Your Next Hire May Be AI

For twenty years, companies have bought more software while continuing to hire people to operate it. Evalon believes the next transformation is fundamentally different: software will no longer simply help people work – it will increasingly execute the work itself.

Founded by enterprise AI specialist Javier Santos together with Iberica Capital Partners, Evalon is building a new category of corporate infrastructure: specialised AI employees capable of executing complete business workflows across the systems companies already use.

The proposition is deliberately simple: instead of buying another AI tool, companies can add operating capacity.

For entrepreneurs, SMEs and mid-sized organisations, that could fundamentally change the economics of building and scaling a company.

FROM SOFTWARE TO EMPLOYEES

Javier, companies already have email, CRM, ERP, accounting software and now AI copilots. Why do they need Evalon?

Because companies do not have a software problem anymore.

They have an execution problem.

We have spent twenty years digitising companies. Yet walk into almost any business and you will still find highly capable people spending enormous amounts of time reading emails, copying information, updating CRMs, preparing documents, following up customers, reconciling information and moving data between systems.

The software is there. The intelligence is increasingly there. What is still missing is execution. Evalon adds that execution layer.

We deploy specialised AI employees that can work across the systems a company already uses, execute defined workflows, update those systems, communicate when authorised, verify what has happened and escalate exceptions to humans.

The customer proposition is therefore very simple:

Do not buy another AI tool. Add operating capacity.

Everyone is talking about AI agents. Why do you deliberately call them AI employees?

Because companies do not hire technology. They hire people to perform roles. An agent is a technical concept. An employee is a business concept.

When a CEO says, “I need someone to manage sales follow-up,” he is not asking for an LLM, an API or an orchestration framework. He needs the job done. That is how we think about Evalon.

An AI employee receives a defined role, objectives, company context, authorised systems, operating rules, permissions, performance criteria and escalation points.

It can then execute the recurring operational workload associated with that role. And critically, it can operate proactively.

It does not necessarily need someone to open a chat window and tell it what to do every time. A new lead, an incoming document, a missed deadline, an unanswered customer, a CRM event or a scheduled process can trigger work automatically.

That persistence and proactivity are part of what turns AI from a tool into operating capacity. That distinction is important.

A chatbot answers. A copilot assists. An AI employee executes.

Give me an example that any business owner can immediately understand.

Sales is probably the easiest. Imagine you have 100,000 potential customers in your database.

A human commercial team cannot research, personalise, contact, follow up and maintain the CRM for every one of them consistently. An Evalon sales employee can.

It can segment the database, enrich information, prepare personalised communications, execute authorised outreach, follow up, detect responses, classify opportunities and update the CRM continuously.

When a prospect reaches the point where human judgement, negotiation or relationship becomes valuable, the AI employee hands the opportunity to a person.

We have already operated a commercial workflow processing more than 100,000 contacts in 66 days. That deployment generated approximately €132,000 in incremental revenue, with an observed ROI of around ten times the estimated operating cost of the workflow.

The important thing is not that AI can write an email. Everyone knows that now. The important thing is that the commercial process gets executed.

So Evalon is not trying to replace the software companies already use?

Exactly the opposite. Companies have already invested enormous amounts in CRM, ERP, email, accounting systems, databases, document repositories and specialised applications. We do not want them to throw those systems away.

Evalon operates across them. An AI employee might read information from an email, check data in the CRM, retrieve a document, execute an authorised action in another system and then update the CRM automatically.

That is why we describe Evalon internally as an AI execution layer. But the customer does not need to understand the architecture. The customer needs to understand the outcome: the work gets done.

WHAT DOES AN AI EMPLOYEE ACTUALLY DO?

What jobs can Evalon employees perform today?

We currently have more than 60 AI employees developed across 16 functional modules, supported by approximately 30 proprietary workflows.

They cover areas such as sales and CRM, marketing, customer operations, document processing, reporting, finance, legal and compliance, HR and broader operational processes.

But I would not focus too much on the number. What matters is that they belong to the same operating environment.

A company can start with one commercial employee and later add CRM, marketing, customer support, finance, legal or operations. One employee becomes several. Several employees become a digital department. Eventually, that becomes a digital workforce.

How difficult is it to introduce one of these employees into an existing company?

Exactly like onboarding a member of the team rather than implementing a traditional enterprise technology project.

The company chooses the role. Then we define the objective, process, authorised information, systems, permissions, company knowledge, communication style, escalation rules and success criteria.

The employee is connected to the systems it needs and begins operating within those boundaries.

For standardised workflows, initial activation can be measured in days rather than months. Our current product architecture targets initial workflow activation in approximately two to five days.

That matters enormously for SMEs. They cannot spend nine months implementing an AI transformation programme.

They need something useful. And they need it now.

Can several Evalon employees work together?

Absolutely. And that is where the economics become particularly interesting. Imagine a commercial opportunity.

One AI employee identifies and qualifies the prospect. Another maintains the CRM. Marketing prepares supporting material. Legal prepares the relevant documentation within its authorised scope. Finance initiates the appropriate process.

Customer support takes over once the customer is onboarded. What we are building is therefore not a collection of isolated bots. It is an orchestrated digital workforce.

THE ECONOMICS OF THE AI-NATIVE COMPANY

Is this really about replacing employees?

It is about redesigning how work is allocated.

Some repetitive tasks will undoubtedly disappear. Pretending otherwise would not be credible.

But companies are not simply collections of repetitive tasks.

People create extraordinary value through judgement, relationships, leadership, negotiation, creativity, empathy and responsibility.

The problem is that we currently consume a huge amount of that human capacity on mechanical execution.

AI changes that. The objective is not to remove people from companies. It is to remove unnecessary manual execution from people.

What does that mean economically for an SME?

Operating leverage.

Historically, if you doubled customers, transactions or geographic coverage, you eventually needed more people.

AI begins to break that relationship. A company can increase operating capacity without increasing headcount proportionally.

That can mean faster response times, more leads followed up, more documents processed, more customers supported and more markets covered with the same human organisation.

This is particularly transformative for SMEs because large corporations could always spend millions building sophisticated automation. Most smaller businesses could not. AI employees can democratise that operating capacity.

Could five people eventually operate a company that traditionally required fifty?

In certain businesses and functions, absolutely. Not every business. You are not going to manufacture aircraft with five people because you have AI.

But in digital businesses, professional services, commercial operations, administration, marketing, customer communication, research, reporting and many document-intensive processes, company capability and company headcount will become increasingly disconnected.

That is one of the biggest changes I expect over the next decade. We will stop asking:

“How many people does that company employ?” and increasingly ask: “How much operating capacity does that company control?

And what about the famous “one-person company”? Could one entrepreneur genuinely operate a company with an AI workforce?

Yes. And I think we will see extraordinary examples.

A founder can increasingly orchestrate specialised employees across sales, CRM, marketing, support, administration, reporting and parts of finance and legal preparation.

But I would not define Evalon around the one-person company. That is the spectacular example.

The much larger transformation is the AI-native company. A business with 10, 50 or 500 human employees could operate with the capacity traditionally associated with a much larger organisation.

That is economically far more important.

TRUST: THE QUESTION EVERY CEO SHOULD ASK

Companies are experimenting with ChatGPT and dozens of AI tools. But many CEOs are uncomfortable putting sensitive company or customer data into them. Is that becoming one of the biggest barriers to enterprise AI?

Absolutely. The question for a CEO is no longer simply: “Can AI do this?” Increasingly, the question is:

“Can I allow AI to do this inside my company?” Those are two completely different questions.

A consumer AI tool can be extraordinary for individual productivity. But once AI starts accessing customer information, contracts, financial records, CRM data or internal documents – and especially once it starts taking actions – the requirements change completely.

You need to know where the information is, who can access it, what the AI is authorised to use, what actions it may take, what is recorded, what requires approval and who remains accountable.

That is why Evalon is not simply an interface giving employees access to AI.

It is a governed environment for deploying AI inside a company.

So is data protection actually part of the Evalon product rather than an additional compliance layer?

Exactly. And I think this distinction is extremely important. We do not believe companies should first deploy AI and then try to add governance afterwards.

Governance has to be part of the architecture. Evalon is designed as a controlled execution environment in which each organisation operates within its own governed context.

Access can be defined by role. Credentials are controlled. Permissions determine which systems and information an AI employee can access. Workflows determine what it can do. Approval rules determine what requires human intervention. Execution is logged so that actions can be traced and reviewed. Data environments can be separated. Information can be encrypted. And the organisation retains control over the boundaries within which each AI employee operates.

So the proposition is not: “Here is an AI model. Trust it with your company.”

It is: “Here is an operating environment in which you decide exactly how AI is allowed to work inside your company.”

That is a fundamentally different proposition.

Does that mean a company can use Evalon and automatically be GDPR or AI Act compliant?

No technology platform can make every company automatically compliant simply by switching it on.

Compliance depends on the company, the data, the purpose, the workflow, the sector and the specific AI use case.

What Evalon can do is provide the technical and governance infrastructure that allows companies to deploy AI in a controlled way and implement the safeguards required for compliant operation.

That distinction matters. GDPR, for example, is built around principles such as purpose limitation, data minimisation, security, transparency and accountability.

The European AI framework increasingly requires companies to think about risk, transparency, traceability, documentation and human oversight.

Those principles map naturally into the architecture we are building.

A company can define what information an AI employee needs for its role rather than exposing everything. It can control access. It can establish approval points. It can maintain records of execution. It can apply different governance standards to different workflows. And it can demonstrate much more clearly how AI is actually being used.

We do not replace the company’s legal responsibility. We give companies infrastructure that helps them exercise that responsibility properly. That is a much more serious way to think about enterprise AI.

What does “data protection by design” actually mean when an AI employee is doing real work?

It means starting from the opposite assumption to unlimited access. An AI employee should receive the minimum information and authority necessary to perform its role.

If it works in sales, it does not automatically need access to payroll. If it processes invoices, it does not automatically need access to HR files. If it prepares a document, it does not automatically receive permission to send or sign it.

Every role can have its own boundaries. That is important technologically, but it is also important from a data-protection perspective. The objective is to create an environment where access is intentional rather than accidental.

Right employee. Right data. Right permission. Right action.

And everything important should be traceable. That is what data protection by design means to us in operational terms.

One concern companies have is losing control of their proprietary information. Does using an AI employee mean giving away the company’s knowledge?

No. And this is one of the questions every CEO should ask before adopting any enterprise AI system.

A company’s customer information, commercial strategy, documents, internal processes and operating knowledge are corporate assets. The architecture therefore has to be designed around controlled access and separation.

An AI employee should operate against the information it has been authorised to use for a particular purpose – not treat the entire company as an unrestricted data pool.

That means separating customer environments, controlling credentials and permissions, defining which sources are available to each workflow and maintaining traceability over how those sources are used.

The fundamental principle is very simple: Your company data remains company data.

AI should create value from corporate information without turning corporate information into an uncontrolled asset.

Does all this governance make the AI slower? Companies also want speed.

That is precisely the engineering challenge. Enterprise AI cannot force companies to choose between speed and control. They need both.

Traditional corporate processes often become slow because controls are manual: someone checks an email, someone validates a document, someone copies information into another system, someone requests an approval and someone updates the record.

A properly designed AI workflow can automate much of that sequence while preserving the controls.

Rules can be evaluated instantly. Permissions can be checked automatically. Data can be retrieved immediately. Routine actions can execute without waiting. Only exceptions or higher-risk decisions need to move to a human.

So governance does not necessarily make execution slower. Good governance can make controlled execution faster. That is one of the major opportunities of enterprise AI.

Could Evalon therefore be particularly relevant for regulated businesses and professional firms?

Yes. Potentially very much so.

Professional-services firms, financial organisations, healthcare operators, legal businesses, insurance, real estate groups, institutions and other data-intensive organisations often have exactly the combination Evalon is designed to address: large amounts of information, repetitive processes, valuable professional time and significant requirements around control.

Historically, those organisations sometimes faced a trade-off. They wanted the productivity benefits of AI, but they could not simply allow uncontrolled systems to operate over sensitive information or business processes.

A governed execution layer changes that equation. The AI employee can be given a very precise operating perimeter. It can execute what is permitted. It can record what it has done. And it can escalate what requires professional judgement or formal authority.

That makes AI potentially more useful precisely in environments where trust matters most.

Could compliance itself become a competitive advantage for European AI companies?

I believe so. Europe is sometimes portrayed as being disadvantaged in AI because it regulates more.

But there is another way of looking at it. If artificial intelligence is going to become operational infrastructure inside banks, professional firms, healthcare companies, industrial groups and public institutions, trust will become enormously valuable.

Companies will increasingly ask: Where is my information? Who has access? What model is being used? What happened to my data? What did the AI do? Why was that action permitted? Can I audit it? Can I stop it? Can a human intervene?

A platform designed to answer those questions has a commercial advantage. So our ambition is not simply to comply with European expectations. It is to turn European-grade governance into a product advantage. Because the more powerful AI becomes, the more valuable trust becomes.

Is this ultimately the difference between experimenting with AI and actually running a company with AI?

Exactly. Experimentation is easy. You can open an AI application this afternoon and do something extraordinary with it. Enterprise deployment is different.

The moment AI starts performing real work, you need identity, permissions, data boundaries, security, monitoring, traceability, approval rules, exception handling and accountability.

That infrastructure is what allows a company to move from:

“Our employees are trying AI”

to:

“AI is now part of how our company operates.” And that transition is the market Evalon is building for.

WHY EVALON?

Javier, there are thousands of AI startups today. Why should an entrepreneur believe Evalon will still matter when the technology changes again next year?

Because Evalon is not built around one model. And neither is my professional experience.

I have spent approximately fourteen years working across data, machine learning and enterprise AI – designing architectures, building systems, taking them into production and leading multidisciplinary teams.

I have worked in industrial, automotive, retail, financial, healthcare, energy and public-sector environments.

At Gestamp, for example, the enterprise AI platform supported more than 50 active production models and hundreds of enterprise users across several countries. In another environment, I led a multidisciplinary team that transformed a pricing process that previously took approximately a month into a process completed in less than a day.

That experience taught me something important. The difficult part of enterprise AI is rarely producing an impressive answer. The difficult part is making technology operate reliably inside a real company. Evalon is the productisation of that experience.

So this company did not begin with the ChatGPT boom?

No. Evalon is being built during the generative-AI revolution, but the operating knowledge behind it predates that revolution by many years.

I have spent fourteen years seeing the same problems repeatedly: integration, fragmented data, permissions, workflows, governance, deployment, monitoring, economics and production reliability.

Foundation models have changed what is technically possible.

What Evalon does is combine that new intelligence with the enterprise operating disciplines required to turn possibility into work.

That is a very different starting point from building a wrapper around the latest model.

Evalon was founded by you together with Iberica Capital Partners. Why was that combination important?

Because building a technology product and building an international technology company are two different challenges.

Iberica Capital Partners (founded by Yves Horoit and P. A. Anderson), is Evalon’s institutional co-founder and strategic structuring and financial partner.

ICP brings a complementary corporate dimension to the company: strategic structuring, financial architecture, capital strategy, corporate governance, institutional relationships, market access and international business development.

Its role has been to help transform a powerful technological proposition into a properly structured company capable of attracting capital, entering strategic markets, developing institutional relationships and scaling internationally.

That creates an important balance.

Evalon combines deep technical leadership and product ownership with sophisticated corporate structuring, financial discipline and international market orientation.

The technology has to work. But to build a major international company, technology alone is not enough.

WHY NOT CHATGPT, COPILOT OR AN AUTOMATION CONSULTANT?

Every major technology company is now building agents. Why does the world need Evalon?

Because businesses do not buy agents. They buy outcomes. The major AI laboratories are building extraordinary intelligence.

Enterprise software companies are embedding AI into their ecosystems. Automation platforms are excellent at connecting tasks. Consultants can build bespoke workflows. All of those have value.

But the problem we are solving sits above those individual components.

Evalon packages intelligence, workflows, integrations, permissions, state, exception handling, traceability and human oversight into operational roles.

We are model-agnostic because the underlying models will keep changing. Our value sits in turning intelligence into reliable business execution.

Models will change. Businesses will still need the work done.

25. Why not simply become a highly profitable AI consultancy?

Because consulting scales people. Products scale repeatability.

A consultant can create an excellent automation for one customer. But if every new customer requires another bespoke project, you have built a services business.

Our objective is different. We are productising recurring business roles.

Reusable AI employees.
Reusable workflows.
Reusable integrations.
Reusable governance.
Repeatable onboarding.
Predictable recurring pricing.

Configuration will always exist because companies are different. But the underlying operating system should become increasingly standardised. That is how you go from deploying AI projects to building infrastructure.

WHAT DOES SUCCESS LOOK LIKE?

How should a CEO measure whether an AI employee actually works?

Exactly as they would measure any other operating capacity.

What did it accomplish?
How accurately?
How quickly?
At what cost?
And what happened to the business as a result?

For sales, that might mean opportunities created or revenue. For support, response time and resolution.

For administration, hours released. For finance, processing time and exceptions. For document workflows, throughput and accuracy. AI should not be judged by how impressive the demo looks. It should be judged by what changed inside the business.

What does Evalon still need to prove?

Repeatability at scale.

We already have the technology, workflows, deployments and operating evidence. Now the challenge is turning that into a highly repeatable commercial machine. Customers should be able to activate quickly. They should see measurable value. They should remain.

And, most importantly, once the first employee works, they should want the second, third and tenth.

That is when the economics become particularly powerful. The strongest proof of an AI employee is not a futuristic video. It is a customer saying:

“This is now simply part of how our company operates.”

THE COMPANY OF TOMORROW

What will the organisational chart of a normal company look like five years from now?

I think it will contain two kinds of operating capacity: human employees and AI employees.

A sales director may manage people together with AI employees responsible for prospecting, CRM administration and follow-up.

A CFO may have AI employees continuously preparing reconciliations, reporting and monitoring.

Operations may have AI employees managing routine processes and escalating exceptions.

The manager therefore changes too. Less time chasing execution.

More time defining objectives, establishing authority, reviewing exceptions and making decisions.

AI will become less like software you open and more like capacity you manage.

Could AI employees eventually communicate directly with AI employees in other companies?

Yes. And that could become one of the most interesting developments in the digital economy.

Imagine your procurement employee communicating with a supplier’s commercial employee.

Documentation is exchanged. Requirements are verified. Availability is checked. Information is updated.

Routine terms are processed.

Humans enter when genuine judgement or negotiation is required. A large proportion of B2B activity is structured information exchange. Some of that will inevitably become machine-to-machine. The critical questions will be identity, authority, security and accountability.

What is the ultimate ambition for Evalon?

To become the execution layer of the AI-native company.

Companies already have systems of record that store information. They increasingly have systems of intelligence that interpret information. The next major layer will be the system that acts on that information.

If we succeed, companies will eventually stop thinking of AI employees as something futuristic. They will simply become part of the organisation. That is when the transformation becomes profound. Software will no longer be only something a company uses.

Software will become part of the workforce through which the company operates.

THE QUESTION EVERY ENTREPRENEUR WILL ASK

Javier, if I own a company and I am reading this interview today, what should I do tomorrow morning?

Do not start by asking: “How can I use AI in my company?” That question is too broad. Ask instead:

“What recurring job in my company consumes time every single day and has a measurable result?”

Maybe it is following up every sales opportunity.

Maybe it is keeping the CRM updated.

Maybe it is customer support.

Maybe it is processing documents.

Maybe it is reporting.

Maybe it is administrative coordination.

Start with that job.

Give it clearly defined authority.

Measure what happens before and after.

And if the first AI employee earns your trust, hire the next one.

That is how the transformation will happen.

Not through a gigantic AI project.

One employee becomes a team.
A team becomes a digital workforce.
And a digital workforce changes the economics of the company.

“The future of enterprise AI may not be another piece of software. It may be a new kind of workforce. And for thousands of businesses, the first question may soon be surprisingly simple: what should my first AI employee do?”

Billions Magazine is a media partner of EVALON.