Madrid · Ciudad de México · Buenos Aires
ENESCareers
AnálisisArtificial IntelligenceIntermediate level

Are we configuring software in a way that will soon stop making sense?

If AI agents can understand the business and work on it, how much should we keep investing in configuring systems the way we always have?

I have spent the last few days working on how to organise our sales and marketing processes. Databases, funnels, fields, automations, tools...

And one question keeps coming back to me: how much of all this will still make sense three years from now?

From the best CRM to a "customer relationship harness"

A few days ago, HubSpot announced that its founder, Brian Halligan, is returning as CEO. But what is truly interesting is his intention to refound the company around a new concept: the customer relationship harness.

It is no longer simply about building the best CRM. It is about creating an environment in which AI agents can understand the business and work on it.

And I think the implications are huge.

Until now, we were the ones explaining everything to the software

Until now, we have designed software so that people explain to the computer what it has to do.

We create fields to store information. We build tables to relate it. We design workflows to automate tasks. And then we train employees to learn how to use all of it.

What if the next step were exactly the other way around?

An example: from configuring a system to explaining a goal

Imagine I want to identify companies that have just hired a sales director, research their expansion plans and prepare opportunities for my sales team.

Today I can do it with Clay, a CRM, data providers and a series of automations.

But will I need to configure that whole system?

Or will it be enough to explain to an agent what I want to achieve, give it access to the right sources and set some rules of engagement?

Does it still make sense to structure so much information by hand?

I even wonder whether it makes sense to keep manually structuring so much information.

A conversation with a customer contains far more context than the twenty fields we fill in afterwards in the CRM. AI can interpret that conversation, extract the data it needs and keep the system up to date.

This does not mean databases will disappear. We will still need reliable information, relationships between records, controls and traceability.

But perhaps people will no longer have to worry about how much of that information is stored, related and processed.

The question for those of us who run companies

And this is where a question arises that I find especially relevant for those of us who run companies.

For years we have bought software and hired specialists to configure it, integrate it and maintain it.

Now we have to ask ourselves how much of that investment will keep adding value when agents can do a growing share of that work.

I don't think tools like Clay are going to disappear. In fact, they are evolving quickly towards this very model.

But I do think we are approaching an important shift:

Moving from companies that have to adapt to software to software that adapts to the way companies work.

My question is no longer which tool I should add to our tech stack.

It is how much we should keep investing in configuring systems the traditional way.

Are we building the infrastructure of the future, or perfecting the infrastructure of the past?

Igor Fernández, CEO of Xternus

Frequently Asked Questions (FAQs)

What is HubSpot's customer relationship harness?

It is the concept HubSpot has used to frame its refounding after Brian Halligan returned as CEO in November 2026. The idea is to move from a CRM where people record and manage information to an environment where AI agents can understand the business and work on it. The company has not yet detailed its specific capabilities.

Will AI agents replace CRMs and databases?

Not necessarily. Companies will still need reliable information, relationships between records, controls and traceability. What may change is who does the work of capturing, structuring and maintaining that data: increasingly, AI agents able to interpret conversations, emails and other sources and keep the system up to date.

Should companies keep investing in configuring their CRM and automations?

It is worth reviewing where the effort goes. Investing in quality data, clear business rules and well-governed access makes sense, because that is the foundation agents will work on. Over-building fields, workflows and customisations that could soon be handled by explaining to an agent what you want to achieve makes less sense.

Artificial IntelligenceDigital TransformationAutomationMarket Trends & Insights