A resident uploads a pay stub to apply for benefits. A few clicks, a confirmation screen, and the document is submitted. In most, if not every case, a caseworker still needs to open the file, read it, find the income information, and type it into an eligibility system.
While the upload might only take a few seconds, the actual work is just getting started.
Across thousands of applications, renewals, and verifications, all of these steps add up. The paperwork has moved online, but much of the processing still lands on someone’s desk. For residents waiting on food assistance or healthcare coverage, that processing time can feel long and stressful. Processing these applications stands in the way of people trying to get access to basic necessities – food, medical attention, a benefit claim, and a lot more and they often add to their already cumbersome circumstances.
Following the enactment of H.R. 1, states administering SNAP benefits face major challenges that significantly increase operational risk and administrative burden.
So, the question is, how can states rise above this challenge to ensure that their citizens are not suffering, while they continue to process the hundreds of thousands of applications that come their way? Additionally, how do states ensure that fraud is prevented so they remain below the 6% payment error rate enforced by the Federal Government in order to avoid paying massive fines?
The Missouri Department of Social Services is addressing this challenge with Hyperscience and its broader technology partners, improving how incoming documents become usable information within its existing eligibility environment
To discuss the details of how they are approaching this issue, the CTO of our Public Sector business, Trevor DeLew and our leader for Public Sector, GTM, Carl Wright, sat down with Toi Wilde, Chief Information Officer at the Missouri Department of Social Services, for a webinar hosted by our partner, the American Public Human Services Association (APHSA).
Missouri’s experience reinforces a conviction behind our work and makes it clear that document processing belongs at the heart of benefits modernization. Improving the information entering eligibility systems gives agencies a practical way to address accuracy and staff capacity while larger modernization efforts continue.
The application is online. The document work still exists.
Across Missouri’s SNAP, Medicaid, and TANF programs, applications, renewals, recertifications, and income verification generate a steady stream of documents. They arrive as paper forms, uploaded PDFs, scans, and phone photos. As Wilde put it, “If it’s not paper, it’s still a document, right?”
About half of Missouri’s applications arrive electronically and half on paper, according to Wilde, with a goal of moving toward 80% electronic submissions. Even as more residents apply online, eligibility specialists still need to find and review the information inside each submission.
Hyperscience is helping Missouri tackle that workload so caseworkers can spend less time preparing information and more time moving cases forward. Getting information into the eligibility system faster is part of the job. Getting it right matters just as much.
Payment accuracy starts before anyone makes a decision
A misread income figure can cause trouble long before a case reaches determination. Downstream systems may process incorrect information exactly as designed, applying the right rules to the wrong facts.
As DeLew explained:
“Everything downstream inherits whatever happens there.”
For Missouri, getting its SNAP payment error rate below 6% is a clear priority. Wilde described examining every part of the process, including how documents are sorted, classified, and prepared for caseworker review. Missouri reviewed error trends and root causes alongside policy and training needs to identify where technology could help.
Wilde summed up the stakes:
“Speed without accuracy right now is just faster noncompliance.”
For us, that is a useful standard for benefits automation. Accurate information needs to reach the people and systems responsible for eligibility. Otherwise, faster processing can simply move a problem further downstream.
Improving that information gives states a way to address accuracy and workload without waiting for a larger technology overhaul.
Modernization within the systems already in place
Replacing an eligibility system is a major undertaking. Existing platforms contain years of business logic and support services people depend on every day. States need those systems to keep working while they introduce new requirements and improve operations.
Missouri chose a modular approach, focusing on how applications and supporting information are received, processed, and delivered into its eligibility systems. Hyperscience works alongside the state’s eligibility vendor and other partners to support that approach.
With Hyperscience Hypercell, agencies can automate document classification, data extraction, and validation before information reaches their eligibility systems. Uncertain information is routed for human review, while trained staff retain responsibility for eligibility decisions.
“We’re designed to plug into the world of system integrators, eligibility vendors, and not to compete with them,” DeLew said.
Wilde highlighted collaboration between Missouri’s eligibility vendor and Hyperscience as a source of positive outcomes and smaller wins that helped build momentum. For other states, the lesson is practical: identify what can improve within the current environment and make those changes part of the broader modernization plan.
Small wins can return serious capacity
Modernization roadmaps tend to look years ahead. Caseworkers are looking at the queue in front of them. Missouri is working on both.
The state prioritized high-frequency documents and lower-complexity forms for early automation. One example is paper SNAP applications. Missouri automated reading of these applications so that, when the required components are present, they can flow into the eligibility system and initiate the queue for subsequent steps, including interview scheduling.
Separately, Wilde described a paper-application workflow in which manual processing and review would take two to three times longer than the automated approach she outlined, which prepares information for subsequent staff oversight and review.
The capacity opportunity extends across Missouri’s broader automation efforts. Through other automations and additional APIs within its eligibility environment, the state freed close to 30 full-time-equivalent staff to return to application and verification processing.
That broader result shows what recovering capacity can mean: more staff available for benefits delivery. It also gives teams tangible progress to build on while larger projects continue.
Build for the people doing the work
Eligibility specialists handle complicated household circumstances, conduct interviews, and navigate hardship exceptions. These responsibilities draw on judgment and experience built over years of service, often while helping residents through difficult moments.
As Wilde explained:
“I don’t think there will ever be a replacement for human touch when it comes to serving others in some of their most vulnerable places.”
Giving staff more time for those interactions connects document automation directly to the agency’s mission: helping eligible residents get the support they need.
Getting the workflow right requires more than a requirements document. Missouri expects its technology partners to observe eligibility specialists and see how the systems are used. Wilde’s advice to other state leaders was direct:
“If you’re a state CIO or a state CTO, go shadow the field.”
Spend a day alongside an eligibility specialist, and the bottlenecks become harder to miss: information entered twice, documents that are difficult to find, handoffs that leave a case waiting. Those observations help teams choose improvements that fit the way services are delivered.
That is how we believe agencies should evaluate automation: does it give staff information they can trust and more capacity to serve residents?
Document processing is core to benefits modernization
Missouri’s experience makes the opportunity concrete. States can address document bottlenecks, strengthen incoming data quality, and return staff capacity while continuing broader modernization.
Hyperscience helps agencies put that approach into practice through intelligent document processing. Hypercell for SNAP applies these capabilities to the varied, complex documents involved in SNAP applications and ongoing eligibility workflows, supporting faster processing, better data quality, and more effective use of staff time.
The starting point is the work already in front of the agency: which documents consume the most time, where do errors enter, and what would change if staff received accurate information sooner?
Missouri is demonstrating how to act on those questions, turning less administrative work into more capacity to deliver the support residents depend on.
Watch the SNAP document processing demo or hear directly from Missouri in the full APHSA webinar.