Use case · Finance & payments

Fraud, payments and KYC on one memory

Enhance fraud detection, streamline payments and accelerate KYC with memory-powered AI that learns from every transaction. Today, most of that learning is thrown away the moment a check completes.

The mixture

01 · Persistent memory: Uses
02 · Current context: Light
03 · Traceable decisions: Uses
04 · Shared context: Leads
05 · Scoped retrieval: Uses

01 · What breaks

Every check starts from scratch

Legacy financial systems struggle with efficiency, speed and context retention. Three places it shows up on the bill.

01
Repetitive fraud checks
Legacy systems reprocess the same entities and devices every time. The device cleared yesterday is scored from zero today, which is redundant and expensive.
02
Rigid payments logic
There is no memory of past issuer quirks or merchant behavior, so the same avoidable declines keep happening. Higher declines, lost revenue.
03
Fragmented case data
KYC, device and transaction information live in silos. Agents cannot act fast, because none of them can see the full context of the case.

02 · What it does

Intelligent financial features

Transform financial operations with memory-powered AI that learns and adapts from every transaction.

Persistent memory for fraud prevention

Remembers devices and prior entities, so duplicate checks are avoided and token and infrastructure costs fall.

Checkout that learns over time

Alchemyst retains issuer quirks and merchant-specific logic, for faster, more accurate authorizations.

Smart credit & BNPL routing

No more sending full loan ledgers. Eligibility checks become 70% cheaper, and smarter.

Unified view of financial cases

KYC, payments and user fingerprints stitched into one memory, for faster resolution and fewer reruns.

03 · What it's made of

Mostly shared context. Then three others.

Every application on Alchemyst is a different mixture of the same five jobs. That mixture is what makes this a different piece of software from the one next to it, even though the API underneath is identical.

Leads · Shared context

One version of the business, however many agents read it

Fraud, payments and KYC agents read and write one case context, scoped by customer, merchant and device. What the fraud agent cleared, the payments agent already knows, and what KYC verified, the credit agent does not ask for again.

One case, one context, every agent.
Cleared entities stay cleared.
No full ledgers in the prompt.

Uses · Persistent memory

Devices and issuers, remembered

Prior entities, device fingerprints and issuer quirks persist between checks.

Uses · Scoped retrieval

Only the ledger lines that matter

Credit and BNPL checks intersect applicant and product scopes instead of sending the full loan ledger.

Uses · Traceable decisions

Why the case was flagged

Every decision carries the entities, rules and transactions behind it, ready for review.

That is four of the five. The fifth, current context (superseded versions are subtracted before the model ever sees them), is what leads in Employee support, Contract & legal ops and Financial ops instead. Same API, different mixture.

04 · What feeds it

The sources you already have

Bring sources in through the data-source integrations (PostgreSQL, MongoDB, Google Docs, Google Sheets, Amazon S3), an n8n workflow, or a direct context.add call. Each one lands scoped, so retrieval can intersect it with everything else.

Transaction streamsKYC recordsDevice fingerprintsIssuer responsesCase managementPostgreSQL & MongoDB

01 · Ingest

Scope what you ingest

Every document lands with a groupName: the sets it belongs to. Those sets are what retrieval intersects later, so the structure you choose here is the precision you get there.

context.add

import AlchemystAI from "@alchemystai/sdk";const client = new AlchemystAI(); // reads ALCHEMYST_AI_API_KEYawait client.v1.context.add({  context_type: "resource",  scope: "internal",  source: "kyc",  documents: [{    content: "KYC verified 2026-06-11: identity and address match. Risk tier: low. Linked device dev_8f21, cleared.",  }],  metadata: {    fileName: "cust_55102-kyc.json",    groupName: ["finance", "cust_55102"],   // the sets this belongs to  },});

02 · Write

Write what the check concluded

A cleared device or a known issuer quirk is worth keeping. Write the conclusion against the entity, so the next check starts from it instead of repeating it.

context.memory.add

await client.v1.context.memory.add({  sessionId: "merchant_4471",  contents: [{    role: "system",    content: "Issuer 5318 soft-declines first attempts over INR 50,000 at night. Retry with 3DS succeeds.",  }],  metadata: { groupName: ["finance", "payments", "merchant_4471"] },});

03 · Search

Search before scoring

Search intersects the scopes, subtracts superseded and duplicate content, and ranks what survives. Only that reaches the model, and the whole decision is recorded as a Context Trace.

context.search

const { contexts } = await client.v1.context.search({  query: "Score this transaction from dev_8f21 for cust_55102",  scope: "internal",  similarity_threshold: 0.8,  minimum_similarity_threshold: 0.5,  metadata: { groupName: ["finance", "cust_55102"] },   // ∩ narrow scope});// − superseded, deduplicated → ranked → into the window// Every search is recorded as a Context Trace.const reply = await llm.respond(message, { context: contexts });

npm install @alchemystai/sdk or pip install alchemystai, both ship the same client. Full reference in the docs.

05 · What it costs

Cost optimized. Out of the box.

Run agents with more power, at lower cost.

Does the heavy lifting of memory offloading by itself.
Optimizes token expenditure with efficient prompts.
Highly capable of handling longer tasks.

Get started

Tokens used, average per taskMemory connected

Standard agent

3M tokens

High token usage, short context window

Alchemyst-enhanced agent

1.8M tokens

Shorter prompt optimized with Alchemyst

06 · Where it runs

Financial data has the strictest boundaries

Transaction, KYC and device data are regulated in every market. Keep them scoped per customer and merchant, encrypted at rest, traced on every retrieval, and on dedicated or self-hosted infrastructure when your regulator requires it.

Security & compliance

Deployment

Managed cloud
Encrypted in transit and at rest, isolated per organization, and scoped at write time.
Dedicated infrastructure
Enterprise
Single-tenant, with VPC peering when your data cannot share a network boundary.
Self-hosted
On-premise
Run the context layer on your own infrastructure, with OpenTelemetry for observability.

07 · The rest