Start Here
A reference atlas of the enterprise systems FDE-DEs integrate with on every deployment. The taxonomy, the political map of an enterprise data stack, and the per-entry format used in each chapter.
Scope & format
Each chapter covers a category (ERPs, procurement, CLM, warehouses, BI, identity, document sources, observability). Within a chapter, each major system has the same six sub-sections:
- What it is — the system's purpose in the enterprise stack.
- Who owns it — which team at the customer, the political weight, who you'll work with.
- How to integrate — API shapes, SDKs, connectors. With code snippets where they earn their place.
- Auth patterns — service accounts, OAuth, SAML, PrivateLink. The minimum the customer's IT team will let you do.
- Gotchas — the things that bite. Customizations, rate limits, undocumented constraints, the system's specific failure modes.
- Escalation routes — when something blocks, who at the customer can unblock it.
Format is intentionally consistent so you can skim a chapter, find what you need, and move on. None of these entries are complete API references — they're enough to brief yourself before a customer call, with pointers to authoritative docs for depth.
Taxonomy
| Category | Examples | What you'll do here |
|---|---|---|
| ERP / financial system of record | SAP, Oracle, Workday, NetSuite, Dynamics | Read supplier master, PO, GL postings. Write back rarely, with care. |
| Procurement / S2P | Coupa, Ariba, Jaggaer, Ivalua | Pull contract refs, supplier profiles; push renewal alerts, risk flags. |
| CLM | Ironclad, DocuSign CLM, Icertis, Conga, Agiloft | Pull signed contracts as the document source. Sync status when needed. |
| Warehouse | Snowflake, BigQuery, Databricks, Redshift | Deliver extracted data as the analytics-ready layer. |
| BI | Looker, Power BI, Tableau, Hex, Mode | Surface insights where business users already look. |
| Document sources | SharePoint, Box, Drive, NetDocuments, iManage | Where contracts live before they reach the platform. |
| Identity | Okta, Entra ID, Ping, Auth0 | SSO, SCIM provisioning, service accounts. The long pole of every deployment. |
| Observability & ops | Datadog, PagerDuty, Splunk, Grafana | Where pipeline metrics and alerts go; embedded-ops integration. |
The political map of an enterprise data stack
Every system is owned by a team that has opinions about who can touch it. Knowing the map saves weeks of confusion. The default ownership pattern at large enterprises:
- ERP — owned by Finance IT (sometimes called "the SAP team"). Gates writes aggressively, gates reads less so. Change Advisory Board approvals are the rule, not exception.
- Procurement suite — owned by Procurement Operations or Procurement IT. More willing to grant API access than Finance IT is to ERP.
- CLM — owned by Legal Operations. Usually open to integration if the platform doesn't disrupt their workflow.
- Warehouse — owned by Data Engineering / Analytics Engineering. Your peer. Where you'll spend most of your collaborative time.
- BI — owned by Analytics or BI team. Embedded in business functions sometimes.
- Document sources — owned by Corporate IT (file shares) or Legal Ops (NetDocuments / iManage). File-share access is usually easier than CLM access.
- Identity — owned by Security or IT Identity team. Smallest team, most overworked. The long pole.
- Observability — owned by SRE / Platform Engineering. Will want their tooling, not yours.
The implication: when you discover a new integration is needed, you're discovering a new team to work with. Their lead times are independent of your engineering velocity. File access requests early.
How to use the atlas
- Before a discovery call: skim the chapters for systems the customer has named. 15 minutes of skimming saves you from looking unprepared in front of their data team.
- When scoping an integration: read the system's entry end-to-end. The gotchas section is the most valuable — the things that bite are predictable.
- During implementation: code snippets in the integration sections are starting points, not finished code. They're calibrated to make the right pattern obvious; you'll adapt to your platform.
- When stuck: the escalation routes section tells you who at the customer to find. "I need to talk to your IT identity team" is more useful than "we have an Okta blocker."
Caveats
- Not exhaustive. Enterprise tooling is vast. This covers the systems FDE-DEs at contract-intelligence / document-AI platforms meet most often. There are 20+ ERPs we don't cover; if your customer runs one of them, treat this atlas as the framework and substitute the specifics.
- Snapshots, not real-time. Enterprise tooling evolves. API endpoints change. Best practice this year is yesterday's best practice in two years. Use the atlas to know what to look for; check authoritative docs for current details.
- US/Western Europe-leaning. Customers in other regions may run regional alternatives (e.g., SAP variants in Germany, Workday alternatives in Asia). Same patterns apply; specific vendor coverage is biased to the markets the contract-intelligence segment serves most.
- Code snippets are illustrative. Not production-ready integration code. They're the smallest snippet that makes the right pattern clear.