How to Audit Your Encompass SDK Usage Before the 2026 Deadline
Every SDK migration starts the same way, whether you run it yourself or hire it out: you count what you have. ICE’s own migration guidance puts “audit current SDK usage” as step one, before anyone talks about APIs, and for good reason — almost every lender we’ve looked at was running SDK dependencies nobody on the current team knew existed.
With the December 31, 2026 transition deadline under six months away, here’s how to build an inventory you can actually plan from.
What counts as “SDK usage”
The audit needs to cover more than the plugins your admin remembers installing:
- Smart Client plugins — desktop plugins loaded into Encompass, whether built in-house or installed by a vendor years ago.
- Custom input forms with code-behind — desktop forms are already a rebuild-for-web story, but forms whose logic calls SDK APIs are doubly exposed.
- Standalone SDK applications — console apps, Windows services, and scheduled batch jobs running on your servers that log into Encompass through the SDK. Nightly data pulls, auto-disclosure jobs, pipeline sweeps.
- Vendor-supplied integrations — third-party tools that connect through the SDK rather than through Developer Connect or Partner Connect. You may not have the source code, but the exposure is still yours.
- Legacy service ordering — the old-style vendor ordering flows (appraisal, title, flood, verifications) share the same December 31, 2026 deadline and move to Encompass Partner Connect.
Start with Customization Reporting
ICE’s prescribed starting point is the Customization Reporting module, which inventories the customizations living inside your Encompass instance — the plugins, forms, and instance-level artifacts your migration has to account for. Run it first: it turns “we think we have a dozen plugins” into a named list, and it’s the same list ICE expects you to be working from.
But understand what an instance-side report cannot see.
The manual sweep Customization Reporting can’t do
SDK applications that run outside Encompass — on a server, a VM, or that one machine under a former employee’s desk — authenticate in from outside. An instance-level customization report is not designed to enumerate external processes on your infrastructure. So pair the report with a sweep of:
- Scheduled tasks and services on every server your IT team associates with mortgage operations. Look for anything referencing Encompass assemblies or credentials.
- Service accounts in your Encompass user list. An account named
svc-batchorautomationthat logs in nightly is an SDK dependency wearing a name tag. - Vendor contracts and invoices. If you’re paying for an integration, find out how it connects. Ask the vendor directly: “Is your Encompass integration on the SDK, and what is your EPC migration date?”
- Institutional memory. Ask the longest-tenured ops people what “just happens automatically.” Automations only get noticed when they stop.
Build the inventory spreadsheet
One row per component. The columns that matter:
| Column | Why it matters |
|---|---|
| Component & type | Plugin, form, batch job, vendor integration, service ordering flow |
| Business function | What it does in plain English — “auto-orders flood cert at submittal” |
| Owner | Who answers questions about it (this column will have blanks; those are your risks) |
| Blast radius | What breaks in loan production the day it stops working |
| Migration target | Developer Connect API, EPC, native business rules / workflow, Plugins 2.0, or retire |
| Risk rank | Production-critical first, nice-to-have internal tooling last |
The “migration target” column is where the audit turns into a plan. Not everything gets rebuilt — a surprising share of old plugins can be replaced with native Encompass functionality or retired outright, which is the cheapest migration there is.
What lenders typically find
Across audits, the same surprises come up: a departed admin’s batch job still running on a forgotten VM, a vendor plugin from a tool the company stopped paying for, three custom forms that duplicate each other because nobody knew the first one existed. Plan for your inventory to be longer than your memory says it should be.
That’s not a reason to dread the audit — it’s the reason to do it now. Finding the forgotten dependency in July is a line item. Finding it in January, when it stops working, is an incident.
We do this audit as a free assessment — you get the written inventory and a fixed-scope migration plan, yours to keep either way.