AI Is Rewriting the Rules of Political Office Work

The Quiet Shift in How We Work

Recently, my social feed has been buzzing with talk about Markdown and HTML. It’s not the usual tech chatter. Three months ago, Claude Code engineer Thariq argued that HTML is a better output format for AI than Markdown. The post sparked a debate, and now HTML is back in the spotlight for AI products.

Andrej Karpathy, a prominent AI figure, also noted that human-computer interaction is forming a new input-output paradigm. Humans are better at expressing needs through voice, while AI excels at presenting complex information visually. This idea is starting to resonate in everyday office work, especially in politics.

Political offices, whether on the campaign trail or in government, are drowning in documents—policy briefs, press releases, budget spreadsheets, and meeting notes. The shift from text to visual interfaces is not just a tech trend; it’s becoming a necessity. But generating these new formats is only half the battle. Editing, collaborating, publishing, and syncing data still keep most people locked out of the web page era.

From Static Documents to Living Files

WorkBuddy’s recent update to its knowledge base attempts to bridge this gap. It’s bringing AI-native formats like HTML and smart Markdown into the fold of office tools. Previously, WorkBuddy introduced a “human-machine co-writing” feature that let AI work inside Word, Excel, and PowerPoint. Now, it’s expanding to HTML and Markdown, with a “light app” that gives HTML data storage and multi-device sync.

The first phase was about getting AI into Office. The second phase is about making the files AI generates actually useful in our daily workflows. For political teams, this could mean the difference between a static PDF and a living document that updates itself.

Let me walk you through how it works. The knowledge base is split into “My Documents” and “Team Space.” The former is for personal files, the latter for collaboration. You can save a document, Markdown, HTML, or CSV after a task. The next time you work, you can point an agent to that file, and it will read and modify it. When a team needs to maintain something, you move it to Team Space and set permissions.

This turns the end of one AI task into the start of the next. The knowledge base isn’t just storing final files; it’s preserving context for future work, both for humans and agents.

A Test: Building a Project Dashboard

To see if this works in real life, I ran a quick test. I input data for five projects—a car launch, a website overhaul, a store expansion, and others—with details like owner, department, status, budget, and spend. WorkBuddy organized this into a structured data table, which I could filter and sort. The generated HTML looked like a lightweight project management dashboard.

But here’s the kicker: I could link this dashboard to a CSV file and a related spreadsheet. Then I typed a prompt: “Change the dashboard to read data dynamically from the linked table.” After that, the dashboard pulled data from the CSV. When I updated the spreadsheet, the dashboard refreshed. No need to regenerate the HTML. This front-end/back-end separation is powerful for political operations where budgets and schedules change constantly.

Of course, the first draft isn’t always perfect. AI-generated HTML might not be designer-level, but that’s not the point. The real test is editing. With WorkBuddy, I could select a section—say, the risk module—and ask the AI to lighten the background and add a note: “Need to hold a special meeting by Friday to resolve this.” The change only affected that selection, leaving the rest intact. I didn’t have to touch the code.

For non-programmers, this is a game-changer. Previously, editing a web page meant regenerating it or diving into code. Now, it’s like editing a document. This is crucial for political staff who need to tweak a briefing or a public-facing page quickly.

Publishing Without the Headaches

After editing, the next hurdle is sharing. For coders, there’s deployment, servers, and domains. For a staffer preparing tomorrow’s briefing, that’s overkill. WorkBuddy lets you publish HTML directly as an online page. You get a link, and colleagues can open it on their phones—no downloads. They can even comment on specific words, just like annotating a document.

This opens up new possibilities. A marketing plan for a new policy could have an interactive timeline. An operations report could let managers filter data by department. A product proposal could include videos and interactive demos. And if changes come up mid-meeting, you can select text and ask the AI to adjust it on the spot.

Markdown and HTML: A Natural Division of Labor

WorkBuddy also added an AI-native review mode for smart Markdown. When a team writes weekly reports, PRDs, or proposals, they can focus on content in Markdown. AI suggestions appear as tracked changes, and you approve them one by one. Once the text is final, you convert it to HTML for a polished presentation.

This creates a natural split: Markdown for lightweight collaboration and content maintenance; HTML for visual display and presentations, also collaborative. In a political context, this could mean drafting a policy brief in Markdown, then converting it to an interactive page for public consultation or a stakeholder meeting.

From a One-Person Workbench to a Collaborative Space

But the knowledge base upgrade goes further. The “light app” feature lets web pages save and sync data. I tested it with a bakery’s inventory CSV—product names, SKUs, stock levels, costs. WorkBuddy generated an inventory dashboard that flagged low-stock items and allowed category filtering. The underlying CSV remained the source, so it wasn’t a one-off report; it was a mini workbench.

For personal use, I created a “life battery” dashboard from a CSV tracking my daily energy. But the possibilities are endless. Imagine a campaign team tracking volunteer hours, donation amounts, or event RSVPs in a CSV, with a dashboard that updates in real time. Or a legislative office managing constituent cases with a similar setup.

One real-world example from WorkBuddy’s official materials is a bakery called Sun Beibei. They stored monthly cost reports by business theme in the knowledge base. The next month, when doing analysis, they could pull historical reports and have the AI build on previous conclusions, then save the new results back to the same directory.

This adds a layer that traditional folders don’t have. It’s not just archiving; it’s preserving context for future work. For political teams, this could mean accumulating institutional memory—every policy analysis, every campaign strategy, every constituent interaction—so that AI can help with the next iteration.

Why This Matters for Politics

In politics, information is power. But it’s also a burden. Campaigns generate terabytes of data—polling, donor lists, social media sentiment, volunteer schedules. Governments produce endless reports, memos, and briefings. The ability to turn static files into dynamic, collaborative tools is not just a convenience; it’s a strategic advantage.

Consider a city council member who needs to track budget amendments. With WorkBuddy, they could have a live dashboard that updates as council members submit changes. Or a congressional office managing constituent cases, with a shared database that agents can query and update.

The shift from text to visual interfaces, from static documents to living files, is redefining how political work gets done. It’s not about replacing Word, Excel, or PowerPoint—they’re still there. It’s about adding a new set of tools that are native to the AI age.

In the end, the file format is more than a suffix. It dictates how we produce, edit, and share information. In the past, Office defined how people worked in the computer age. Now, AI is starting to redefine what files should look like—and what our workday should feel like. For political professionals, embracing this change could mean the difference between a campaign that’s stuck in the past and one that’s ready for the future.