too much of your time.

You don’t need to share confidential information or prepare a complex project. It is enough to describe the problem, how you currently handle it and what you would like to improve.

From there, you can assess which part can be simplified, which tools could help and what the most affordable way of testing it would be.

Fewer promises. Less complexity. More real improvements, step by step.

Because the best way to start with artificial intelligence is not to buy more technology.

It is to solve a problem you already have, and solve it better.

Initial automation diagnosis for SMEs

Would you like to find out which tasks in your company could be simplified with AI?

Tell me what your company does, which process takes up the most time and which tools you currently use.

We will analyse one specific process and explore possible improvements, starting with the simplest and most affordable options.

There is no need to transform your whole company at once. First we identify an opportunity; then we decide whether it is worth investing.

Request a first conversation →

Want to apply this in your company?

Tell us about your project and our team will help you turn it into a clear plan.

Talk to our team →