Virgílio 'V' Bento on building and investing in the current AI moment
Apr 4, 2025 with Virgílio Bento
Key Points
- Bento argues healthcare costs rose while tech prices fell because the industry locked into a fully labor-intensive model, unlike consumer electronics where automation drove down costs.
- Sword's Phoenix AI system writes patient messages so warm and personalized that evaluators rated them more human than messages from burned-out clinicians.
- Sword's pelvic health product Bloom grew from $500,000 in 2024 revenue to $25 million in 2024, proving the expansion model despite initial board skepticism about niche markets.
Summary
Virgílio 'V' Bento, co-founder and CEO of Sword Health, makes a structural argument about why healthcare costs kept rising even as technology penetrated the industry: unlike consumer electronics, where technology replaced human labor and drove prices down, healthcare doubled down on a fully labor-intensive model. A TV that cost $3,500 thirty years ago now costs $500. Healthcare went the other direction. The fix, in Bento's framing, is shifting care from a human-first to an AI-first model — letting AI do part of the clinician's job while keeping the clinician in the loop at scale.
Sword started with physical pain, where the US spends $560 billion per year, with the bulk going to surgeries that Bento argues should be replaced by conservative, non-invasive care. The problem is access: three PT sessions a week for three months can resolve many conditions, but when that's unaffordable or unavailable, patients end up on an operating table instead. Sword's pitch is making that conservative-care pathway as accessible as running water.
The Phoenix finding
Sword's AI system, Phoenix, analyzes patient home-session data and drafts messages from clinicians to patients — personalized check-ins that reference session performance and patient goals. When Sword ran a blind test of 50 clinician-written messages against 50 Phoenix-written messages, evaluators consistently mislabeled them: the AI messages were rated as more human, and the clinician messages were flagged as AI-generated. The reason was straightforward. Clinicians operating near burnout write dry, concise notes. Phoenix writes warm messages with long memory, surfacing things patients mentioned weeks earlier — a goal to hike with their kids, a specific recovery milestone. The AI, in effect, made the clinician's communication more human than the clinician could manage under load.
Product expansion
Bento uses pelvic health as the clearest proof point for Sword's expansion model. The Bloom product, targeting conditions like post-childbirth urinary incontinence, launched in 2022 to near-universal skepticism from his board, who called it too niche. It generated $500,000 in revenue that year, $25 million in 2024, and is on track for $50 million in 2025. Mental health is next — Bento flags talk therapy, currently structured around once-a-week human sessions, as a direct analogy to where physical pain care was before Sword intervened.
Sword is also deploying AI voice agents for patient enrollment, predictive models to identify members likely to seek surgery within six months so clinicians can intervene earlier, and motion-quantification tools. Bento frames the roadmap simply: any area of healthcare still delivered through 100% human labor is a target.