How I Built the Salesforce Infrastructure That Powered a PBM's Core Promise
- Company
- SmithRx — pharmacy benefit manager
- Role
- First Salesforce hire (Aug 2022 – Feb 2025)
- Scale
- ~650,000 members at exit
- Stack
- Service Cloud, Experience Cloud, Apex, Pardot, ZeroBounce, Twilio, Talkdesk, DocuSign CLM, CoverMyMeds, Salesforce Shield
The business problem
SmithRx sells self-funded employers a specific promise: “Had you been with us last year, you'd have spent significantly less on drug spend.” That promise is backed by a 15-page analytical report — but the report is only the sale. Delivering on it is harder.
The mechanism is member-level intervention. SmithRx finds members on expensive brand drugs — say, Biktarvy at $8,000/month — and guides them through prior authorization, prescriber coordination, and patient assistance programs. If it works, the member pays nothing and the employer saves $8k/month per member.
The entire model depends on one thing: can you find the member, reach them, and guide them through the process?
What I was asked to do
I joined as the first Salesforce hire in August 2022. No existing org, no team, no playbook. My mandate: build the operational infrastructure for member outreach, agent productivity, broker relationships, and compliance — in a HIPAA-constrained healthcare environment.
What I built
Contact Center (Service Cloud + Talkdesk)
Integrated Talkdesk CTI directly into Salesforce — screen pops on call connect, member record up before the agent says hello. Consolidated a fragmented tool stack (DocuSign, SMS-Magic, Opero Fax, Cloudingo) into a unified workspace.
Result: +20% agent productivity (calls per agent per day).
Email Verification at Scale (ZeroBounce + Apex)
The one that mattered most. Employer HR data arrives messy — wrong, missing, outdated emails. If you cannot reach the member, the intervention never happens and the savings promise breaks. I built a scalable verification service: ZeroBounce API via queueable Apex with intelligent batching to stay inside governor limits. 600,000+ emails verified. Silent failures in healthcare data have real downstream consequences, so the error handling had to be thorough.
Result: +30% outreach effectiveness, +13% program enrollment.
Phone Verification (Twilio)
Same principle, different channel. Twilio phone verification API with comprehensive error handling for PHI-adjacent member communications. Prescription outreach has zero tolerance for wrong numbers.
Prior Auth Modernization (CoverMyMeds)
Agents were submitting prior-authorization requests by fax. I integrated CoverMyMeds directly into Salesforce — electronic PA submission to physicians from inside the member record. Replaced a manual, error-prone workflow with a structured, auditable one.
Contract Infrastructure (DocuSign CLM)
Built contract lifecycle management with automated workflows and approval routing.
Result: −60% contract processing time.
Member Engagement Infrastructure (Pardot)
Wired the Salesforce–Pardot connector with sync rules scoped by enrolled program and cadence. Members segmented by drug savings program — wrong segment meant wrong message, which meant confused members and broken trust.
Broker Portal (Experience Cloud)
Brokers bring the employer clients. I built a stopgap Experience Cloud portal for 10–15 brokers while engineering built the full solution — case management scoped by account-contact relationships, with client-specific savings reporting by program. Honest about scope: an internal bridge tool, not a flagship product — but it kept broker relationships intact during the build.
Compliance Infrastructure (Salesforce Shield)
Every integration touched PHI. Before config started, I scoped each piece against HIPAA with legal and compliance. Implemented Salesforce Shield: PHI encryption, Field Audit Trails, Event Monitoring. Contributed to SOC2 readiness.
Reporting Performance
Indexed custom fields and composite external IDs. Executive dashboards went from 12 seconds to under 3 seconds.
The hard part
Building operational infrastructure in a regulated healthcare environment means every decision has a compliance dimension. You can't move fast and fix it later when the data is PHI.
The ZeroBounce work is the example: 600,000 records, Salesforce governor limits, and a business that breaks down if members go unreached. Getting the batching logic right — without silent failures — took real engineering attention, not just configuration.
I did this as the sole Salesforce person. No team, no senior SF resource to escalate to. Legal, compliance, and clinical ops all had to be in the loop before anything went to production.
Metrics
- Agent productivity
- +20% calls/day
- Emails verified
- 600,000+
- Outreach effectiveness
- +30%
- Program enrollment
- +13%
- Dashboard load time
- 12s → <3s
- Contract processing
- −60%
Have a system like this to build?
I do this kind of work end to end — audit, build, and run. If you have a Salesforce or RevOps problem that needs owning, let's talk.