The First Draft Doesn't Have to Be Perfect

In 2020, I was on the communications team for the COVID-19 response at the Colorado Department of Public Health and Environment. Our data teams were building charts and visualizations on critical COVID data, sometimes hours or just minutes before the Governor's press conferences.

The data was always there. What wasn't always there was a consistent, trustworthy way to present it to a general public audience, fast, under pressure, with no room for error.

So I built a data visualization style guide. A set of standards for chart design and colors that removed the design decisions so anyone on the team could build a clear, credible chart without starting from scratch. A starting point that helped our teams move quickly without sacrificing clarity or credibility. During a period when public trust in data was fragile and contested, consistent branding wasn't just aesthetic. It was part of how people learned to trust what they were seeing.

I didn't know it at the time, but that guide was my answer to the blank page problem.

Every civic team I talk to now, local government analysts, nonprofit data leads, public health communicators, runs into the same wall. The data exists. The analysis is done. Someone has a clear sense of what it means and why it matters.

But then they hit the blank page.

Data work has a clear process. You know the steps for cleaning, analyzing, and verifying. But synthesizing that work into something that actually communicates — finding the main point, translating it into plain language, framing it for a specific audience — that's a different kind of work. It doesn't have the same kind of checklist. It just takes time, and time is the one thing most civic teams don't have enough of.

That's where the blank page comes from. Not a lack of skill or knowledge. Just the reality that turning data into clear, public-facing communication is its own discipline, and even when there's a communications team in the mix, the synthesis work still has to happen before anything can be handed off.

Every organization uses data. Private companies, nonprofits, government agencies — everyone is expected to back decisions with numbers and communicate findings to people who need to act on them. The pressure to turn data into something useful isn't unique to public sector work.

But the accountability is different.

When a civic team publishes an analysis, when it goes to a city council meeting, gets cited in a budget conversation, or supports a public health campaign, it answers to the public. Not shareholders. The people affected by those decisions didn't opt in. They're constituents, community members, taxpayers. That's exactly why the data underneath the work has to be defensible. The synthesis and communication can evolve through drafts. The data integrity can't.

Here's what I've come to believe after years of turning complex public health data into campaigns, press releases, and public-facing materials that had to hold up:

The first draft doesn't have to be perfect. It just has to exist.

Reacting to something imperfect is so much faster than generating from nothing. A rough structure with the right data underneath it, something you can push back on, reframe for your specific audience, and fill in with your own judgment and context, is not a shortcut. That's how good public-facing work actually gets made.

This is where AI enters the conversation in a way I think is genuinely useful for local government and nonprofit teams: not as a tool that produces finished work, but as something that clears the blank page so the human work can begin. The interpretation, the local context, the framing for a specific audience, the accountability that can only belong to a person — none of that goes away. AI just removes the paralysis that happens before any of it is possible.

Afia Tyus at the Southern Economic Advancement Project runs a monthly data series called Visualizing the South, built on the thesis that the story of Black community progress in America is, in large part, the story of southern progress. The two are tied together in ways that often get missed in national conversations. Each piece in the series makes that case with data.

One project started with a map she found on Reddit. The map made a compelling point, but she had no way to verify where it came from or trust what was underneath it. So she used Sidekick Build to recreate it — starting from a prompt, working with verified data, getting to a first draft she could actually evaluate and refine. The idea she had in her head became something she could look at, test, and stand behind.

That's not a story about AI doing the work. It's a story about AI clearing the path so the work could happen. The insight behind Visualizing the South — the argument it makes, the communities it centers — that came from Afia. The starting point just didn't have to come from a blank page.

The teams I see breaking through aren't waiting for more staff or bigger budgets. They've changed what they expect from a first draft and what they reach for to get there.

Solving for the blank page isn't a productivity hack. It's the difference between getting critical information to the people who need it, in a way they can actually understand, and not reaching them at all. I learned that firsthand with COVID data for Coloradans. It's just as true for the civic teams doing this work today.

Recently, Afia joined us for a webinar on civic data work, including a live look at how she builds public-facing data resources as a team of one. Watch the recording here.

And if you haven't already, read our earlier piece on why AI hallucinations are a governance problem, not just a tech problem.

If you're navigating AI adoption in the public sector, let's connect.