Capture.
The foundation your AI stands on.
Your AI is only as good as the data it sees.
Fragmented identities. Stale exports. Unverified signals. Feed those to an AI agent and it makes fragmented, stale, and unverified decisions. Flash Capture builds the foundation that changes that — a single verified member profile, real-time behavioral signals, and live warehouse data, unified before the AI ever touches it.
Loop
What Capture produces
Beyond a CDP — the AI-facing semantic layer.
A customer intelligence layer your AI can talk to.
A CDP unifies the profile, then hands your AI raw tables to guess at. Capture goes further: a governed semantic layer built on an object spine — identity-resolved members, orders and products — with three projections hanging off it: Tags (who), Intent (why now) and Metrics (how much).
Ask in plain language and it reads a certified answer, never an invented one — the number an agent acts on is the number on your screen.
The foundation · the noun layer
The object spine · the nouns everything hangs from →
who & whatIdentity-resolved business objects — members, orders, products — and the relationships between them. The three projections below all attach to this spine: a Tag, an Intent and a Metric for the same member resolve to the same object. It references, never duplicates — the spine supplies the objects; each projection keeps its own definitions.
This object graph is also where downstream governed actions attach — what can be done — but that lives in the Decide node, not here. See the complete semantic layer.
The three projections · semantics attached to the objects
Tags →
Stable, explainable labels of who a member is — lifecycle, value, preference, consent, risk.
Inference · “why now”Intent →
Hypotheses inferred from the live behavior stream — confirmed before action, never treated as fact.
Aggregate · “how much”Metrics →
Certified, single-definition business meaning over a population — measurement and feedback.
Capture → Decide
The AI brain reads the semantic layer this node builds.
Node ② (Decide) is where AI makes next-best-action calls — who to contact, with what offer, on which channel. It forms every call by reading Capture's governed semantic layer — the object spine and its three projections: Tags (who), Intent (why now) and Metrics (how much) — never by guessing at raw tables. Capture is what makes that layer trustworthy.
What the layer is built from
- ›One verified member identity across every channel
- ›Fraud-checked purchase history with no duplicates
- ›Real-time behavioral timeline, linked to the member
- ›Live warehouse signals without another data pipeline
Resolved into one governed semantic layer — an object spine with three projections (Tags, Intent, Metrics) — the single interface the AI reads it all through.
Decide input
The governed semantic layer — the object spine and its three projections (Tags, Intent, Metrics) — plus the last cycle's results, ready for the AI to read and form a next-best-action.
What changes
What the AI reads — raw tables, or a governed layer.
Without Capture, an agent guesses at scattered records. With it, the same agent asks the governed semantic layer in plain language and reads a certified answer.
Raw tables
Two rows for the same person, read as two strangers
A governed layer
One governed identity — the Tags projection reads them as one member
Raw tables
Last night's stale export, read as today's truth
A governed layer
Live, certified Metrics the moment the AI asks
Raw tables
Anonymous events that connect to no one
A governed layer
A linked timeline the Intent projection reads to infer why, now
The capabilities that build the foundation
Capture is not one black box — it is a set of governed, production capabilities. Each one contributes a piece of the foundation the AI stands on.
Build the member golden record with One ID
Member, tags, events, segments and relationships sit on the same One ID profile. Tags label who a member is, events capture what they do over time, and relationships connect them to other members and to products — governed projections of the same customer, read the same way by every agent, dashboard and API.
Members →
One verified, field-encrypted member record fused across every channel into a single behavioral timeline. Each profile also shows an expected next purchase date and a click likelihood — computed only when the member's own history supports them, never invented.
Tags →
The state layer of the AI-facing semantic layer: governed, explainable labels for who a member is — lifecycle, value, preference, consent and risk — the certified 'who' every agent, dashboard and API reads and acts on.
Events →
Raw behavior abstracted into an event model — a per-member timeline of what happened and when, the signal every agent reads to understand why, now.
Segments →
Turn One ID profiles and tags into governed audiences: rule-based, query-checked, and ready for activation. Lookalikes of your best members explain themselves — a “Why they're similar” summary shows exactly what each match is built on. Predictive conditions — each member's likelihood of clicking, and how close they are to their expected next purchase — make high-responder and replenishment-reminder audiences a single filter.
Relationships →
Orders and referrals abstracted into relationship models — a member-to-member referral graph and product co-purchase / category affinity — so reach and recommendations read from real connections.
Sync your business data
A curated set of enterprise-grade connectors, by category — eCommerce, ERP & finance, order management, CRM and your data warehouse — bringing the systems that run your business onto the one verified customer record through a single governed API.
eCommerce & Marketplaces →
Online orders, carts and first purchases
ERP & Finance →
Verified transactions, refunds and master data
Order Management →
The order lifecycle and fulfillment
CRM →
Accounts, contacts and service history
Data Warehouse & Marketing / CDP →
Propensity, churn and LTV signals
Point of Sale →
Earn and redeem at the register
Product Catalog & Industry Models →
An AI-ready PIM that flexes to every vertical — serialized luxury items, apparel size matrices, F&B combos — synced from ERP/Shopify and wired to order line items for category-level personalization.
Receipts (Offline) →
Turns an in-store receipt photo into fraud-checked loyalty points — AI reads it, rules route the unclear ones to a person or SoClaw, and a fraud agent guards the points.
Zero-Copy Data Sources →
Query Snowflake, Databricks, Salesforce DC and Delta Sharing live — no ETL, no permanent copy, governed by scope and a per-query cost budget.
Capture digital & physical behavior
Track digital behavior with Behavior Tracking, member-keyed store visits with Location Marketing (LBS), and turn physical foot traffic into owned members with branded Stores & QR sign-up — all attached to the same member timeline.
Behavior Tracking →
Capture behavior across email, the loyalty portal, your app, and your own site via SDK, fused into one identity-resolved member timeline.
Store & Location Marketing (LBS) →
Member-keyed store visits + a randomized holdout that proves incremental lift; privacy-by-construction location consent and proximity re-engagement — captured as a first-party location signal.
Stores & QR Sign-up →
Branded QR codes and a public capture page turn anonymous foot traffic into owned members: no app, no POS swap.
See the foundation through dashboards & AI insights
The governed data doesn't just feed agents — it feeds you. Read the business through live dashboards, and query the same certified metrics in plain language as AI insights, so people and agents work from one shared, trustworthy view.
Metrics — the AI interaction layer →
The aggregate layer of the AI-facing semantic layer: the certified, single-definition business meaning over any population or segment. Each metric defined and computed once on one engine — the single interface every AI agent, dashboard and API reads the business through, so they never disagree.
Data Governance →
Trustworthy numbers, member data handled responsibly, access under control and clean data going in — governed inside the platform, with no separate catalog, lineage or monitoring stack to run alongside it.
Data Compliance →
Meet the standards you're held to — audited to SOC 2 Type II and ISO 27001, and built to support your GDPR and CCPA/CPRA obligations with encryption, access control, data residency and versioned consent.
Take your members off-platform as ad audiences
The golden record isn't only for your own channels. Turn any segment into a privacy-safe, hashed audience for Google and Meta — consent-filtered server-side — so you retarget existing members and find lookalikes of your best customers on the ad platforms, all from the same governed data.
Extend Flash with your own data
Make the platform hold exactly the data your business runs on — no fork required, two ways. Add your own attributes to the members, stores and products Flash already tracks — a membership number, a preferred store, a child's birthday. And create whole new business objects Flash has never heard of — repair jobs, service appointments, warranty claims — with no code.
Custom Fields →
Add your own attributes to the members, stores and products Flash already tracks — a membership number, a preferred store, a child's birthday — and use them everywhere: segments, personalization and import.
Custom Objects →
Track something loyalty software has never heard of — repair jobs, service appointments, warranty claims — as your own no-code business object: name it, add the fields, collect records, and build audiences from them like any other object.
Developer topics
How Capture is built, governed, and secured.
Capture is the data foundation for AI decisions, so its architecture must be explicit — and the trust around it just as much. Start with the architecture, then see how the data is kept compliant, governed, and secure.
Data Architecture
How the foundation is built end to end: business facts, behavior events, change data capture and event streaming — plus live, zero-copy warehouse access with no second pipeline — served through a high-performance layer that turns raw data into a real-time foundation the AI reads.
Trust — compliance, governance & security
Compliance & certifications
SOC 2 Type II and ISO 27001, GDPR and CCPA — see how Capture meets each framework's requirements, and request our audit reports.
Visit the trust center→Data Security Management
Review the controls for API and data ingestion, event transport, processing, storage, application access, marketing activation, and audit response.
Review the controls→Data Governance
Trustworthy numbers, member data handled responsibly, access under control and clean data going in — governed inside the platform, with no separate catalog, lineage or monitoring stack to run alongside it.
See how it's governed→Related reading
Keep exploring the pages most related to this one.
Members
One verified, field-encrypted member record fused across every channel into a single behavioral timeline. Each profile also shows an expected next purchase date and a click likelihood — computed only when the member's own history supports them, never invented.
Read more CapabilityCustom Fields
Add your own attributes to the members, stores and products Flash already tracks — a membership number, a preferred store, a child's birthday — and use them everywhere: segments, personalization and import.
Read more CapabilityCustom Objects
Track something loyalty software has never heard of — repair jobs, service appointments, warranty claims — as your own no-code business object: name it, add the fields, collect records, and build audiences from them like any other object.
Read more CapabilityEmbedded Pages
Embed your own https-hosted pages and screens as permissioned menu items inside Flash — with single sign-on identity, per-role access and sandboxing. Early access.
Read more CapabilityConnectors & Notifications
Keep the rest of your stack in sync: real-time notifications the moment something happens in Flash, plus no-code Zapier and Make connectors into thousands of apps.
Read more CapabilityTags
The state layer of the AI-facing semantic layer: governed, explainable labels for who a member is — lifecycle, value, preference, consent and risk — the certified 'who' every agent, dashboard and API reads and acts on.
Read more