FinTech

FinTech Compliance Training LMS

FinTech companies and lenders train their people on rules that change, and they have to show who was trained on what, and when. Syllabium is an enterprise learning management system (LMS) for regulated industries, made by Athantic AI LLC.

FinTech is its field for lending: for banks, credit unions, mortgage companies and lending platforms, and for the loan officers, underwriters, processors and compliance staff who work in them.

What Is the FinTech Field in Syllabium?

The FinTech field sets Syllabium up for lending organizations, on the same platform as every other field, with the same sign-in, roles, authoring studio, versioned libraries and AI authority. It is designed around five kinds of lending, each with its own training.

Kind of lendingTypical loans
Consumer lendingPersonal loans, credit cards, auto loans and home equity lines of credit
Commercial lendingSmall business loans, equipment finance and commercial real estate
Mortgage originationResidential mortgages, jumbo loans, refinancing and government-backed loans
Small Business Administration (SBA) lendingSBA 7(a), 504 and microloans
Specialty and asset-based lendingFactoring and lending against assets

A field is a setting of the organization, fixed when it is created: it picks the organization’s defaults, grants no access, and claims no compliance. Regulated industries lists every field.

Which Courses Come with Syllabium?

The courses are your organization’s own: each organization owns its content, and your compliance team writes it in Syllabium Studio. Syllabium is built to SCORM 1.2 and 2004, the long-standing format for course packages, so packages your team already owns run beside the courses it writes. A library’s published releases can be shared with other organizations, inside a tenant or across tenants under a platform license, read-only or read-and-copy, and each share is recorded and revocable. A lending platform can train the institutions it serves from one library this way, while each institution keeps its own records. Courseware describes the studio and its libraries.

How Does Training Keep Up with Rule Changes?

When your team revises a course, the change passes one publishing gate that publishes, holds or refuses it and records the decision, and the author sees the outcome before clicking. Publishing workflows set reviewer chains, time limits and escalation for each kind of change. Each curriculum is told when a newer version of a library entry exists, and its owner accepts the update, defers it, compares the two block by block, or keeps a local copy. Learners stay on their version until the next one is published, and a verified question goes stale, and is no longer drawn, when a slide it cites changes. Lessons open on the cohort’s schedule, and a pre-course survey and prerequisites gate what follows. Each day an AI agent counts every learner’s overdue lessons, missed reflections and days unseen, and raises a suggested intervention that waits for a staff member; it calls no model and reads no learner’s writing.

How Is AI Governed Here?

AI in Syllabium acts through the same recorded route as people, under authority an administrator grants capability by capability.

ControlWhat it does
Twelve checksEvery agent action passes twelve checks in a fixed order; a refusal names its check, is recorded, and never reaches a model
Graded authorityFor each agent and capability: Allowed, Human review, Admin review or Blocked, and unset means refused
LimitsPer agent, actions and tokens an hour and spend a day and a month, with a cost cap on every call
Emergency stopRead first by every agent action; it covers all AI, one agent, one kind of work or one type of content
Second administratorLoosening a safety-critical setting, or any change an agent proposes, waits for a different administrator
Action logEvery attempt with the check that decided it, and every model call with its model, tokens and cost, kept seven years

No person’s identity is sent to a model, and authoring actions send content text only. Syllabium uses Claude models, through Amazon Bedrock and Anthropic's API. An organization can also bring its own model (BYOM). The route to a model is set for the organization, else for its tenant, else for the platform, and with no route no call is made.

Syllabium’s agentic RAG, retrieval-augmented generation that looks up sources before it answers, is grounded in each field’s own library. In FinTech it reads the FinTech field’s library and what the organization may read, each piece labeled with its source, and it never reads another field’s library. Each answer cites its paragraph and version, and “not enough evidence” is an allowed answer. Governed AI, Your own models, Agentic RAG and AI agents in training go further.

What Records Does a Compliance Review Get?

A compliance review gets tamper-evident records: each entry carries a fingerprint of itself and of the entry before it, and changing, deleting or reordering one makes every later fingerprint stop matching. An administrator exports one archive for a period, with each record, a verification report and a list of file fingerprints, and every export is logged.

A reviewer asksWhere Syllabium keeps the answer
Who completed which course, and whenThe learning record, kept as a tamper-evident chain, and each learner’s transcript as a CSV file
How each person scoredEach attempt, with the questions as asked and the pass mark applied
Why a result was changedA recorded reason for every flag, reset and clearance for credit
Who changed a course, and who approved itThe publishing gate’s decisions, which download as CSV, Markdown or PDF once closed
Who signed in, and who read whose recordThe sign-in record and the record of each reading, kept two years
What AI didEvery AI attempt and model call, in the action log

Most records are kept seven years, and a nightly purge records what it removes. Evidence and Training records for audits describe the records in full.

How Is Each Lender’s Data Kept Apart?

Each lender works in its own organization, and the database lets a session read only the platform’s shared rows and its own tenant’s and organization’s rows. A tenant is an owning company, licensed for a number of organizations, and AI settings at a lower level can only be stricter than those above. Record chains are kept per tenant, organization and record. The application cannot touch the integrity records directly, only through fixed functions. Any read across tenants by Athantic AI LLC is read-only, per session, and carries a recorded reason. The database refuses to delete an organization that still holds records. Platform shows the three levels in full.

Which Standards Is It Built To?

Syllabium is built to ten standards, and these four speak to security and to AI governance.

Built toWhat it covers
SOC 2 Type IISecurity controls for a service provider
ISO/IEC 27001An information security management system
NIST CSF 2.0 with SP 800-63BSecurity risk and sign-in strength
ISO/IEC 42001An AI management system

“Built to” is a statement about design: it names a standard the platform is designed and built to. Standards explains each of the ten, and ISO/IEC 42001 the standard for AI management.

Get Started

Enroll your organization online, see the pricing, or talk to us. Syllabium is made and operated by Athantic AI LLC.

Contact us