How a few plugins shook the tech world
We may think software companies represent the future of business. For years, software-as-a-service (SaaS) has been one of the great places to make money in the stock market. However, at the beginning of the month, we were reminded that every industry faces creative destruction after Anthropic’s Claude Cowork was released. The question with Cowork is not whether disruption will happen, but how quickly.
Claude Cowork is an AI agent that runs on your computer and does actual work rather than just answering questions or creating lewd pictures of Sir Keir Starmer. Released with plug-ins in late January 2026, it sits in your desktop application as a separate tab from regular chat. You point it at a folder, describe what needs doing, and it executes the task autonomously while you make a cup of tea.
The difference between Claude and regular AI chat is stark. Traditional AI responds with suggestions you type manually. Cowork reads files, creates spreadsheets with working formulas, generates presentations, analyses data, reorganises folders, and produces finished documents without constant supervision. It plans, executes, and coordinates multiple sub-tasks in parallel. Early users report using it for expense reports from receipt screenshots, assembling first drafts from scattered notes, and file organisation that would normally take hours. The experience resembles leaving work to a competent junior colleague who does not complain.
This capability extends beyond document work. Through plugins released this January, Cowork can now automate legal contract reviews, sales prospect research, marketing workflows, compliance checks, and data analysis. These plugins represent Anthropic moving from foundation models into the application layer, competing directly with specialised software vendors rather than powering them.
Claude Code was originally released as a developer tool in February 2025, but users found it could do a lot more than write code – it could handle documents and files. When Anthropic realised this, they repackaged Claude Code as Claude Coworker after just 11 days of development; by contrast, OpenAI’s ChatGPT took more than four years to develop, including the GPT-1, GPT-2 and GPT-3 iterations.
Markets responded with panic selling that traders immediately dubbed the “SaaSpocalypse”. Legal and publishing stocks bore the initial brunt. Thomson Reuters fell 16%, RELX (owner of LexisNexis) dropped 14%, the London Stock Exchange Group declined 13%, and analysis companies like Gartner plummeted 21%. Broader software indices followed, with the S&P North American Software Index dropping 15% in January, its worst monthly decline since October 2008. An estimated $285bn in market value disappeared in a single trading session.
The selling spread beyond obvious targets. Indian IT services giants like Infosys and Wipro saw their American-listed stocks fall 5% on concerns that AI threatens the traditional billing model. Microsoft, despite solid earnings, dropped 10% as investors expressed concern about its Productivity and Business Processes division. This division includes its high-margin Microsoft 365 and Dynamics software business, which generates over 50% of its profit today.
This reaction appears indiscriminate; Microsoft is arguably one of the best-positioned companies to benefit from this shift rather than suffer from it. Microsoft has already embedded AI into the core workflow of knowledge workers — and is charging for it at multiple levels (take GitHub Copilot, Microsoft 365 Copilot and Teams Copilot as cases in point). In our own operations, we have realised very significant productivity gains across the various iterations of Copilot, even as Microsoft monetises usage through layered subscriptions and additional token consumption. In other words, the same dynamic that threatens narrow SaaS vendors could expand Microsoft’s revenue opportunity. Yet the market initially treated it as a casualty rather than a beneficiary.
Software companies with a specialised, narrow business, relying on user subscription models, face immediate pressure if AI automates tasks previously requiring multiple users and revenue shrinks. Companies like specialised SaaS tools performing boring, rules-based work – such as contract management, expense processing, compliance checking and data formatting – appear particularly vulnerable. Professional services face similar threats. The traditional full-time equivalent billing model breaks down when AI can execute complex multi-step tasks. Clients could increasingly pay for outcomes rather than hours worked.
The speed of disruption for software providers remains uncertain, but the direction is clear. If a share titan like Microsoft is falling, then what chance do smaller business process companies have? Companies like DocuSign (e-signatures), Salesforce (customer management), Adobe (document workflow) or Atlassian (project management). That said, not all software faces equal risk. Companies with proprietary data, controlling both distribution and infrastructure, with strong customer relationships or unique positions, may be able to absorb AI agents as features rather than be displaced by them. Indeed, if customers are already seeing measurable productivity improvements, willingness to pay can increase even if pricing complexity rises. The risk therefore lies less with diversified platforms like Microsoft and more with single-function tools whose value proposition can be automated away.
Claude Cowork is a classic case of innovation causing Austrian economist Joseph Schumpeter’s “creative destruction”. As old, redundant technologies disappear and free up capital, others will emerge in unimaginable ways. For once, the Chinese saying “may you live in interesting times” is an invitation rather than a curse.
Please be aware that the value of investments may go up or down and you may receive back less than you invested originally. Past performance is not a guide to the future.
This document contains general information only and does not provide any advice or guidance specific to your personal circumstances.
Callanish Capital is authorised and regulated by the Financial Conduct Authority to manage investments and authorised to provide advice on investments, pensions, credit, protection and mortgages.
This email and any accompanying attachment are issued by Callanish Capital Ltd, a company registered in England and Wales under company number 13182424. Registered Office: 45 Pont Street, London, SW1X 0BD.
Callanish Capital Ltd is a Discretionary Fund Manager, directly authorised and regulated by the Financial Conduct Authority (“FCA”) under registration number 955992.