Prepared specifically for Dr. Ha & the team

Built for Quantum Pain & Spine.

I'm Anthony Kheng, an AI automation builder with six years in audited healthcare IT. You're building an AI-enabled operations practice on eClinicalWorks. So instead of just sending a resume, I mapped your bottlenecks, drafted a 90-day plan, and built a working AI tool for your intake workflow.

The bottlenecks I see

A scaling pain & spine clinic on eCW loses the most time in three places. Each one is automatable.

Prior authorization

The #1 pain-clinic drain. Manual payer portals, faxes, resubmissions. Prime target for AI-assisted drafting + status tracking.

Eligibility & benefits

Re-verified by hand per visit. eCW data + a verification workflow cuts denials before they happen.

Intake & triage

New-patient notes read and routed manually. The demo below shows how AI structures this in seconds.

My first 90 days

Listen first, ship a visible win fast, then build the measurement layer. Every step paired with a compliance check and a metric.

Days 1–30 · Learn & quick win
  • Shadow each department; map workflows to find the real #1 bottleneck.
  • Confirm the compliance baseline (BAAs, where PHI lives, eCW access).
  • Ship one small automation people feel immediately — proof I move.
Days 31–60 · Automate
  • Tackle prior auth / eligibility with AI-assisted drafting + tracking.
  • Stand up safe AI workflows — PHI only in BAA-covered tools, never consumer chatbots.
  • Document SOPs so the team runs the same play every time.
Days 61–90 · Measure & scale
  • Build dashboards/KPIs so leadership sees throughput, denials, SLAs.
  • Prove the metric moved on the first automation; pick the next one.
  • Hand Dr. Ha a prioritized roadmap of what to automate next, by ROI.

Live demo — eCW intake → AI triage

A working AI agent I built for this page. Paste a patient intake note; it returns a structured chief complaint, suggested ICD/CPT codes, and a prior-authorization draft. The kind of step that saves staff minutes per patient.

Output appears here. Powered by Claude, called securely server-side — the API key never touches the browser.

Demo runs on synthetic data only, no PHI. Output is a decision-support draft, not clinical or coding advice; in production a person reviews every draft before submission, and any real patient data would run only through a BAA-covered, HIPAA-compliant tier.

Built for your Microsoft stack

Your environment is Microsoft-based, with HubSpot for the marketing funnel. Here's the kind of KPI dashboard I'd stand up in Power BI, and how I'd wire it together with Power Automate while keeping PHI inside the compliance boundary.

Operations KPIs Power BI · illustrative sample data
Prior-auth turnaround 2.1days ▼ 34% vs. baseline
Denial rate 6.4% ▼ 21% vs. baseline
Days in A/R 38 ▼ 12 days vs. baseline
No-show rate 11% ▼ 18% vs. baseline
Denials by payer
Aetna
UHC
BCBS
Cigna
Medicare
Days in A/R · last 6 months
How it connects
eCW (EHR) Power Automate Dataverse / SharePoint Power BI KPIs

AI layer (Copilot / Claude) used only on a BAA-covered tier for PHI-adjacent work.

PHI / BAA boundary
HubSpot · CRM & marketing funnel non-PHI only

Proof I've done this

Not just theory.

~2 hrs/wk

Built an AI agent, unprompted

Custom agent (OpenAI API + MCP) that killed manual status entry.

−30%

Cut incident resolution time

Ran vulnerability management via Splunk SIEM + real-time triage. Data-driven, measured, repeatable.

6 yrs

Audited healthcare IT

Two healthcare orgs under active compliance audits — SOC 2, NIST 800-53. I build AI and keep it compliant.

Why me, why this role

I've spent six years going deep on healthcare IT and compliance, and the part that lit me up most was the AI and automation work building agent systems and advising leadership on how to actually put AI to work. This role lets me do that full-time, at the business-strategy level, with direct exec exposure and a clear path toward AI Operations leadership.

I'm an AI automation builder who already operates in audited healthcare. You're building AI-enabled operations infrastructure. That's exactly what I've been building, piece by piece, and I want to do it full-time here.