Conciseness is essential in the legal world, and few know that better than an assistant who has spent two years at the firm of Jorge Morell, which advises 400 tech companies in Spain and Latin America. His name is Claude, and he is an AI chatbot created by former ChatGPT developers under the start‑up Anthropic. “It writes precisely and with few words, while other AIs embellish sentences and create confusion,” says the 43‑year‑old lawyer from Mallorca in Spain. In five minutes, Claude can draft the infamous “terms and conditions” — those 10‑page documents that flood the internet even though no one reads them. “A person would need two hours,” Morell adds.
The chatbot’s French name has been everywhere in recent days. Not only because it advertised during last Sunday’s coveted Super Bowl slot, but because an update that improved its legal‑assistance capabilities triggered an extraordinary shockwave in the markets: $1 trillion in market value evaporated from the U.S. software sector; the mighty Adobe lost 15%. Claude, however, began long before that.
Anthropic’s chatbot was launched in March 2023, and for years it has been familiar to Morell and to many entrepreneurs across sectors thanks to its focus on practical applications. But for the stock markets, this was a revelatory moment about a branch of AI far less known than large language models and far less ambitious than artificial general intelligence — yet with a clear purpose and a concrete use.
What is Claude?
On the surface, the chatbot resembles any other model: it responds conversationally to simple natural‑language instructions or to text, audio, or video files. Underneath, it is different: instead of selling an agent to the general public, Anthropic has trained its model to perform specific tasks for companies. It offers dedicated functions for sales management, advertising, customer service, and, among others, legal assistance.
Claude also has an Excel add‑on that allows companies to create, analyze, correct, or automatically update their spreadsheets. If Microsoft’s program reduced the need for programming knowledge, Anthropic’s tool has done the same for spreadsheets. For example, Claude can instantly update a company’s entire financial plan based on potential interest‑rate hikes or changes in input prices, simply by receiving an instruction in plain language. Or it can create a sheet from scratch using publicly available data.
“You notice the advantages in the most typical day‑to‑day functions of a company, like writing, editing, or programming,” says Morell. In fact, beyond its legal or financial applications, Claude also includes specialized models to assist with computer programming: one for expert developers, called Code, and a simplified version that doesn’t require writing a single line of code, Cowork — another of January’s new features.
What is it used for?
For the past year, Spanish entrepreneurs Manuel López Rivero, 54, and Manuel López Aragón, 18, have been using Claude — specifically its Code version. Father and son, the two Manuels run a company in Santa Cruz de la Palma, in Spain’s Canary Islands, that develops applications for businesses.
“Before, we might spend years and thousands of euros on developers, but now we can get it done in about three months,” the son says. In that same timeframe, the duo has just launched a platform to find and hire staff for event organization.
Anthropic’s chatbot also allows entrepreneurs to react quickly. When the European Commission approved a set of proposed amendments to its digital legislation late last year, Morell created an interactive calculator to help entrepreneurs understand the new rules.
“It was a way to simplify the understanding of a somewhat complex regulation,” explains Morell. With the tool, integrated into his firm’s website, a company developing a high‑risk AI system — such as those used in biotech — could simulate when certain articles of the regulation would begin to apply to them.
How popular is Claude?
As OpenAI CEO Sam Altman said in a message to employees, “ChatGPT is the AI for most people,” but entrepreneurs are not part of that group. In the business world, the landscape is completely different. Anthropic — which plans to go public this year — has raised more than $60 billion in funding rounds over its trajectory, backed by giants like Google, and has reached a valuation of $380 billion. That’s almost a third of OpenAI’s, yet still an enormous figure.
With three paid subscription plans and one free tier — the latter with daily usage limits — Claude is trained using a method in which its developers define the principles that should guide the system as if they were a constitution, rather than relying solely on user interactions.
This approach leads the model to prioritize clear, direct answers instead of simply trying to please, a tendency that often results in misunderstandings. “Claude rambles less, and that matters in business,” says Morell.
What is the future of software?
Claude’s emergence beyond the business world comes at a time of growing turbulence in the markets, which have become increasingly volatile as investors grow impatient to see how tech companies will profit from a technology whose costs keep rising. In this context, Anthropic has an advantage and, unlike its competitors, is already projecting profitability — starting in 2028. Its revenues surpassed $9 billion in 2025, half of OpenAI’s but more than double what had been forecast at the end of the first half of the year. Growth is accelerating. According to an HSBC survey from late last year, Anthropic accounts for 40% of corporate AI spending, compared with OpenAI’s 30%.
On the stock markets, major data‑analysis software companies — including S&P Global and Moody’s, two of the world’s leading credit‑rating agencies — have fallen 15%. FactSet, another giant in business data analytics, has plunged nearly 30%.
Despite the hit, analysts maintain that the sector will adapt and hold up. “We expect some short‑term relief after the recent wave of selling, but dispersion among companies is likely to increase depending on how much AI can disrupt their business models,” Citi said in a report last week.
J.P. Morgan strikes a similar tone: “Although the rapid evolution of AI is a legitimate concern for the software industry, current investor sentiment may be overly pessimistic. Despite weak sector fundamentals, limited insider buying and stagnant hiring, our CIO survey work does not suggest the imminent death of the broader software landscape.” They already know who they can turn to for reviewing their plans.
Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition
