What state education leaders should be asking about ESA program operations
Doug Dewey, Co-Founder & CGO
Education savings account programs are usually debated as policy and delivered as operations, and the gap between those two things is where families experience them.
A well-designed program with a phone line nobody answers is, from a parent's point of view, a program that does not work. Eligibility rules written carefully become unclear when the person explaining them has been trained for three days. Approved vendor lists are only useful if someone can answer questions about them in the enrollment window when every family calls at once.
Ahead of the 2026 National Summit on Education, these are the operational questions worth asking about an education savings account program, whether you are standing one up, reviewing one, or evaluating who should run it. EdChoice counts 18 states operating a total of 21 ESA programs (EdChoice), which means there is now enough operating experience that these questions have real answers.
Who actually contacts the program, and about what?
The first question is deceptively simple and often unanswered: who calls, and why?
An ESA program serves at least three distinct audiences, and they are not interchangeable. Families ask about eligibility, how to enroll, what expenses qualify, and where their reimbursement is. Schools and providers ask about approval status, payment, and documentation. Vendors in the marketplace ask about onboarding and transactions.
Each needs different knowledge from whoever answers. A support model designed around families only will handle vendor calls badly, and vendor frustration reaches legislators quickly. Our guidance on what ESA administrators should demand covers this from the procurement side. Ask for the contact mix by audience and by reason, ranked. If a program cannot produce it, the support model was designed on assumption.
Is the support model built for the enrollment curve?
ESA contact volume is not steady. It concentrates hard around enrollment windows, reimbursement deadlines, and any change in eligibility rules or approved expenses.
This is a staffing design question, not a capacity question. A team sized for the annual average will be overwhelmed in the enrollment window and underused in February. A team sized for the peak is expensive for ten months. The workable answer is a core team plus a planned surge, with the surge trained and certified before it is needed rather than recruited when the queue builds.
Ask specifically: what is the ratio between peak week volume and average week volume, and what is the staffing plan for that ratio? And ask how far in advance surge staff are trained, because a surge team that arrives untrained arrives as a quality problem.
Who wrote what the agents are telling families?
This is the question that gets asked after something goes wrong.
ESA programs carry real complexity around eligibility, qualifying expenses, and documentation, and the answers agents give become, in practice, the program's public interpretation of its own rules. That content needs a named owner inside the agency, an editorial review process, and a version history. When a rule changes, the change has to reach the knowledge base, the agents, and the public-facing content within a defined window, and someone has to confirm it did.
Ask who signs off on agent-facing content, how long a rule change takes to reach the floor, and what the process is when an agent gives an answer that turns out to be wrong. That last one distinguishes a mature operation.
What is being measured, and against what standard?
Program contact centers are often measured on answer speed alone, which is the easiest number and not the most important one.
Three measures matter more. Resolution on first contact, because a family that has to call three times about eligibility is experiencing a barrier to using the benefit. Written response time for email and portal contacts, which is how a great deal of ESA volume actually arrives. And resolution across the full contact mix, so vendor and school contacts are not hidden inside a family-weighted average.
Standards should be contractual rather than aspirational. On Mpathic's published work for a statewide education benefits program, the contractual first call resolution standard was 80 percent and delivery reached 93 percent, with average speed to answer at 12 seconds, email response within eight hours, and customer satisfaction at 4.88 out of 5.0, reaching go-live in roughly 90 days (State agency contact center case study). Those are the categories to write into an agreement. How fast the phone is answered matters less than whether the family got an answer.
Who is answering, and where are they?
For a state program, staffing is a policy question as much as an operational one.
The agency is accountable for every answer given in its name, which makes the training, supervision, and quality review of whoever answers a matter of public accountability rather than vendor management. Our account of running a call center for a state benefits program describes what that requires operationally. Programs handling family and student information also carry data handling obligations that belong in the contract explicitly.
There is a second consideration that state leaders raise consistently, which is where the jobs go. A program serving families in a state can be staffed by people in that state. On the engagement above, Mpathic recruited within the state served where practicable, with an embedded trainer and quality assurance specialist inside the delivery team and professional services covering the contact platform, CRM, knowledge base, and reporting under the same contract.
At the summit
Mpathic will be at ExcelinEd's 2026 National Summit on Education in Washington, D.C., December 2 to 4 (ExcelinEd). Doug Dewey, Gary Hale, Paul Szerszen, Andrew Craven, and Lisa Roder will all be there.
If you are attending, we would love to connect with you. We are glad to compare notes with state leaders and program administrators on what has worked operationally, and what has not.
Policy design decides what an education savings account program offers. Operations decide whether families can actually use it. The questions worth asking are concrete: who contacts the program and why, whether staffing is built for the enrollment curve rather than the annual average, who owns the content agents rely on, what is measured and against what contractual standard, and who is answering.
If you are reviewing a program now, start with the contact mix by audience and reason. Most operational problems in ESA programs are visible in that one report before they are visible anywhere else.
See our work with government agencies, or talk to our team
Frequently asked questions
What is an education savings account program?+
An education savings account program provides publicly funded, government-authorized accounts that families can use for approved educational expenses, which depending on the program may include tuition, tutoring, online programs, therapies, and instructional materials. EdChoice counts 18 states operating 21 such programs.
What kind of support does an ESA program need?+
Support across at least three audiences: families asking about eligibility, enrollment, qualifying expenses, and reimbursement; schools and providers asking about approval and payment; and marketplace vendors asking about onboarding and transactions. Each requires different knowledge, and volume typically arrives across phone, email, chat, and portal channels.
How should ESA program contact volume be staffed?+
Around the enrollment curve rather than the annual average. Volume concentrates sharply during enrollment windows, reimbursement deadlines, and rule changes. A core team with a planned surge works when surge staff are trained and certified before the peak rather than recruited once the queue has already built.
What should a state measure in an ESA program contact center?+
Resolution on first contact, since repeat contacts about eligibility indicate a barrier to using the benefit; written response time for email and portal contacts; and performance across the full contact mix so vendor and school contacts are not obscured inside a family-weighted average. Speed of answer alone is insufficient.
Who should own the content agents use to answer ESA questions?+
A named owner inside the agency, with an editorial review process and version history, because the answers agents give become the program's working interpretation of its own rules. Rule changes need a defined window to reach the knowledge base, the agents, and public-facing content, with confirmation that they did.
