"Revolutionary" is ambiguous here, depending on the context; it could mean either backward-looking (the bourgeois/capitalist revolution overthrowing feudalism) or forward-looking (the socialist revolution overthrowing capitalism) or both.
To make things a little more concrete, consider Adam Smith's discussion of a pin factory. In a pre-industrial economy, each pin is made by a single skilled individual. (Smith gives a list of steps that's something like drawing out the wire, measuring and cutting it, pointing one end and capping the other.) Putting a bunch of pin-makers in a single room together doesn't really make them more efficient or productive, because they're each still working individually, one pin at a time.
A factory, by contrast, is organized in what Henry Ford dubbed an assembly line. Each worker is responsible for exactly one step of the process. This (according to Smith) allows for more efficiency and increases production. (For some critical discussion and context, see here.) In the nineteenth century, assembly lines made effective use of steam engines (cf https://en.wikipedia.org/wiki/Cotton_mill#/media/File:Cotton_mill.jpg). This was a revolutionary increase in production relative to the old feudal system of piece-work manufacturing.
That was the backward-looking sense of "revolutionary." For the forward-looking sense, consider the social requirements for organizing a factory. An individual craftsperson doesn't need to coordinate their work with anyone else — they work at their own speed, according to their own schedule (more or less). But a factory run as an assembly line requires workers to be coordinated. Each worker needs to work at a certain speed for the assembly line to run smoothly, and workers need to start and stop their work according to strict schedules. Zooming out, whole industries need to be coordinated in the same way — raw materials have to be extracted, finished goods have to be sold to consumers, and everything needs to be transported in between. At the highest level, the entire global economy should be organized to run smoothly. So capitalism requires an elaborate infrastructure for coordinating and organizing productive activity and the transportation of goods, linking the work of each individual worker to the entire global economy.
But who controls this infrastructure? Clearly it would be developed by capital, over time, as ever-greater efficiency and productivity is sought for the sake of ever-greater profits. But, once developed, there's no reason why this infrastructure can't be controlled by the workers, collectively, for the good of all. Furthermore, the workers can use this infrastructure to coordinate their own political activities. (Think of the way socialists used the printing press, cheap newsprint, the mail system, telegraph, radio, and so on to spread revolutionary ideas.) Here's how Robert Paul Wolff put it:
Marx expected, for sound reasons, that the technology of production,
communication, and management required for the central planning and
control of an entire economy would develop first within capitalist
firms, in direct response to the pressures of competition and the
demands of profitability. And so they have. An immediate consequence
of this process is the transformation of economic calculations into
political decisions within the firm. Thus, if by socialism we mean
the rationally coordinated planning of an entire national economy in
such a way as to transform the major economic choices of the society
into political choices, responsive to the will of the people, then it
is true that socialism has been growing within the womb of capitalism,
or at least that the technical pre-conditions for socialism can be
seen to be developing there.
So the association (productive organization) of workers is also revolutionary in this forward-looking sense of creating a critical piece of infrastructure for a socialist society.
Is the forward-looking sense what Marx and Engels had in mind when they wrote the Manifesto? That's probably a stretch: they were only about 30 and 28, and industrialization was still a very new process. It's somewhat more plausible as an interpretation of Marx's Contribution to the Critique of Political Economy, which was written 11 years later. (Wolff's focus in his essay is on the Contribution.)