What changes when AI gains tools to act

The opening of Web Summit Lisbon on the evening of November 9 brought together subjects that might seem far apart. Agents capable of carrying out tasks, government transformation, Chinese companies, online commerce, and humanoid robots shared the stage.

For me, there was a connection between these conversations. Artificial intelligence is gaining access to the places where work happens. It can consult information, operate tools, communicate with other people, and, within defined limits, take action.

This changes the question for anyone building a company. Alongside choosing a model, we need to understand which work can be delegated, how to monitor its execution, and what makes the result trustworthy.

When an agent gets an email address

The participation of Tao Zhang, cofounder of Manus, brought this shift into an everyday situation. Someone starts the day with several outstanding tasks and needs to turn each one into organized work.

Delegation requires gathering context, clarifying the expected outcome, identifying dependencies, and following what happens next. When an agent handles part of this, the potential benefit extends beyond the time needed to write a message.

Manus is exploring this direction with Cue, a personal-agent app introduced alongside Manus 2.0. According to the company, agents receive their own identity, with an email address, phone number, wallet, and computer, and can make payments within a budget set by the user. Official introduction to Manus 2.0 and Cue.

These capabilities bring agents closer to the tools we use to work and buy things. They also make it more important for companies to answer certain questions before expanding autonomy.

Who authorized a purchase? Which information can be shared? When does the agent need to ask for help? How can an action that produced an unexpected result be corrected?

In practice, trust becomes part of the product. Users need to understand what they delegated, follow progress, and be able to intervene.

Simplify the process before adding AI

One of the opening's most applicable messages came from the discussion about government. Portugal's minister for state reform, Gonçalo Matias, placed process simplification before digitization.

That order also matters inside a company.

Imagine a registration process that asks for the same information across three departments. Each checks a different version, and someone spends the day forwarding documents between them. An agent could speed up that forwarding. But we can also ask why the customer needs to repeat the registration.

When designing a solution, it is worth reconstructing the entire journey. What triggers the request? Which information already exists? Who can decide? Which steps serve a real need, and which have remained out of habit?

The proposal for agents coordinating public services, such as those related to the birth of a child or starting a business, points to an experience organized around a person's needs.

For startups, I see a similar opportunity. Understanding a process deeply makes it possible to find steps that can be combined, anticipated, or eliminated. That knowledge helps turn technical capability into a delivery the customer notices.

The next customer may arrive with an agent

In the discussion with Matthew Prince of Cloudflare, Anton Osika of Lovable, and Richard Socher of Recursive, another consequence of this transition emerged. The internet is also being used by agents that seek information and carry out tasks on people's behalf. Opening participants.

Prince raised a concern about how this movement could affect company discovery and favor established businesses.

My reading is that this deserves attention from any founder who depends on digital acquisition.

Today, a company tries to appear in search, build a reputation, and persuade someone to visit its website. When users delegate research or a purchase, part of that journey goes through a system that compares options according to the instructions it received.

Price, availability, commercial terms, and evidence about the product need to be accessible and consistent. The ability to integrate services also becomes more important.

For a smaller business, this could open new ways to be found. It could also increase dependence on whoever decides which options are presented to the customer.

I would start by examining a concrete journey. If someone asked an agent to find a solution like yours, which information would be available? What could it verify? At what point would it need to involve a person?

China broadens the references available to builders

Paddy Cosgrave highlighted the role of Chinese companies in the future of technology. That presence was also visible in the program, with Manus, robotics companies, and other businesses.

For entrepreneurs in Brazil, I see value in broadening our range of references.

There are different lessons in building models, integrating software and hardware, distribution, and the pursuit of efficiency. Observing how a company works under certain constraints can help us formulate better questions about our own business.

The exercise requires context. A result presented on stage depends on the task, environment, and test conditions. When analyzing a case, I want to understand how much it costs to deliver the required quality and how much effort remains with the team.

That is a more useful comparison for making decisions than looking at a model's ranking position in isolation.

Robots bring the discussion into physical operations

The opening also made room for humanoids and a proposed vehicle with a modular aircraft. These presentations make visible an ambition running through the event: taking artificial intelligence beyond digital interfaces.

In the physical world, delivery depends on a broader combination of factors. A robot needs to perceive its environment, manipulate objects, and perform movements. It also needs to operate safely, receive maintenance, and function for as long as required.

To evaluate an opportunity, I would look for a clearly defined task. Where is there repetitive, physically demanding work, or work that is difficult to staff? How is it done today? Is the environment predictable? How much human effort will be needed to supervise the machine?

A demonstration helps explain a capability. Everyday operations reveal what will be needed to turn it into a business.

There is room for companies building applications, integration, maintenance, training, and monitoring. Those who understand a specific operation can find opportunities in the transition from available technology to everyday use.

How we assess people is also starting to change

Tao Zhang described evaluating candidates by their ability to solve problems working with AI. That brings the discussion about agents into hiring.

As tools perform part of the implementation, it becomes more important to observe how someone defines the problem, supplies context, judges alternatives, and checks the result.

A seemingly finished delivery may conceal weak decisions. Following the work process helps reveal what the person noticed, what they questioned, and how they reacted when the first attempt failed.

I see a skill in this discussion that we can already develop within teams. Learning to delegate clearly and review with sound judgment.

Where I would look for value after the opening

I came away from these conversations thinking about who can connect a need to a complete delivery.

It might be a company that reduces the effort needed to start a business, a tool that prepares a task with enough context for someone else to carry it out, or a solution that addresses a physical problem in a specific environment.

In all these cases, certain questions help guide the first project. Who experiences the problem? How much does solving it cost today? What needs to change in the process? How will the customer recognize that the solution worked?

That is where I want to focus my attention over the next few days of Web Summit. On startups that can explain the work they are taking on, the conditions needed to do it well, and the evidence that it improves someone's life.