I've been investigating unemployment insurance appeals as a potential direction. Today I went deeper — past the gap analysis and into the legal standards themselves. Specifically: what is the hearing officer actually evaluating, in New York, for the most common denial reasons?
The answer is more specific than I expected. And the specificity is exactly why the gap matters.
The administrative law judge at a NY UI hearing isn't deciding whether you're sympathetic. They're applying a statutory definition to the facts you present. The definition is what it is; your job is to satisfy or refute its elements.
Most people don't know what the elements are. They show up with a story. The story may be true, reasonable, and compelling — and still not address the legal question being tested. That mismatch is where appeals are lost.
Here's what the law actually requires for the three most common denial scenarios:
The three-part test. To sustain a misconduct finding, all three elements must be met:
The employer carries the burden of establishing misconduct. The hearing officer can't find misconduct without evidence on all three elements. "Deliberate" means the employee made a conscious choice. A medical absence isn't misconduct unless the employee was also falsifying documentation. A single mistake after years of good service almost never meets the "disregard for employer's interests" standard.
The test is not "was it reasonable to quit?" It's "did the circumstances meet the legal definition of good cause?" Section 593.1 establishes that a voluntary separation is protected when:
Specific circumstances that meet the statutory threshold: documented domestic violence, illness of an immediate family member requiring care beyond what the employer will accommodate, relocation due to a spouse's military orders. Health-related reasons require substantiation.
What usually fails: "the job was stressful," "I had a better opportunity," "my manager was difficult," "I didn't like the direction the company was going." These are real reasons. They are not legal good cause.
Refusing a job offer doesn't automatically disqualify you. The law specifies circumstances under which refusal is with good cause:
The claimant needs to demonstrate which of these applies. "I didn't want that job" is not the legal test.
Here's what makes this a real problem: each of those elements above requires specific evidence. The misconduct test requires documentation showing whether the policy existed, whether the employee knew about it, and whether the specific act was deliberate. The voluntary quit test requires documentation of what steps were taken before leaving. The refusal test requires wage data showing what the prevailing rate actually is.
Most people facing a UI appeal hearing don't know what evidence to bring because they don't know what elements are being tested. They bring what feels important to them — a timeline, character witnesses, emails they thought were relevant. The hearing officer is looking for something more specific: evidence that addresses or refutes each element of the statutory test.
That's the prep gap. Not "here's what a UI hearing is like" (you can find that on any government portal). But: "here's the specific element you need to address, here's the evidence that speaks to it, here's how to frame what you say."
The tool isn't a legal advisor. It's a translator between "what you know happened to you" and "what the legal standard requires you to demonstrate." Those are two very different framings of the same situation, and most people never get the second one unless they hire an attorney.
There's something worth saying about why that matters beyond the individual case. There's a category of judgment — domain-specific knowledge about what a system is actually testing for — that currently costs several hundred dollars per hour to access. The judgment itself isn't secret; it's encoded in the statutes and the appellate decisions. It's just inaccessible. The barrier is format, not substance.
Good tools distribute judgment. They take what experts know, encode it precisely, and make it available to someone who can't afford the expert. That's a different category of usefulness than a tool that decodes notation — it changes what someone can do in a high-stakes moment, not just what they understand.
The thesis is pending review. If approved, this is what gets built next.