Welcome
Welcome to Corbel. I'm glad you're here, and I'm especially glad you're here now: you start the same month we kick off Kittredge Health System, and you'll be leading it.
You already know how to run an implementation. What you don't know yet is our customers: why each one bought, what went wrong where, and who to call when it goes wrong again. That history is spread across three years of Slack, Jira and Confluence. This is the short version, with links to the long one.
READ THIS TWICE The Wrenfield section. It's the reason our implementation playbook looks the way it does, and Kittredge's supply chain team will ask about it if they've heard the story. They probably have.
- Shadow Gabriel on the Lakemont wave 2 cutover
- Sit in on Kittredge kickoff prep with Beth Arnaud
- Read the security review standard answers before your first customer call
- Don't commit to a go-live date with anyone yet
Our first 1:1 is Tuesday at 11. If something here doesn't match what you're hearing, tell me. That's a gap in this document, not in you.
Your role
You're joining a team of eight implementation leads who take health systems from signed contract to live replenishment. Corbel is live at 23 health systems and 141 hospitals. Every implementation follows the same arc: item master cleanup, the interface build with the customer's ERP, then go-lives in waves.
That last part wasn't always true, and the reason is **Wrenfield Regional Health**. Wrenfield went live at all four of its hospitals on the same weekend in February. Eighteen percent of its active items had no validated unit-of-measure conversion, so orders that should have gone out by the case went out by the each. In the first week, one distributor shipped eleven times the expected volume on a set of high-use items. Three weeks after go-live, Wrenfield paused the rollout.
THE TWO RULES No go-live until 98% of active items have a validated unit-of-measure conversion, signed off by the customer's supply chain lead. No wave larger than two hospitals.
Gabriel Osei led the recovery and wrote most of the playbook changes. Getting all four hospitals back to live replenishment took four months. Wrenfield is stable now, and cautious, and its renewal comes up next year.
**Lakemont Health**, nine hospitals, took fourteen months from first meeting to signature. Seven of those months were the security review: SAML single sign-on against Lakemont's identity platform, a third-party penetration test, US-only hosting, and how we store and rotate the credentials we use for its ERP.
**Penhallow Health** has the one custom interface we maintain, to the homegrown warehouse system at its consolidated service center. It was a condition of the 2024 deal, and the decision record that followed says we don't build custom interfaces anymore.
YOURS TO LEAD Kittredge Health System signed in September: six hospitals, sold by Beth Arnaud. Kickoff is in your second week.
14 months From first meeting to signature at Lakemont Health, seven of them in security review
The people you'll work with
These are the people you'll work with most in your first ninety days.
- Renee Castellano (Director of Implementation) — Runs the implementation team and approves every go-live date before it goes to a customer. Ran supply chain operations at a health system for six years before joining Corbel. Your manager: weekly 1:1 on Tuesdays and the Thursday go-live review.
- Gabriel Osei (Senior Implementation Lead) — Led the Wrenfield recovery and wrote most of the playbook changes that came out of it. Running the Lakemont wave 2 cutover now. Your onboarding buddy, and the person to shadow for your first go-live.
- Lin Zhao (Senior Integration Engineer) — Owns the ERP and EDI interfaces, including Penhallow's custom interface. Knows which customer ERPs send non-standard purchase order acknowledgments, because she mapped most of them. The person to pull in before you tell a customer an interface question is simple.
- Sanjay Iyer (Security and Compliance Manager) — Runs every health-system security review and maintains the standard answers library. Took Kittredge's review from questionnaire to sign-off in five weeks. The person to ask before you answer any security question in writing.
- Beth Arnaud (Account Executive, Health Systems) — Sold Kittredge and owns the commercial relationship through the first go-live. Knows who championed the deal inside Kittredge, and who didn't. The handoff call with her is in your first week.
- Kendra Moss (Director of Customer Success) — Takes each customer from hypercare into steady state, and owns the Wrenfield relationship today. The person to ask what a customer is like after the implementation team leaves.
- Nadine Sorensen (VP of Customer Operations) — Leads implementation, customer success and support. Signed off on the Wrenfield recovery plan and the two-hospital wave rule. Your skip-level; she joins the kickoff for every new health system.
Gabriel ran the hardest four months this team has had and came out of it with a better playbook. Ask him about Wrenfield. He'd rather you heard it from him.
Reporting structure
You report to Renee Castellano, Director of Implementation. Renee reports to Nadine Sorensen, VP of Customer Operations, who also runs Customer Success and Support.
Implementation leads own a set of customers from contract to the end of hypercare, then hand them to a customer success manager on Kendra Moss's team. Integration engineers report through engineering, but every implementation has one assigned. For Kittredge, that's Lin Zhao.
- Weekly 1:1 with Renee, Tuesdays at 11
- Go-live review, Thursdays at 1pm: every date is approved here first
- Kittredge internal stand-up, Mondays and Wednesdays
- Skip-level: Nadine Sorensen, who joins every new customer kickoff
Escalations depend on what's broken. Interface failures go to #interfaces and the integration on-call. Security questions from a customer go to Sanjay before anyone answers. Anything that sounds like a scope change goes to Beth or the account's AE.
START HERE The implementation playbook (v3) and the Wrenfield retrospective are the two required reads. Most of the rest of this document assumes you have read both.
Tools & systems
**Jira is where implementations live.** Every customer has an epic in the IMPL project, with a cutover checklist per wave. Interface work sits in the INT project, owned by the integration team. If it isn't a Jira ticket, it isn't on the plan.
**Confluence** holds the playbook, the retrospectives, the security answers and the interface catalog. The catalog lists every standard interface we support: purchase orders, acknowledgments, advance ship notices and invoices over EDI, plus connectors for the ERPs our customers run. If a request isn't in the catalog, it's a scope conversation.
The one exception is Penhallow's custom interface. Lin spends about one sprint a quarter maintaining it, which is the clearest explanation of why we don't build another.
**Salesforce** is the customer record: contract terms, hospital counts, the implementation package each customer bought, and the account team. Check the opportunity before any kickoff; it says how many on-site weeks the customer paid for.
- Jira: an IMPL epic per customer, INT for interface work
- Confluence: the playbook, retrospectives, security answers and the interface catalog
- Salesforce: contracts, hospital counts, implementation packages and account teams
- Slack: one internal channel per active implementation
- #implementation (34 members): the whole team, and go-live announcements
- #kittredge-deployment (11 members): your implementation, internal only
- #lakemont-deployment (14 members): the wave 2 cutover
- #interfaces (26 members): EDI and ERP questions, and the integration on-call
- #security-reviews (9 members): customer security questions, routed to Sanjay
- #go-live-review (18 members): the Thursday review agenda and decisions
23 Health systems live on Corbel, across 141 hospitals
Culture & norms
THE WRENFIELD LESSON We would rather tell a customer "not yet" a month early than "we have a problem" a week after go-live.
**Security answers come from the library.** Every health-system security review asks some version of the same questions: single sign-on, where data is hosted, how ERP credentials are stored and rotated, the latest penetration test, breach notification terms, and exactly what data the platform holds. The approved answers live in one Confluence page that Sanjay maintains.
**Dates go to customers only after the go-live review.** Renee put this in writing in July, after a customer was given a go-live date in a status email before its item master had passed the readiness gate. The date slipped, and the customer's supply chain team found out from their own storerooms.
Until the item master passes the gate, the plan has gates, not dates.
Renee Castellano, Director of Implementation, in #implementation
**Every go-live gets two weeks of hypercare.** The implementation lead is on site for the first week, with a daily huddle with the customer's supply chain lead, and on call for the second. Plan on travel for every wave.
- If a security question isn't in the library, send it to Sanjay; don't improvise
- No go-live date in writing until it passes the Thursday review
- On site for the first week of every go-live, on call for the second
- Scope questions go to the AE before you answer them
Active projects
You'll run two things at once in your first ninety days: your first go-live as a shadow, and your first implementation as the lead.
**Kittredge Health System is yours.** The implementation epic is open, with kickoff in your second week and the first wave planned as two hospitals in the spring.
THE RISK Kittredge's first item master extract arrived last week: 61% of active items have a validated unit-of-measure conversion, against the 98% the readiness gate requires. That's normal for a first extract. It's also the Wrenfield pattern, which is why it gets your attention first.
Kittredge's ERP sends purchase order acknowledgments in a non-standard format. Lin has scoped the mapping, and it has to be finished before interface testing starts.
**Lakemont Health wave 2** is the go-live you'll shadow. Gabriel is running the cutover for two hospitals in your sixth week, using the checklist that came out of Wrenfield.
- Week 1, read and meet: the playbook and the Wrenfield retrospective, the Kittredge handoff call with Beth, and introductions to Lin and Sanjay.
- Week 2, kickoff: run the Kittredge kickoff with Nadine and Beth in the room, and walk Kittredge's supply chain lead through the readiness gate and why it exists.
- Weeks 3 to 5, the item master plan: agree a cleanup plan that gets Kittredge from 61% to 98%, with an owner on their side for every category.
- Week 6, Lakemont wave 2: on site with Gabriel for the go-live and the first week of hypercare.
61% Of Kittredge's active items with a validated unit-of-measure conversion. The gate is 98%.
That's where things stand this week. Welcome to Corbel.
Open questions
- Kittredge bought two on-site weeks per wave, and Lakemont bought three. Is two enough for a six-hospital system, or should we plan to ask for more?
- Who on Kittredge's side owns the item master cleanup? The kickoff deck names a sponsor but not a working owner.
- Wrenfield's renewal is next year. Is anything from the recovery still open with them?
- Has Kittredge said whether its ERP upgrade lands before or after our first wave?
- When a health system asks to speak to a reference customer, who do we offer, and is Wrenfield ever one of them?
- Does the two-hospital wave rule ever bend for small hospitals, or is it absolute?