Old COBOL. Excel macros. Stored procedures. Email threads. Slack. The 80% of your business logic that doesn't live in your warehouse — extracted and bound.
The actual rules your company runs on are not in your data warehouse. They are buried in a 1998 stored procedure that excludes vendor 4711 from AP aging, the VBA macro that computes DSO with a 90-day cutoff, the ABAP routine that classifies revenue recognition, the email from the controller in 2019 explaining why RSU vests are excluded, the Slack thread that established the new escalation policy, the Confluence page nobody updates but everyone references. Catalog tools embed a link next to the table and call it done — the text body, the code body, the macro body is never actually read. Eisberg parses every source. Extracts the entities, the rules, the exceptions. Binds them to your business ontology as governed first-class facts. Signs the provenance chain. Cites them in every agent answer. "The stored procedure last touched 2008-04-12 excludes vendor 4711 from AP aging — that rule is now bound to the AR-Aging metric and replayable from the audit log" is a different product than "the database schema is in this catalog." Every other platform ships the second. Eisberg ships the first.
Every source
Legacy code, stored procs, macros, email, Slack, Teams, Confluence, docs, transcripts, runbooks
Auto-bound
Every extracted rule tied to your business ontology — tier-graded confidence
Governed
Same seven-policy plane + Birth Certificates that protect structured data
What ships, in detail.
The sources where tribal knowledge actually lives
Not just Slack. Legacy code — COBOL, MUMPS, RPG, SAS, Fortran, ABAP, PowerBuilder, Delphi, classic ASP. Stored procedures, SQL views, triggers, materialized views. Excel formulas, macros, VBA scripts. SSIS packages, Informatica mappings, dbt models, SAS programs. Email archives, Outlook PSTs, Gmail vaults. Sales call transcripts (Gong / Chorus / Otter / Read.ai). Slack + Microsoft Teams (channels, threads, DMs the workspace owner consents to). Confluence + Notion + Google Docs + Office 365 + SharePoint + internal wikis. GitHub + GitLab pull-request bodies, code comments, commit messages, ADR documents. Runbooks, incident postmortems, change-management tickets. Support ticket histories. Audit walkthroughs and SOX evidence narratives. Regulatory filings. Onboarding documents. Every place a rule, exception, or definition was ever written down.
How we extract and bind it to your business
LLM-driven extraction over every ingested code body, message, document, transcript, and macro. Stored-procedure parsers walk SQL ASTs. VBA + macro parsers extract conditionals + thresholds + named ranges. Legacy-code parsers (COBOL / ABAP / SAS / RPG) extract the conditional logic that defines your business. Each extracted fact is bound to entities in your auto-discovered business ontology (Customer.AR-Aging.Definition · Vendor.Exclusion-List · Metric.Revenue-Recognition.RSU-Exception · Sprint.Velocity-Adjustment-Rule · Claim.Subrogation-Exception). Tier-graded confidence — high-confidence bindings auto-attach; medium queue for human approval; low surface for exploration. The platform asks before it acts.
Governance by construction — same policy plane, same agent identity
Tribal knowledge is some of the most sensitive data in your company. Every ingestion passes through the same seven-policy plane (role-based + attribute-based + row-level + dynamic masking + purpose limitation + consent + temporal) that protects your structured data. Every agent that queries a tribal fact references a signed Birth Certificate scoping which data classes it can read. Every retrieval logs an audit event. Consent is per-source, revocable per-source, and time-bounded.
Cited in every answer
When an analyst asks 'Why does our AR aging exclude these vendors?', the platform returns the structured-data signal AND cites the stored procedure from 2008 + the email from Finance that explains why. When an agent acts on a customer-impact escalation, the action record cites the Slack thread + the Confluence page + the SSIS package that handles the exception. Every answer is auditable to the human knowledge — and the legacy code — that informed it.
Cross-domain wiring across code + comms + business
Combined with the business ontology, tribal knowledge stitches across domains nobody else stitches. The 2008 stored procedure binds to the modern Salesforce.Account record it filters. The Slack thread in #revenue-ops binds to the Confluence page and the Jira ticket it references. The VBA macro that calculates DSO binds to the Tableau dashboard that displays it and the Slack thread where Finance approved the cutoff. Eisberg sees relationships across legacy code + modern systems + comms that no single tool could surface — because no single tool spans all three.
Privacy and retention guarantees
Customer-owned data plane means tribal knowledge sits in your bucket, behind your KMS keys, in your cloud. Default retention policies match the source system; per-source overrides are honored. Right-to-delete propagates through every bound fact. Eisberg never stores your tribal knowledge on Eisberg infrastructure.
Other capabilities that compound with this one.
Want to see it live?
A 30-minute demo against your real data. We'll show you this capability end-to-end and answer any architecture or security question your team has.