Case study · Healthcare
Building an Agentic Platform for a National Specialty Pharmacy
A national specialty pharmacy partnered with Teleperson to replace a patchwork of disconnected staff tools and patient touchpoints with a single, agent-ready customer experience platform.
At a glance
National specialty pharmacy
- Vertical
- Specialty pharmacy (anonymised)
- Engagement
- Admin & Agent Portal + Patient Portal, built on a shared agent-ready data model
- Surfaces shipped
- Unified agent workspace · self-service patient portal · role-based admin console
- Why it's a Teleperson story
- The data model, APIs, and human-facing surfaces share one vocabulary — so an AI agent uses the same call as a human one.
The challenge
Staff stitching together four systems for every call.
Specialty pharmacy operations are uniquely demanding. Patients are managing complex, often life-sustaining therapies. Staff are fielding refill questions, prior authorisation issues, shipment inquiries, and clinical concerns — frequently for the same patient, across multiple channels, in a single day.
Before partnering with Teleperson, the client's agents were doing what most pharmacy teams do: stitching together information from a dispensing system, a communications log, a shipping portal, and a billing tool. Every patient call started with the same friction — who is this person, what are they on, what did we last say to them, and where is their order? That lookup tax compounded across thousands of interactions a week.
The client needed three things at once
A unified agent workspace
Collapse the lookup work into a single pane of glass.
A patient self-service experience
Deflect routine inquiries without sacrificing the personal feel of specialty pharmacy.
An agentic foundation
The platform itself had to be ready for AI agents to read, write, and act on the same data their human counterparts use.
The approach
Agentic isn't a feature you bolt on. It's an architecture.
Teleperson's thesis is that to get to an agentic enterprise, the data model, the APIs, and the human-facing surfaces all have to share the same vocabulary — so an AI agent looking up a patient's refill status uses the exact same call as a human agent clicking into the dashboard.
We built the platform in two tightly coupled surfaces: an Admin & Agent Portal for staff, and a Patient Portal for patients. Both read from the same patient graph: prescriptions, communications, receipts, profile, and clinical context.
That symmetry is what makes the platform agent-ready — anything an AI does on the patient's behalf is auditable in the same workspace the human agent already trusts.
The solution
Two surfaces, one source of truth.
A unified workspace for pharmacy agents
The Admin & Agent Portal was designed around the moment that defines specialty pharmacy: an agent picking up a call and needing context in seconds. Agents sign in and land at a search bar that accepts the three things they actually know about an inbound caller — last name, date of birth, or phone number. As they type, a live dropdown surfaces matching patients. One click opens the full patient workspace.
Inside the patient workspace
Patient Dashboard
Active and inactive prescriptions, recent communications, and recent receipts in one view.
Communications
Full thread of messages and call transcripts. Agents never have to ask 'what did we tell you last time?'
Prescription History
Every Rx on file with status, prescriber, fill dates, refills remaining, and carrier tracking links.
Receipt History
Every order with line items, totals, and downloadable PDFs.
Patient Profile
Editable in place: name, contact info, allergies, current medications, known conditions, profile photo.
Context that follows the agent
An agent's own profile surfaces the last opened patient's prescriptions and receipts for quick reference between calls.
Admin tooling that scales with the team
For pharmacy operations leaders, the platform includes a full admin console: a filterable user directory, role-based access (Admin and Agent), and the ability to add, edit, activate, deactivate, or remove staff accounts. Sensible guardrails — admins can't deactivate or delete their own accounts — prevent the kind of small mistakes that cause big outages.
A patient portal that earns its keep
On the patient side, we built a self-service experience that genuinely reduces inbound volume rather than just displaying information. Patients sign in with last name and date of birth and land on a dashboard that does the work for them: a time-based greeting, their active prescriptions front and centre, and tabs to switch between active, inactive, and on-hold lists. Each prescription card surfaces the details patients actually call about — prescriber, Rx number, dosage, quantity, last filled date, expiration, remaining refills, refill history, and a live carrier tracking link when a shipment is in transit.
Beyond prescriptions, patients can
- Review recent communications, including call transcriptions, searchable by date range or phone number.
- Pull up receipts with full order detail and download them as PDF or CSV.
- Search globally by Rx number, receipt number, or medication name.
- Maintain their own profile — addresses, delivery preferences, allergies, current medications, conditions, emergency contact — so the pharmacy works from current information.
Every one of those self-service actions is a call the agent team didn't have to take.
The agentic foundation
Why this is a Teleperson case study and not a portal build.
Every interaction in both portals — a patient updating their allergies, an agent reviewing a call transcript, a refill status check, a profile edit — flows through the same APIs and the same data model. That symmetry is the precondition for agentic AI in customer experience.
Same data path
An AI agent answers a patient's refill question using the same data path the human agent uses, with the same authorization model.
Same workspace
An AI's draft follow-up message lands in the same Communications tab the human team monitors.
Same accountability
An AI agent's actions are visible, reviewable, and reversible in the same workspace the team already lives in.
The pharmacy isn't replacing its team with AI. It's giving the team a platform where AI can take on the repetitive lookup-and-respond work, and human agents can spend their time on the clinical and emotional conversations specialty pharmacy demands.
Outcomes
Operating from a single source of truth.
Working with Teleperson, the pharmacy now operates from a single source of truth across its agent and patient experiences. Qualitatively, the team reports:
Faster time-to-context
Agents open a patient workspace and have prescriptions, communications, and receipts in front of them before the patient finishes their first sentence.
Fewer 'let me check on that' moments
Carrier tracking, refill counts, and recent communications are visible in the same view, eliminating the cross-system swivel.
Higher-quality patient self-service
Patients resolve common questions — where's my order, when was my last fill, how many refills do I have — without picking up the phone.
Cleaner administrative surface
Operations leaders manage staff access, roles, and account status without engineering involvement.
Audit-ready by default
Every action — patient or staff — flows through the same workspace, so reviewing what happened is a search, not an investigation.
A platform ready for what's next
The same architecture that powers today's portals is the architecture AI agents will use tomorrow.
What's next
Three pillars of expansion.
The pharmacy and Teleperson are continuing to extend the platform along three pillars:
- 01
Deeper survey and feedback workflows
Capturing patient sentiment alongside the operational data, in the same workspace, so trend signals are visible to both human and AI agents.
- 02
Full contact center integration
Every channel — phone, SMS, secure messaging, email — writes back to the same patient record, so context is unified regardless of how the patient reached out.
- 03
CRM-side automation
Turning patient signals into proactive outreach: refill reminders that know the carrier ETA, follow-ups that know the last clinical conversation, escalations that know which provider to loop in.
Each pillar reinforces the same thesis — that the agentic enterprise isn't a separate AI layer, but a customer experience platform where humans and AI agents work from the same information, in the same workspace, on behalf of the same patients.
Building toward your own agentic platform?
We work with a small cohort of design partners building the same architecture in their own vertical. Healthcare, banking, telecom, insurance — wherever a unified workspace is the precondition for agentic AI.