Media Operating System

After seeing the trends in generative ai, workflows, browser/computer use, and data services, it seems we're at a tipping point to redesign some job/department processes from the ground up. In looking back to my earlier career in brand communications, there was a LOT of tedious tasks.

Some things were fun, but often I had so much on my plate as a content strategist, manager, and UX designer that it felt unfortunate some job skills like analytics could never be better explored & improved.

Because:
1) There was no additional staff budget for valuable but low frequency tasks
2) Some tasks (esp. creative) took extra monotonous time to perform since you'd have to click around massive programs to GTD, copy/paste, save/download/upload file here/there.

We've all been there and just dealt with it when there was no other choice, and were happy if we got to use a new unified software that improved turnaround time and reduced annoying task switching and cobbling.

In learning fullstack SWE after 15 years of light frontend work, I got to glimpse how OSS, APIs, and capable self-reliance could allow me to achieve more, albeit as a generalist web developer. Deep down though, I really rather wanted to improve marketing and stakeholder communications.

While it's possible to do both at large brands, of course you'd be limited to the confines and boundaries of what software was used and whether it was extendable. The arrival of simpler software like Canva, Resend, Dub.co, Plausible Analytics, and platforms like Sanity made ops easier.

And while these products have appropriately adapted to the current times by making their services available via CLI or MCP, I have to imagine how departmental systems can become even more automated so people can do more review, guidance, and analysis over boring UI clicking.

To this end, maybe I'm wrong but it seems like for marketing (strategy, customer service, promotion, product) there'll be a "media operating system" that'll include:

A knowledge graph, asset graph, agent team, human team, distribution layer, and analytics layer. It'll take some time to setup and in some cases it'll probably make sense to invest in creating custom software over designing RPA, browser/computer use agents that then need wiring and adjustments.

It's an early thought process for me, but I think there's a lot of exciting opportunities ahead for engineers embedded into functional departments to see how automated, ai-assisted workflows could at least improve both outcomes and yield higher job satisfaction & quality-of-work. Heck, that’s what “forward-deployed engineers” are doing as vendor and agency consultants today.