Behold.ai vs Oxipit
These two define the autonomous end of imaging AI, the only pair in the index that removes normal studies from human review entirely, both CE or UK and neither cleared for autonomous use in the US. So the buyer's real question is who stands behind an auto reported study that no clinician read. If you want the cleanest published autonomy evidence and CT head coverage, start with Behold.ai, but read its continuity history first. If corporate stability and the first ever autonomous reporting clearance matter more, start with Oxipit, now owned by the imaging IT vendor Sectra. Behold's original entity entered administration in early 2025 and a successor acquired the assets. Oxipit was acquired intact. Both publish strong operating detail; the difference is who will still be there.
- The cleanest published autonomy evidence at this frontier: peer reviewed work reporting 15 percent of chest X-rays auto reported normal at a 0.33 percent error rate, publishing the auto report rate and its error rate together, which is what a defensible autonomy claim requires.
- CT head as well as chest X-ray: it removes more than 60 percent of CT head scans from the reporting workload at a 93 percent negative predictive value, coverage Oxipit's autonomous product does not match.
- A concrete commercial benchmark: it positions explicitly against outsourced radiology reporting, so a buyer can price it against the per study outsourcing rate it displaces.
- The first regulatory approval anywhere for autonomous diagnostic reporting: ChestLink holds CE Class IIb across 32 European countries, and the company publishes the fullest operating detail in the category, over a year of supervised operation on more than 500,000 images at a 99 percent sensitivity threshold, with an audit trail.
- Stable corporate footing: acquired by the imaging IT vendor Sectra in April 2026, so procurement, roadmap, and support run through an established organization rather than a relaunched entity.
- A broader safety netted suite: 75 findings plus a quality module that continuously re checks AI negative studies, a standing false negative control.
Side-by-Side
| Axis | B Behold.ai |
O Oxipit |
|---|---|---|
| AI Centrality | ||
| Autonomy and Oversight Model | ||
| Model and Technology Transparency | ||
| Clinical and Operational Evidence | ||
| AI Safety and PHI Stewardship | ||
| HIPAA and BAA Posture | ||
| Security Certifications and Trust Center | ||
| FDA and Regulatory Status | ||
| AI Governance and Bias Disclosure | ||
| EHR and Interoperability Depth | ||
| Deployment Model and Data Residency | ||
| Commercial Transparency | ||
| Setting and Specialty Coverage |
Both sit at the autonomous frontier and remove normal studies from human review, so the governing questions are accountability and subgroup performance, and neither publishes demographic subgroup analysis, the pointed gap when a tool removes studies from review entirely, since any group the model underperforms on disproportionately receives an unread scan. Neither is available for autonomous use in the United States; Oxipit is CE only and Behold's autonomous operation is a UK and European deployment, with its FDA clearance covering triage only. Behold's regulatory standing is unsettled: NICE withdrew red dot in February 2025 after the original company entered administration, and a buyer must verify which authorizations, support terms, and indemnity transferred to the successor entity. Behold's materials do not address who holds clinical responsibility for an auto reported study no clinician read. Neither publishes pricing or a security attestation.