Can one developer realistically improve a small WordPress agency?

Yes, but don't treat it as one giant transformation. Reporting, WordPress workflows, hosting and finance are separate projects with different users and constraints.

Pick one repetitive task, find out how much time it takes, and look for an existing tool before building anything. Try a small change, see whether the team actually uses it, then decide what comes next. A useful reporting improvement might have more immediate value than redesigning the whole hosting setup.

Get agreement on the time you're going to spend. Infrastructure cleanup can take a lot of work without producing something clients or coworkers can immediately see.

Talk through the problem, the expected benefit and what other work might be delayed. That gives you a better shot at keeping the time available once the first urgent client request comes in.

I'd standardize what you can before building a custom cloud platform. Work out what access, backups and maintenance each site needs, then compare that with the hosting tools already available to you. Missing SSH access may limit some workflows, so check that early rather than designing around access you don't have.

For WordPress, reusable workflows and existing plugins are a reasonable starting point. Learn PHP as you tackle specific gaps, and only write a custom tool when the available options genuinely don't fit.

This could be a good way to rebuild your development skills while solving real business problems. Just leave room to learn the PHP and WordPress parts instead of relying on generated code to do all the thinking. You don't need to become an expert before making the first improvement.

A small team can make it easier to change a process without getting approval from layers of management. That said, people may not like every change, and one developer can't fix everything at once. Set a modest scope and expect some adjustment along the way.