Crayon Data Tangram AICapability Map
Demo portal/Platform & capabilities
The Specialized AI Agent Platform

Every product is the same platform, recombined.

These are the components we have already built. Pick a solution and step through exactly how it is assembled — the problem it answers, the specialized agents it draws on, which of your systems it connects to, and the workflow wrapped around it. Recomposition, not a rebuild.

PLATFORM MODULES

Connectors, guardrails, models, knowledge, prompts, memory, observability.

+

SPECIALIZED AGENTS

Capabilities, each on its own purpose-built framework.

+

DOMAIN WORKFLOW

Ticketing, RBAC, approvals, SLA clocks, evidence — per product.

+

CROSS-CUTTING ARCH

Human gates · immutable audit of every step.

=

A SOLUTION

Recomposition, not a rebuild.

The method

How a product gets built — in six moves

Build →

The components

What the six moves draw on

Integration is repeatable because it is organised by system class, not by project. Name your system — the connector class is already certified.

S1

Core banking & transaction systems

Finacle · Temenos · FLEXCUBE · iCore · CASA · cards switch · payments hub · loan origination
Read: balances, transactions, limits, customer master, product master. Write (gated): flags, holds, status updates.
REST/SOAP APIs, ISO 20022/8583 messages, DB views, batch files (SFTP), CDC streams
S2

Data platforms & warehouses

Oracle · SQL Server · Snowflake · Databricks · Teradata · Hadoop · S3 / object data lake
Schema discovery, governed query execution, feature and metric computation, lineage capture.
JDBC/ODBC, native SQL drivers, warehouse APIs, Parquet/Delta on object store
S3

Communication channels

Exchange / Microsoft 365 · Gmail · SMTP/IMAP · SMS & WhatsApp gateways · Teams / Slack
Inbound listening and screening, outbound drafting and dispatch after human approval.
Graph API, IMAP/SMTP, webhook subscriptions, provider REST APIs
S4

Document & content repositories

SharePoint · network shares · DMS · scanner intake · regulator portals · email attachments
Ingest, classify, extract, compare and archive documents with full provenance.
CIFS/SFTP, SharePoint API, S3 events, watch-folder polling, OCR pipeline
S5

Workflow, ticketing & case management

ServiceNow · Jira · BPM engines · in-house case tools · Crayon-native case module
Create and update cases, route to queues, hold SLA clocks, capture approvals and evidence.
REST APIs, webhooks, event bus (Kafka), native module when no system exists
S6

CRM & customer engagement

Salesforce · MS Dynamics · campaign managers · mobile / net banking app · push & offer surfaces
Serve ranked decisions and enriched entities into the surfaces customers actually touch.
REST APIs, Dynamic Lists API, SDK / widget embed, batch export
S7

Identity, access & entitlement

Active Directory / LDAP · Azure AD / Entra · Okta · SSO (SAML, OIDC) · bank RBAC matrices
Who may see which data, invoke which agent, and approve which action — enforced, not advisory.
SAML / OIDC federation, SCIM provisioning, LDAP bind, row-level security mapping
S8

Risk, compliance & regulatory

AML / fraud engines · GRC platforms · sanctions & watchlists · regulator circulars & portals
Pull rules and watchlists in; push alerts, evidence packs and regulatory responses out.
REST APIs, secure file exchange, portal automation, scheduled circular ingestion
S9

External & third-party data

Merchant & entity web sources · app stores · review sites · telematics · market and offer feeds
Enrich internal records with governed external signal — scraped, verified and compliance-checked.
Licensed APIs, governed crawlers, partner feeds, file drops
S10

Automation & RPA

Blue Prism · UiPath · Automation Anywhere · scheduler / orchestrator
Reach legacy systems that have no API — agents decide, RPA executes the keystrokes.
RPA queues, bot triggers, orchestrator APIs
S11

AI runtime & knowledge stores

vLLM / NIM inference · open-weight & hosted models · vector DB · feature store · cache
Where the intelligence actually runs — on the bank's own infrastructure when required.
Model endpoints (OpenAI-compatible), vector DB clients, GPU scheduling, container runtime
S12

Observability & audit sinks

SIEM / Splunk · ELK · Prometheus / Grafana · immutable audit store · data warehouse
Every agent step, prompt, retrieval, decision and human approval — traced and retained.
OpenTelemetry, syslog / HEC, log shipping, append-only audit tables

Seven modules in four layers. M1 maps one-to-one onto the twelve system categories; M2 is the core; M3–M5 are the auxiliaries that serve it; M6 and M7 are cross-cutting architecture — the guardrails, human gates and audit that make every output defensible.

Integration layerMaps one-to-one onto the twelve client system categories.
M1

Connector fabric

S1S2S3S4S5S6S7S8S9S10S11S12

The integration layer — and it maps one-to-one onto the twelve client system categories. Every category has a certified connector class: auth, schema mapping, rate limits, retries and lineage solved once, reused by every agent.

  • One certified connector class per system category — not per project
  • Credential vaulting with per-connector scopes
  • Schema-drift detection with lineage capture
  • Read/write separation — every write passes a human gate
Core selection layerRoutes work to the right specialized agent — and the right model beneath it.
M2

Specialised agent selection

S11

The core of the platform: route each unit of work to the right specialized agent — and, beneath it, the right model for that task. This is where token efficiency, long-context handling and multi-agent orchestration are decided.

  • Task-level routing to the agent whose framework fits the job
  • Model routing beneath the agent: small models for extraction, larger only where reasoning earns it
  • Token efficiency — retrieval and summarisation collapse context before it reaches the model, so cost scales with work, not with data volume
  • Long context handled by decomposition: state objects and scoped retrieval instead of one enormous prompt
  • Multi-agent orchestration and swarming — many specialized agents run in parallel over partitioned work, then reconcile
  • Open-weight default, hosted burst optional; published accuracy per task
AuxiliariesSupply the corpus, the prompts and the state that the selected agent runs on.
M3

Knowledge-base selection

S2S4S8S11

Auxiliary — supplies the right corpus for the question at hand: policies, circulars, schemas, product manuals, indexed with retrieval tuned per agent.

  • Per-domain corpora with access scoping
  • Chunking and embedding tuned per document class
  • Freshness policies and re-index triggers
  • Source citation on every grounded answer
M4

Specialized prompt libraries

S11S12

Auxiliary — versioned, domain-tuned prompt assets: the encoded output of the ~100 iterations an enterprise would otherwise run itself.

  • Versioned and diff-able, tied to eval results
  • Jurisdiction and language variants
  • Role and persona framing
  • Regression-tested before promotion
M5

Memory & state management

S5S11S12

Auxiliary — gives every unit of work a durable state object with an owned lifecycle, so agents can hand off, resume and swarm without losing the thread.

  • Durable state per case, not per chat session
  • Replay and resume after failure
  • Short- and long-term memory scoping per agent
  • API-driven, inspectable at every transition
Cross-cutting architectureApplied identically to every agent in every product: guardrails, human gates, audit.
M6

Guardrails

S7S8S12

Cross-cutting architecture — applied before an agent sees data and before any action leaves the platform, identically in every product.

  • PII detection, masking and redaction on ingress
  • Prompt-injection and jailbreak screening
  • Allow-listed action catalogue per agent
  • Confidence thresholds that force human review
  • Mandatory human gates on consequential actions
M7

Common observability

S7S12

Cross-cutting architecture — one way to trace, evaluate, audit and report every agent in every product. The layer a bank's risk function actually buys.

  • Step-level traces: prompt, retrieval, tool, decision
  • Ground-truth evals and accuracy dashboards
  • Immutable audit of every human approval
  • Export to the bank's SIEM and warehouse

Generic capability is a commodity. Each agent carries its own framework — the framework is the IP; the agent is its packaging.

A1

Document intelligence

Parse, classify and extract structured fields from enterprise documents at production grade.

Specialized because: Document-class-specific extraction schemas; confidence per field; layout + language robustness.

S4S1
A2

Legal clause extraction & mapping

Pull clauses from legal and regulatory documents; map mandatory vs non-mandatory obligations.

Specialized because: Obligation taxonomy tuned per jurisdiction — the capability an enterprise needs ~100 iterations to reach.

S4S8
A3

Document comparator

Compare versions, policies and contracts; flag material differences.

Specialized because: Materiality scoring, not diffing — what changed that matters.

S4
A4

Image intelligence

Compliance-grade image checks including letterhead, signature and Shariah screening.

Specialized because: Domain classifiers over generic vision; audit-ready verdicts.

S4S8S9
A5

Email listening & screening

Ingest, authenticate and triage inbound mail; screen scam and unauthorized senders.

Specialized because: Sender authenticity scoring; routing before reading costs money.

S3S7
A6

Email & response drafting

Draft jurisdiction-correct replies and notices for human approval.

Specialized because: Templated + generated hybrid; never dispatches without a gate.

S3S5
A7

Schema & DB intelligence

Connect any database; synthesize and understand tables, fields and relationships.

Specialized because: Metadata harvesting + semantic layer built automatically.

S2S1
A8

Text-to-SQL

Natural language to governed queries over live schemas.

Specialized because: Hybrid RAG + text-to-SQL framework: deterministic plan checker, self-correcting validator, read-only by construction.

S2S11
A9

ETL / transform agents

Describe, profile and transform tables into one standardized model with lineage.

Specialized because: Rule + pattern hybrid; lineage emitted as a first-class artifact.

S2S10
A10

Synthetic data generation

Generate governed, zero-PII test data preserving referential integrity and business rules.

Specialized because: Schema-driven generation with a five-state inventory lifecycle.

S2S10
A11

Anomaly detection

Rule-based and pattern-based detection across the full transaction population.

Specialized because: Full population, not 2–5% sampling; known and emerging deviations.

S1S2S8
A12

Entity enricher

Research and enrich merchants, offers and product entities from external sources.

Specialized because: Governed crawling plus compliance screening built in.

S9S6
A13

TasteGraph recommender

Transactions → affinity scores → ranked, explainable recommendations.

Specialized because: Two-tower embeddings; 2,888M customer-merchant affinities; sub-1.5s at scale.

S2S6
A14

Inference & decisioning

Score, rank and decide with explainability attached to every output.

Specialized because: Deterministic rules where required, models where they earn it.

S2S6S8
A15

Action & dispatch

Execute the approved action into the system of record — API or RPA.

Specialized because: Idempotent, gated, fully logged; RPA fallback for API-less systems.

S1S5S10S3
Why specialization — not generic agents — is the product

Capability finds purpose only when mapped to a known problem

Why this scales

Specialization is an economics argument too

Scale

Token efficiency

A specialized agent knows what it needs. Scoped retrieval and staged summarisation collapse context before it reaches the model, and small models handle extraction while large ones are reserved for genuine reasoning — so cost scales with the work done, not with the size of the data estate.

Scale

Long context, handled by decomposition

Rather than pushing an enormous prompt at a bigger context window, work is decomposed into durable state objects with scoped retrieval per step. Accuracy holds as documents and case histories grow, and nothing depends on a model vendor's context limit.

Scale

Multi-agent orchestration & swarming

Because agents are discrete and state is externalised, many can run in parallel over partitioned work and reconcile at a gate — full-population control checks, 50,000 notices a month, 421M transactions. Throughput scales horizontally; the audit trail stays single and complete.

Same modules, same agents — different workflow, different systems. Modules M1–M7 and the human-gate + audit layer apply to every one.

LEANM · Notice Management

Reads, verifies and routes every regulatory & law-enforcement notice — end to end.

Specialized agents
A5 Email listening & screeningA1 Document intelligenceA2 Legal clause extraction & mappingA4 Image intelligenceA6 Email & response draftingA15 Action & dispatch
Client systems touched
S3 Communication channelsS4 Document & content repositoriesS5 Workflow, ticketing & case managementS8 Risk, compliance & regulatoryS1 Core banking & transaction systemsS7 Identity, access & entitlementS12 Observability & audit sinks
Custom domain workflow
  • Seven-stage case pipeline (F1 acquire → F7 dispatch)
  • Mandatory DB-enforced human gate at F6 — nothing reaches a regulator without it
  • Deadline / SLA clocks per regulator and jurisdiction
  • Event-driven agent handoffs, each an audit record
30s notice → classified & routed · 22 languages · 88% eval accuracy · 100% outbound human-gated

Controls & Anomaly Detection

Continuous, full-population control monitoring with automated anomaly detection.

Specialized agents
A7 Schema & DB intelligenceA9 ETL / transform agentsA11 Anomaly detectionA14 Inference & decisioningA15 Action & dispatch
Client systems touched
S1 Core banking & transaction systemsS2 Data platforms & warehousesS5 Workflow, ticketing & case managementS8 Risk, compliance & regulatoryS12 Observability & audit sinksS7 Identity, access & entitlement
Custom domain workflow
  • Nine-stage control lifecycle — design, operate, detect, resolve, assure
  • Exception case management with multi-level approval and evidenced closure
  • Standardized data model across source systems with quality checks
  • Live control-effectiveness dashboards for audit
3M+ transactions monitored daily · 1.2M+ reconciliations/day · <60s breach-to-alert · 4,200+ man-hours recovered monthly

CXO Concierge

Ask the banking business anything — sourced answers, no analyst, no SQL.

Specialized agents
A7 Schema & DB intelligenceA8 Text-to-SQLA14 Inference & decisioning
Client systems touched
S2 Data platforms & warehousesS7 Identity, access & entitlementS11 AI runtime & knowledge storesS12 Observability & audit sinks
Custom domain workflow
  • 12-step question flow with retry loops and validator routing
  • Persona-aware framing — tone, KPIs and suggested questions per role
  • Read-only by construction: no INSERT/UPDATE/DELETE ever generated
  • Row-capped execution with source citation on every answer
$300K/yr analyst reporting cost saved · 500+ daily users · 200+ metrics always current

Test Data Hub

AI-generated, governed synthetic test data on demand — zero PII.

Specialized agents
A7 Schema & DB intelligenceA10 Synthetic data generationA9 ETL / transform agentsA15 Action & dispatch
Client systems touched
S2 Data platforms & warehousesS10 Automation & RPAS7 Identity, access & entitlementS12 Observability & audit sinksS1 Core banking & transaction systems
Custom domain workflow
  • Five-state inventory lifecycle: generated → allocated → reserved → released → consumed
  • Three-tier RBAC (org → product → project) with row-level security
  • RPA-driven orchestration into core systems with retry/failure handling
  • Audit-grade logging of every allocation and access event
Zero-PII by design · referential integrity across domains · ICICI reference deployment

Entity Enrichment (EMP)

Research, operate and distribute banking product entities at scale.

Specialized agents
A12 Entity enricherA1 Document intelligenceA4 Image intelligenceA14 Inference & decisioningA15 Action & dispatch
Client systems touched
S9 External & third-party dataS6 CRM & customer engagementS4 Document & content repositoriesS8 Risk, compliance & regulatoryS12 Observability & audit sinks
Custom domain workflow
  • 13 discrete agents composed by one orchestrator across research → operations → distribution
  • Compliance module per jurisdiction (RBI, IRDAI, MAS, Shariah)
  • Entity lifecycle states with immutable audit records
  • Unified API gateway serving structured entity data downstream
HDFC + ADIB live · 30K+ offers managed · 120K+ images processed incl. Shariah checks

AI Marketplace (Personalised CX)

Convert transaction data into real-time, ranked, explainable decisioning.

Specialized agents
A13 TasteGraph recommenderA12 Entity enricherA14 Inference & decisioningA15 Action & dispatch
Client systems touched
S2 Data platforms & warehousesS6 CRM & customer engagementS9 External & third-party dataS11 AI runtime & knowledge storesS12 Observability & audit sinks
Custom domain workflow
  • Four lifecycle stages: cold start → active → intent → lapsed
  • Six decisioning agents behind one Dynamic Lists API
  • Fatigue monitoring and rotation with anti-repetition rules
  • A/B experimentation built into the serving path
156M transactions analysed/month · 3.6M merchants scored · 37.8% visit-to-view vs 4.8% baseline

Read down a column to see how a product is assembled; read across a row to see how often a capability is reused.

Specialized agentLEANMControls & Anomaly DetectionCXO ConciergeTest Data HubEntity EnrichmentAI Marketplace
A1Document intelligence····
A2Legal clause extraction & mapping·····
A3Document comparator······
A4Image intelligence····
A5Email listening & screening·····
A6Email & response drafting·····
A7Schema & DB intelligence···
A8Text-to-SQL·····
A9ETL / transform agents····
A10Synthetic data generation·····
A11Anomaly detection·····
A12Entity enricher····
A13TasteGraph recommender·····
A14Inference & decisioning··
A15Action & dispatch·

The reuse is the argument: every engagement consumes proven capability and returns hardened capability to the registry.

Why specialization is the product

Agents find purpose only when mapped to a known problem

Generic capability sells nothing

“Document intelligence” sells nothing; “verify a law-enforcement freeze order in 30 seconds across 22 languages” wins a bank.

The 100-iteration argument

LEANM distinguishes mandatory vs non-mandatory legal clauses out of the box. An enterprise building this on a generic platform needs ~100 prompt iterations, domain experts and months of eval work — per use case.

Open source, made fit for purpose

We reuse what exists and innovate on top — architecture, evals, domain tuning. Said out loud: a strength, not an admission.

Differentiated architecture

Evidence, not adjectives

Framework-free by design

Frameworks add abstraction and micromanage state you don't need — and strip the control regulated workflows demand. We own every transition, API-driven. (We moved off LangChain/LangGraph on Controls and LEANM.)

Self-healing, storage-aware retrieval

How data is stored defines how it is retrieved; retrieval architecture, metadata design and self-correction are engineered together, not prompted together.

Event-driven agent handoffs

Every handoff is an auditable event; every notice a state object through a seven-stage pipeline with one mandatory, DB-enforced human gate.

One unique framework per agent

Example — the hybrid RAG + text-to-SQL framework (the Text-to-SQL agent, as used in CXO Concierge): schema-mapping agents, deterministic zero-LLM plan checker, self-correcting validator loop. 14+ agents behind one interface.

Workflow orchestration

Deliberately not a drag-and-drop builder

Real workflows carry semantics

Ticketing, statement management, RBAC, multi-level approvals, SLA clocks and evidence trails — a no-code canvas cannot express any of it.

Customized per product

The same base agents, wired into each solution's regulated workflow. That is the domain build — and the reason our products survive an audit.

From capability to your solution

The same components, wired into your workflow

Every capability above already runs in production. For a given problem we recompose the right agents, connect them to your systems, and wrap them in your regulated workflow — assembled and hardened, not built from scratch. The connector work to your systems is the only genuinely client-specific effort.

See the products live →Client briefings