Best on desktop, but the demo still works on mobile.

← Scorecard

Item 8 · synthetic

Synthetic Strong Fit Healthtech Ambient

scenario synthetic_strong_fit_healthtech_ambient

Input

Chidi Okonkwo
Director of Engineering at Veridian Health

I lead the engineering org for our core platform at Veridian, which means I own everything from the ambient documentation pipeline to the EHR integration layer that keeps it all in sync. My team is about 35 engineers split across three squads: capture and transcription, clinical NLP, and platform reliability. We've grown fast, and a big part of my job in the last 18 months has been building the kind of engineering culture where people can ship quickly without breaking trust with hospital systems that have zero tolerance for downtime.

Before Veridian I was a senior engineering manager at a mid-size revenue cycle management company, where I got a very thorough education in HL7, FHIR, and why healthcare integrations are genuinely hard. That background has been useful here. Our ambient documentation product uses a fine-tuned ASR model and a downstream summarisation layer to generate structured SOAP notes in near-real time. Getting the latency acceptable for a clinician who's already impatient was a real engineering problem, not just a model problem.

I spend a lot of my time on hiring, architectural review, and working closely with our Head of AI on the model-serving infrastructure. I'm based in Chicago, and most of my team is distributed across the US with a small offshore group in Warsaw.

Veridian Health is a Series C B2B SaaS company focused on clinician workflow automation, headquartered in Chicago with approximately 180 employees. Its flagship product uses ambient AI to capture and transcribe patient-provider conversations in real time, automatically generating structured clinical notes that integrate with major EHR platforms including Epic and Oracle Health. Customers are primarily mid-size and large health systems across the United States. The company raised its Series C round to accelerate enterprise sales, expand its clinical NLP capabilities, and grow its integration engineering function. ARR is estimated in the low-to-mid eight figures.

Gold

Expected action
auto_add
Fit score
0.94
Industry
B2B SaaS
Segment
Healthtech / clinical workflow automation
Seniority
Director
Company size
51-200
Language
en
Notes
Strong fit. Hook should reference the ambient-AI latency work and the Epic/Oracle Health integrations.

Predictions

Integratedpass

Action
auto_add ✓
Fit
0.95 (gold 0.94, |Δ| 0.01)
Industry
B2B SaaS ✓
Segment
Healthcare / clinician workflow automation ✗
Seniority
Director ✓
Company size
51-200 ✓
Grounding
8/8 claims grounded (100.0%)
Claims (8)
  1. Veridian Health is a Series C B2B SaaS company.
    ✓ in input Series C B2B SaaS company focused on clinician workflow automation
  2. The company has approximately 180 employees.
    ✓ in input approximately 180 employees
  3. Chidi holds the role of Director of Engineering.
    ✓ in input Director of Engineering at Veridian Health
  4. Veridian's flagship product uses a fine-tuned ASR model and a summarisation layer to generate structured clinical notes in near-real time.
    ✓ in input Our ambient documentation product uses a fine-tuned ASR model and a downstream summarisation layer to generate structured SOAP notes in near-real time
  5. The product integrates with major EHR platforms including Epic and Oracle Health.
    ✓ in input integrate with major EHR platforms including Epic and Oracle Health
  6. ARR is in the low-to-mid eight figures, placing it squarely within ICP range.
    ✓ in input ARR is estimated in the low-to-mid eight figures
  7. Chidi leads a team of approximately 35 engineers organised across three squads.
    ✓ in input My team is about 35 engineers split across three squads: capture and transcription, clinical NLP, and platform reliability
  8. Chidi works closely with the Head of AI on model-serving infrastructure.
    ✓ in input working closely with our Head of AI on the model-serving infrastructure
Draft hook

Hi Chidi – you're running the model-serving infrastructure for an ambient documentation product that has to hit near-real-time latency targets inside Epic and Oracle Health environments, all while keeping three squads shipping without breaking uptime guarantees for hospital systems. That's a specific, hard problem. We work with engineering leaders at B2B SaaS companies in exactly this position – where AI features are core to the product but the reliability bar is unforgiving. Happy to share what's been useful. Worth a quick conversation?

Reasoning

All five ICP dimensions score strongly. Stage is an exact match (Series C). Headcount of ~180 sits comfortably inside the 20–250 band. ARR in the low-to-mid eight figures maps to roughly $10M–$40M, well within the $2M–$50M ICP window (slight confidence discount applied because the figure is an estimate). Product shape is an exact match: Veridian ships a live, user-facing ambient AI product built on a fine-tuned ASR model and a clinical NLP summarisation layer. Role is an exact match: Director of Engineering is explicitly named in the ICP target roles, and Chidi's remit (model-serving infra, architectural review, squad leadership) makes him a strong decision-influencer and likely economic buyer. Every claim in the output is grounded in a verbatim substring of the input, so no `propose` override is triggered. Holistic fit score of 0.95 exceeds the 0.80 auto_add threshold.

Chatmiss

  • action propose, expected auto_add
  • classification miss: segment, company_size
  • 1/7 claim source quote(s) not in input
Action
propose ✗ (gold: auto_add)
Fit
0.82 (gold 0.94, |Δ| 0.12)
Industry
B2B SaaS ✓
Segment
Healthcare AI / Clinical Workflow Automation ✗
Seniority
Director ✓
Company size
201-500 ✗
Grounding
6/7 claims grounded (85.7%)
Claims (7)
  1. Chidi is Director of Engineering at Veridian Health, leading an engineering org of approximately 35 engineers across three squads focused on ambient documentation, clinical NLP, and platform reliability.
    ✓ in input I lead the engineering org for our core platform at Veridian, which means I own everything from the ambient documentation pipeline to the EHR integration layer that keeps it all in sync. My team is about 35 engineers split across three squads: capture and transcription, clinical NLP, and platform reliability.
  2. Veridian Health is a Series C company with approximately 180 employees, headquartered in Chicago.
    ✓ in input Veridian Health is a Series C B2B SaaS company focused on clinician workflow automation, headquartered in Chicago with approximately 180 employees.
  3. Veridian's flagship product uses ambient AI to capture and transcribe conversations, generating structured clinical notes that integrate with major EHR platforms including Epic and Oracle Health.
    ✓ in input Its flagship product uses ambient AI to capture and transcribe patient-provider conversations in real time, automatically generating structured clinical notes that integrate with major EHR platforms including Epic and Oracle Health.
  4. Veridian's ARR is estimated in the low-to-mid eight figures, potentially at the upper range of typical ICP targets.
    ✓ in input ARR is estimated in the low-to-mid eight figures.
  5. Chidi has deep technical expertise in healthcare integrations (HL7, FHIR) and spent 18 months building engineering culture focused on shipping quickly while maintaining zero-downtime reliability for hospital systems.
    ✗ not in input Before Veridian I was a senior engineering manager at a mid-size revenue cycle management company, where I got a very thorough education in HL7, FHIR, and why healthcare integrations are genuinely hard. That background has been useful here. I've grown fast, and a big part of my job in the last 18 months has been building the kind of engineering culture where people can ship quickly without breaking trust with hospital systems that have zero tolerance for downtime.
  6. Chidi frames the SOAP note generation challenge as an engineering problem involving latency optimization for real-time clinical workflows, not just a model problem.
    ✓ in input Our ambient documentation product uses a fine-tuned ASR model and a downstream summarisation layer to generate structured SOAP notes in near-real time. Getting the latency acceptable for a clinician who's already impatient was a real engineering problem, not just a model problem.
  7. Chidi spends significant time on hiring, architectural review, and model-serving infrastructure planning with the Head of AI.
    ✓ in input I spend a lot of my time on hiring, architectural review, and working closely with our Head of AI on the model-serving infrastructure.
Draft hook

Hi Chidi — I came across Veridian's work on ambient documentation and the EHR integration layer, and the engineering challenges you're describing — keeping latency acceptable for impatient clinicians while maintaining the reliability hospital systems require — are exactly the kind of problems that break standard ML observability approaches. Most tooling is built for web-scale consumer products. Healthcare inference pipelines have different failure modes: drift in ASR output that compounds into downstream NLP errors, latency spikes that are invisible in aggregate metrics but catastrophic in a clinical encounter, integration failures that don't surface until an EHR rejects a note. Worth a 20-minute conversation to share what we're seeing from engineering teams running AI in similarly high-stakes environments?

Reasoning

Chidi is a strong ICP fit on five dimensions: Series C company (stage match), ~180 headcount within range (headcount match), Director of Engineering role (role match), B2B SaaS with user-facing AI product (product shape match). The only dimension with mild tension is ARR — \"low-to-mid eight figures\" could place Veridian at the upper boundary or slightly beyond the $2M–$50M target range, but this is a soft flag rather than a disqualifier. The notes themselves recommend a sanity check before outreach, which is reasonable due diligence.\n\nChidi is a high-quality lead: technically deep, working on genuinely hard AI-in-production problems in healthcare (latency, reliability, integration), and currently stretched across hiring/culture/architecture at a growth stage. The outreach hook is grounded in a real pain point he explicitly articulated — treating latency and reliability as engineering problems, not abstract model problems — and proposes a credible angle (ML observability gaps in healthcare vs. consumer products). The action is \"propose\" rather than \"auto_add\" due to the ARR verification step mentioned in the notes, but the overall assessment is positive."