When it comes to languages, our country is a patchwork. Our civic infrastructure hasn’t kept up with more than one or two. There might be hundreds of languages spoken in our country, but they aren’t spoken by our government. Like the polyglot individual, who is fluent in many languages, government bodies and agencies need to become fluent in many languages in order to serve the people. To become a polyglot democracy, we need to design infrastructure that ensures certain patches aren’t left behind.

A crucial first step is to work toward including all eligible voters in the electoral process.

Translation is often framed as a technical problem that can be solved through effective bureaucracy. The assumption is that if the board of elections in a certain county is able to provide translated materials for every language spoken by eligible voters in their county, then we have perfect language access. In principle, I don’t disagree. However, the mere existence of a ballot in Lao, Hindi, or Mongolian is not a sufficient standard for measuring language access.

Language access is a technical problem, but not one that is solved simply by hiring translators and interpreters. Language access is about designing systems that include people in every step of the process. Language is often one of many barriers that voters face. That’s precisely why expert insight and the wisdom of communities are both crucial foundations for the polyglot democracy.

Tanzila Ahmed, whose organizing acumen is a constant inspiration, has applied a decade and a half of experience in Asian American & Pacific Islander (AAPI) electoral organizing to develop a theory of change model where low voter turnout isn’t caused by “voter apathy,” but rather that AAPI voters experience severe barriers when it comes to casting a ballot. She identifies five such barriers:

1. a barrier to voting information;
2. a barrier to the mechanics of voting;
3. a barrier to engagement;
4. a barrier to in-language resources; and
5. a barrier to voting rights.

These five barriers also translate to five critical needs that can’t be ignored when we talk about designing for a sufficient standard of language access.

An expanded standard of access means doing more than providing a written translation of any given ballot available. We also need to:

1. provide voter guides to help voters understand issues;
2. streamline the process of voting, so they can navigate its often complex mechanics; and
3. match them with an actual human from their community who can help make sense of a large volume of brand new information and help troubleshoot problems as they arise.

That’s precisely what we’re trying to do with VoterVOX, the newest tool from the Asian American & Pacific Islander new media organizers 18MillionRising.org. The app, currently in development, will connect Limited English Proficient (LEP) voters with multilingual volunteers to help them understand their ballots.

Communities that include LEP voters already have the expertise needed to include those voters in the democratic process. Creating access isn’t a matter of delivering information from a central source to LEP voters, but a matter of helping communities organize themselves. VoterVOX is as much about community organizing as it is about voting, and one-to-one connections are a vital component. I don’t want to build software that languishes in app stores or online. I want to build a tool that uses the beating heart of our communities to circulate fresh blood to its furthest-flung limbs.

We’re designing VoterVOX to include input from stakeholders—from LEP elders to multilingual high school kids to organizers working at the grassroots level—in order to understand their needs and expectations when it comes to community technology. Regardless of what the outcomes of working with these folks might be, we have some core assumptions about design—and language access more broadly—that guide our efforts to engage them in the first place.

Committing ourselves to language access means committing to providing more than just translated ballots. Translated ballots are just the first of many steps toward trying to change a culture around civic participation. Through a well-designed workflow for ballot translation, we can simultaneously create conditions that foster engagement where discrimination, lack of information, and structural exclusion have previously made participation difficult, if not impossible. When we’re designing to expand access to the ballot box in a landscape of problems, we’re working to right structural wrongs.

Designing for inclusion isn’t easy. In fact, it’s very difficult—otherwise this effort wouldn’t be needed.

Good design won’t restore key provisions in the Voting Rights Act, the key law that has expanded access to the vote for millions of voters, which was gutted by the Supreme Court in 2013. We still need to fight to protect the voting rights of all citizens of this country, in the streets and in the courts. We still need to pressure county boards of elections to do the right thing and obey the law by providing translated voting materials when they’re required to.

That work starts at home, in our communities. By building opportunities for connection between people with expertise and people with need, we’re changing the language around democratic participation. In the one-to-one link between a volunteer translator and a voter, an opportunity for organizing grows. That organizing is the real meat of civic engagement—it’s fuel for the long game of language access in a polyglot democracy. True language access requires a commitment to organizing by design.

Follow the quest to design better tools for a polyglot democracy on Twitter @votervox.

Cayden Mak (@cayden) is Chief Technology Officer at 18MillionRising.org, an organization founded in 2012 to organize Asian Americans and Pacific Islanders online. For the past three years, they have designed, hacked, and deployed tech to better organize people and promote popular education in the AAPI community for civic engagement, racial justice, and transformative structural change.