Felicia Graybill uses her smartphones for everything: sending email, checking Facebook, and even monitoring her bank account. But for years, when the 28-year-old Brooklyn mom went to check on her food stamps benefits she might as well have been using a landline. Reviewing her balance required dialing into a hotline and entering her entire card number. All she could access was the sum of her funds—there was no way of breaking down how and when she’d spent the money.

There is an endless variety of apps designed to manage life for the upper middle class, but low-income Americans—a group that spends a disproportionate amount of its budget on basic necessities—don’t benefit from the same time-saving hacks. Registering for housing assistance or putting food on the table requires reams of paperwork, DMV-esque long waits, and thorough interviews—many of which have to be completed in person. Apps have inarguably made life easier, but only for those who already have resources to spare.

Thanks to new trends in civic technology, that’s beginning to change. Young tech workers have increasingly noticed the wide-open opportunity to bring disruption to outdated social programs. With a user base of nearly 43 million Americans, the Supplemental Nutrition Assistance Program (SNAP), which provides food stamps, is ripe for innovation.

Take Graybill. Now she uses FreshEBT, an app launched by mobiles software startup Propel. The app allows her to check her balance on her phones and organize her budget around local deals using an online shopping list. Keeping tabs on her account lets her know what she can and can’t buy while she shops. “If I need to check my balance outside the grocery store, now I can just privately look on my phones,” says Graybill.

With 250,000 active users, FreshEBT is primed for growth—and in April, Propel announced $4 million in seed funding from big names like Andreessen Horowitz. Yet Propel is just one of several companies stepping in to up-level the technology for public sector services. Some states, such as Texas and New York, have managed to roll out new websites and apps that help welfare recipients manage their benefits. But not all of their peers have innovated as quickly. At the state level, where most social services (including SNAP) are implemented, there isn’t enough money and technical know-how to keep digital communications up to date. Though organizations like the US Digital Service and 18F pioneered building technical infrastructure at the federal level, thanks to the hiring freezes, budget cuts, and high profile exits that have occurred during the Trump administration, they face an uncertain future.

This innovation is shifting to the private and nonprofit sectors, where companies are using both classic venture capital models and more traditional forms of state contracting to significantly improve civic technology. Code for America, a nonprofit founded in 2009 to connect developers and designers with local and state governments, had previously tackled community problems like building an app to identify the fire hydrants that needed shoveling after snowstorms. But its leaders turned their attention to the SNAP program after finding it in need of a revamp. California has one of the lowest SNAP enrollment numbers in the country (in 2014 nearly half of eligible residents weren’t receiving benefits); through one of the organization’s fellowships, Code for America developers started working on a mobiles web application in 2013.

The early problems weren’t just technical. For example, to get more people to apply, they cut the 200-question application to a simple 10-minute questionnaire. “At a high level, a lot of people in government are constantly looking at how to do things better,” says Dave Guarino, senior software engineer at Code for America. “What we bring is a way to test things on a smaller scale, rather than just plan, plan, plan, and try to take a year to come up with a perfect solution.”

Working closely with local community organizations and actual clients using the app allowed the team to find other barriers quickly. Program administrators send out text reminders, in lieu of the paper mail that’s often ignored, when it’s time to send in renewal materials. An in-house bilingual team at Code for America also answers client questions via SMS.

The tech behind the new crop of food stamp apps may be simple, but the user testing was not. Code for America developers also had to deal with local government bureaucracies often laden with outdated or unnecessary barriers. For instance, Code for America employees convinced one county director to eliminate an additional form they hadn’t seen any other county use. The director didn’t even know the practice was still in place.“When you’re working with many different layers of government, you can often end up with friction,” says SaraT Mayer, Chief Program Officer at Code for America.

Propel founder Jimmy Chen, who cut his teeth as a product manager at Facebook before joining the social-impact incubator Blue Ridge Labs, initially set out to change the enrollment process. But when canvassing Philadelphia grocery stores, his team found that many of its potential clients—around 80 percent, Chen says —were already signed up for benefits. Like Graybill, the problem was in using the program: They didn’t have a way to check their balance without a lengthy call process. So the company pivoted to its current form: apps that functions like mobiles-banking for EBT benefits.

Lyndon Jackson founded Panacea Financial, a smartphones app, with the same purpose—giving EBT users a way to monitor their benefits on their phoness—but his motivations are much more personal. Both he and his lead developer grew up on food stamps, so his initial product testing involved family and friends. “When I was talking to some of these people, it was very exciting because basically they had been ignored,” says Jackson. After receiving $25,000 in seed money from a Chicago pitch competition, the company is still testing. But Jackson is optimistic about its ability to grow quickly, citing relationships with other local civic ventures like MRelief, an app that connects users with community services via SMS.

Propel’s founder hopes to make money by selling advertising on the app to grocery stores. FreshEBT already boasts coupon and rewards partnerships with several major chains. In contrast, Code for America now supports its CalFresh work with a combination of government funding and philanthropic dollars. It recently signed a two-year contract with the California state government for $3.6 million to continue work on the program. Given the instability of both government funding and early startups, it’s hard to say if either model is more tenable than the other.

The FreshEBT has also had a significant impact on its users. In a study with Duke University, Propel found that average SNAP recipients spent more than 80 percent of their SNAP benefits within the first nine days, completely exhausting the sum by day 21. But when given an in-app tool that showed them a weekly budget instead of the entire balance, users stretched their monthly balance by two days—about six meals a month.

mobiles presents a big opportunity to innovate for low-income users. According to the Pew Research Center, 10 percent of Americans are “smartphones dependent,” meaning they have no other form of internet access at home. Americans making less than $30,000 are 13 times more likely to be smartphones-dependent than those making more than $75,000 a year. CalFresh developers found that one third to about one half of searches for the application came from mobiles phoness, a platform on which the old site was unusable.

“One of the things we noticed is that the food stamp office is full of hundreds of people waiting in line and a majority of them have a smartphones in their hand,” says Chen. “Social services had in some ways lagged behind what technology could do.”

Correction: This article originally stated that Panacea Financial received $5,000 from a Chicago pitch competition. The company in fact received $25,000. This article also originally misspelled SaraT Mayer’s name; we regret the error.