Back to case studies
MH

Meridian Health

Healthcare

Meridian Health used Brand Pulse and Knowledge Vault to catch AI answers misstating clinical information about its services, turning an unmeasured compliance risk into a monitored, correctable one.

97%

of AI-cited health claims verified as accurate, up from an unmeasured baseline

31

inaccurate AI answers flagged and corrected

+24 pts

visibility score gain on Perplexity and Copilot

18

compliance-reviewed pages shipped via Publishing Center

The challenge

For a healthcare provider, an AI assistant getting a fact wrong isn't a brand annoyance, it's a compliance and patient-trust risk. Grace Whitfield, Director of Digital at Meridian Health, had no visibility into what ChatGPT, Perplexity, or Copilot were telling patients who asked about Meridian's services, insurance coverage, or care protocols, and no way to know when an answer was simply outdated or actively wrong.

The stakes made a wait-and-see approach unacceptable, but Meridian's legal and compliance team also couldn't approve content changes on a hunch. Grace needed evidence: which specific AI answers were inaccurate, what they should say instead, and an audit trail showing the fix was reviewed before it shipped.

The approach

1

Monitor for accuracy, not just presence

Brand Pulse

Brand Pulse was configured around accuracy flags rather than visibility alone, surfacing any AI answer describing Meridian's services, coverage, or protocols in terms that conflicted with Meridian's published, compliance-approved source material.

2

Give the engines a single source of truth

Knowledge Vault

Knowledge Vault crawled and structured Meridian's approved clinical and service pages into a clean knowledge base, closing the gap between what compliance had approved and what was actually easy for an AI engine to find and cite.

3

Route every fix through review before it ships

Publishing Center

Corrected pages moved through Publishing Center with a compliance sign-off step built into the workflow, so every one of the 18 shipped pages had a documented review trail before it went live and was re-crawled.

The results

Brand Pulse surfaced and helped correct 31 AI answers that had been misstating clinical or coverage information, several of which predated Meridian's visibility program entirely and had been circulating unchecked. Verified-accurate citations of Meridian in tracked health-related prompts rose to 97%.

Visibility on Perplexity and Copilot, the two platforms Meridian's patients used most for care-related questions, gained 24 points as Knowledge Vault gave engines a clearer, more citable source than the scattered PDFs and legacy pages they'd been pulling from. All 18 corrected pages carry a compliance review trail through Publishing Center, giving Grace's team the audit history their legal partners required.

This isn't about ranking higher. It's about a patient asking an AI assistant a question about their care and getting an answer that's actually true. Brand Pulse is the first tool that let us prove that, not just hope for it.

MH

Grace Whitfield

Director of Digital, Meridian Health

See how AI talks about your brand.

Start your free 7-day analysis. No credit card required.