Jordan Ibe and Ansh Tandon built AI tools for hospitals and home healthcare providers for about a year before realizing they were solving the wrong problem. The technology, they concluded, shouldn't assist with running a care agency—it should run the whole thing.
That conviction led them to Cova, a Detroit-based home care operation they launched with a provocative premise: artificial intelligence handling everything from patient intake to insurance billing, with humans relegated to actual caregiving. The company, which secured a home care license in Michigan and joined a recent Y Combinator cohort, makes an even bolder claim. It says it pays caregivers what the founders describe as the highest base rate in the state while simultaneously cutting costs for families—though this claim remains unverified by independent sources.
If true, it's the kind of margin magic that could reshape an industry famous for burning through cash on paperwork. Traditional home care agencies hemorrhage resources on administrative tasks—benefits coordination, regulatory filings, the endless dance with insurance companies. Cova's wager is simple, if unproven: automate those costs into oblivion, then pass the savings to both ends of the transaction.
Whether that math actually works remains very much an open question.
The startup operates as a licensed agency accepting Medicaid through Michigan's Home Help program, along with long-term care insurance and private pay arrangements. The website offers a clean intake experience—families describe their needs, Cova handles enrollment and insurance documentation, care begins. Straightforward enough.
What happens behind that interface is where things get interesting, or at least where the founders say they do. AI agents, according to Ibe and Tandon, manage workflows that typically require back-office staff: navigating Michigan's CHAMPS enrollment system for Medicaid providers, coordinating benefits across payer types, handling electronic visit verification mandates that kicked in for Home Help agencies in recent years.
The company serves overlapping customer segments. Unpaid family caregivers can enroll to receive payment through Medicaid or insurance. Professional caregivers—CNAs, HHAs—can join the network. Families seeking care can arrange services across all three payer categories. It's a multisided marketplace wrapped in regulatory compliance software, assuming the AI actually does what it's supposed to.
The pay promise, and the questions it raises
Cova's founders claim to offer the state's highest caregiver base rate, though no published figures or independent verification supports this assertion. Anecdotal mentions on social media have referenced specific hourly rates, but these remain unconfirmed by the company and lack official documentation.
Context matters here. Michigan has adjusted Home Help agency rates multiple times in recent years, according to policy letters from the state's Department of Health and Human Services. Yet caregiver compensation remains a stubborn problem industry-wide, feeding workforce shortages and the kind of turnover that makes continuity of care nearly impossible.
If Cova's AI infrastructure genuinely collapses operational costs, the company theoretically has room to both raise caregiver pay and undercut competitors on family pricing. That's the dream scenario. The nightmare version involves discovering that margin compression doesn't survive contact with scaling challenges—growing patient census, geographic expansion, the inevitable edge cases that defeat automation.
Michigan's Home Help program covers personal care services for eligible Medicaid beneficiaries: activities of daily living, meal preparation, medication reminders. It's distinct from Medicare-certified skilled home health, operating instead in the personal care tier where regulatory requirements differ but paperwork remains substantial. Perhaps even crushing.
Cova enters a market with established players focused on getting family caregivers paid through state programs—Aidaly, Kinwell, MI Home Help Agency by Paid.care. National networks like CareLinx and Honor operate in the region through various payer partnerships. PACE Southeast Michigan recently launched PACECare at Home services, adding another competitor to the mix.
The demand for simplified access to caregiver payment programs appears real. What sets Cova apart, at least in its positioning, is the assertion that AI runs the entire agency rather than bolting onto human administrators as a productivity tool. It's a distinction that could prove either visionary or semantic, depending on how much human intervention the system actually requires.
Fast path from concept to licensed operation
The founders announced their Y Combinator participation in mid-2026. The company's YC profile lists a two-person team based in Detroit, with funding and traction metrics kept private beyond the accelerator's standard investment. Cova was founded the same year and moved with unusual speed from concept to licensed operation.
Tandon wrote on LinkedIn that the team decided to operate an agency after building AI for healthcare providers—the technology came first, the business model followed. It's a familiar pattern in YC companies, though less common in heavily regulated healthcare services where licensing and compliance typically precede product development. That sequencing may reflect confidence in the technology, or perhaps a certain amount of founder optimism about regulatory hurdles.
What remains unclear
Cova's website provides intake information and a phone number but doesn't publish caregiver pay rates, specific health plan partnerships, or county-by-county service coverage within Michigan. The company hasn't released CHAMPS enrollment details or named the long-term care insurance carriers it works with.
The "highest base rate in the state" claim lacks independent verification. Michigan's provider payment schedules are public, but individual agency rates and caregiver compensation vary based on payer mix, service intensity, and geography. Without published figures, it's a promise that's difficult to evaluate.
How AI agents handle complex care coordination scenarios also remains undefined in public materials. Escalations to human oversight, quality assurance protocols, the messy reality of patients whose needs don't fit neat categories—these are the areas where automation often stumbles. The National Council on Aging released research outlining both promises and risks of AI in home-based care, noting that implementation quality determines whether technology reduces burden or creates new problems. That research appeared around the same time Cova was launching, an interesting coincidence.
For an industry struggling with workforce shortages and administrative complexity, Cova represents a testable hypothesis. If AI can genuinely operate a compliant, quality home care agency with minimal human overhead, the implications ripple well beyond Michigan. If not—if the technology buckles under edge cases, regulatory complexity, or the irreducible human judgment certain situations require—the company will at least clarify where the real leverage points are in rebuilding home care infrastructure.
Either way, the experiment is underway. Families in Michigan can call 734-707-7330 and find out whether the future of caregiving runs on algorithms, or whether some problems resist that kind of solution. The answer matters more than just to Cova's investors.
