Is Your HR1 Crisis Engineering Center Ready?

H.R. 1 (formally known as the "One Big Beautiful Bill Act" or budget reconciliation law) is a federal law passed in July 2025 that tightens rules and adds work and check-in requirements for government safety-net programs like Medicaid and SNAP.

Something is going to go wrong with your HR1 implementation. Probably a few things. You don’t know what yet, and that’s fine; you can still be ready for when it does… and not only ready to solve the crisis, but to use every HR1 crisis moment as a chance to make lasting, positive, transformational change.

We’ve run some of the most famous (and most secret) crisis engineering centers in history, including the one for HealthCare.gov, and we have some tips.

1. Designate the room. Your key people have to be together, in person, and that’s not negotiable. Yes, that’s kinda gonna suck on January 1. (Did you watch Project Hail Mary? Leaders from around the world flew in, except for the handful of B-team losers on the video screen. Don’t be a B-team loser.) The room needs power, Wi-Fi with the password taped to the wall, a whiteboard, and a door that anyone in the agency can walk through without a keycard.
2. Designate the incident lead. We have a whole list of characteristics to look for in our book; this should not be the agency head. You want someone who’s comfortable walking and talking across tech, data, eligibility policy, and the call center floor, and who knows exactly who to ask about HR1 is actually being implemented in every one of your agency's silos. If the room is going to run around the clock, you will need more of these people, working in shifts.
3. Pick your crisis team. The test is whether you can make any decision, and begin implementing it, with only the people in that room. Apply the test seriously and you’ll probably end up with an unexpected crew: a mainframe programmer, a county eligibility supervisor, someone from the contact center vendor, and the one person who knows how notices get printed and mailed. That sounds like a great team to us! There’s lots of advice in our book on managing this, but start anywhere. A list with three names on it beats no list.
4. Delegate as much authority to that incident lead as possible, and do it now if you can. Once a crisis is underway, this might mean the lead can call the governor’s office and have it be answered immediately. Ideally, the crisis team in the room can make any policy interpretation or approve any technical change on the spot. They must also have a known, single hop for any approvals necessary. If you can’t delegate authority ahead of time, at least have the request and the approver lined up for when the crisis hits.
5. Set up a status page, off your current infrastructure. Pay particular attention to community-based communications: navigators, legal aid, clinics, food banks, and other people who’ll be fielding questions from your enrollees when a correction or a sudden change needs to go out fast. A second status channel for employees is worth it too, so your own staff get firsthand, trusted information about changes or outages instead of hearing about them from claimants or Teams chat.
6. Make it clear, and widely communicated, how to reach the room. We recommend a telephone bridge, always on, on a speakerphone in the middle of the table. A monitored Teams or Slack channel will do, but only if someone is assigned to watch it and read messages aloud as they arrive, and even then it isn’t as good. Email doesn’t count. The help desk manager who notices an unusual spike in tickets at 11 p.m. has to be able to reach the room without knowing anyone’s cell number.
7. Have your comms ready for multiple stakeholders. To have more of a sense of control, you can pre-write some of these. Tailor them to each audience. For instance, the governor should not be getting information on real-time network traffic, and your engineers should not be getting a plain-language, high-level summary when they need exact error message codes and server names.
8. Set up a dashboard of your key metrics, on a screen, in the room. Some measures worth tracking, current and over time:
   • Enrollment applications: opened, completed, and the completion rate
   • Contact center volume by reason, hold time, abandon rate, and resolution rate
   • Ex parte renewal rate
   • SNAP payment error rate
     Anything you’ll be measured on federally, you should be tracking now, yourself, in this room.
9. Create an action log. Every action taken gets written down, with the hypothesis behind it and the result. Did it work? If there’s no record of what you’ve tried and what you decided, you will find yourselves having the same argument on Thursday that you had on Monday.

If you have downtime, invest it in improving the dashboard. Your goal is to see any blip in the room first, rather than reading about it in the news or intuiting it from a call center deluge.

What does transformational change mean in the context of an HR1 crisis, you Pollyanna weirdo?

"Only a crisis—actual or perceived—produces real change. When that crisis occurs, the actions that are taken depend on the ideas that are lying around." – Milton Friedman

A crisis means you can make decisions and changes in minutes to hours that you might never get done outside the crisis. So: do you have any good ideas lying around?

As an example from a prior engagement, in 2020, California’s unemployment backlog was on pace to clear in 48 years. This crisis presented many new opportunities, including the department turning off manual-review flags that were diverting as many as 20,000 claims a day into a process that could only handle 2,400, because of low-fidelity “fraud” flags like: the applicant used a middle initial but the Social Security database didn’t have one listed. CA EDD selected and deployed a NIST-compliant identity verification service in weeks, stopping billions in organized international identity theft while simultaneously increasing access for everyday Californians. It adopted a data-driven burndown plan that put the most experienced claims processors back on claims, and it stopped letting brand-new hires work overtime (since their overtime hours mostly created rework for the experienced staff). That 48-year backlog was gone in three months. Many of those improvements had been started but languished in bureaucratic muck; it took the crisis to implement them right away.

If you have a wishlist like that, and an action plan to implement it, you’ll be well-positioned when things go sideways and someone needs a solution, pronto. Some candidates:

• Procurement flexibilities: the ability to award, cancel, or modify a contract in days
• The ability to move money, or get more of it
• The ability to hire, fire, or reassign personnel rapidly, grant overtime, and/or bring back retired annuitants
• A novel data sharing effort (say, with corrections) that nobody prioritized or could get approved under normal circumstances
• An authoritative statewide income verification service
• Previously verboten AI to take the busywork off your key staff
• Replacing a homegrown identity verification with an industry standard
• Pushing policy interpretations to the top of their license, so to speak
• Talking to other states, and stealing every good idea you hear
• Getting more development resources for core, overlooked infrastructure like a mainframe and its maintainers

It’s October. The room is a lot easier to set up now than in January.