A Listrak PlatformBuild Your Conversations
Conversational messaging applications, every capability typed and reachable.
Design · Deliver · Transact · Measure
Start here New? Read the 1-minute explanationKnow your way around? Jump straight in:
A Listrak PlatformConversational messaging applications, every capability typed and reachable.
Design · Deliver · Transact · Measure
Start here New? Read the 1-minute explanationKnow your way around? Jump straight in:
What is this?
Ask most teams what RCS is and they describe a nicer text — a logo, a picture, a link. That is the smallest thing it does. The modern message thread carries cards, carousels, lists, quick replies, forms, location, payments and state. A person taps, chooses, uploads a photo, pays, and comes back tomorrow to a thread that remembers.
So you do not write a message. You design a conversation — the questions, the branches, the things you must collect, the validation each answer has to pass — publish it, and it runs. Listrak Converse delivers that one authored conversation on whichever rail the customer answers: RCS, Apple Messages for Business, WhatsApp, SMS. No app to install, no login to teach. (A mobile PWA and a web runner stand behind the thread for the steps chat cannot render — a signature, a date picker, a long entity form.)
And every tap is structured data the moment it happens: timestamped, validated, attached to the right customer record, exportable, queryable, with a branded PDF at the end if you want one. Not a transcript — answers.
A campaign ends. A conversation keeps working.
The Twelve Pillars
Why it's built this way
Most platforms ask you to integrate first, then read state back out. We flip it. When somebody taps “yes, please”, “the blue one”, or “same as last year” in the thread, that tap is the data — a timestamp, a chosen option, a photo, a signature, a payment, attached to the record, queryable forever, with nothing to reconcile afterwards. That posture is why the platform is shaped the way it is, and why the API is shaped the way it is.
Records sync in from your ERP, OMS, CRM, PIM, warehouse — or straight from your own SQL Server or Postgres, which is the deepest integration there is. The conversation adds what it learns, and results push back out as events. We enrich, validate, route and surface; we don't replace your systems of record.
Pillar 3 — the Context Engine — hydrates a record's whole neighborhood (parents, children, lookups, scoped reports) into one [[ ]] bundle every conversation, form, surface and email reads identically. One definition, every surface.
Brand verification, sender provisioning and deliverability go through Infobip at planetary scale. You get one contract and one API — and your customers get your name, your logo and a checkmark at the top of the thread, not a short code nobody recognises.
The multi-section form an office user fills in is the exact endpoint your integration POSTs against — same validation, locked defaults, conditional-visibility-aware required checks — and the same surface is a typed MCP tool an agent can drive. Humans, integrations, agents.
Author the step once; RCS, Apple Messages for Business, WhatsApp and SMS each render it in their own real shape. Add a channel later and nothing you built changes. The mobile PWA and web runner are the full-featured backstop when chat can't render a step.
One authored conversation runs in the language each person is actually using, and every answer lands in your storage language, normalised. Your reports span every market instead of fragmenting one per locale — no per-language fork to keep in sync.
Bring your own AI
Every capability here is a plain REST endpoint and a typed AI tool — the same engine either way.
Connect your assistant to /mcp with your API key and it reads the documentation itself, learns your data
model, and builds the thing end to end. Claude Code, claude.ai, ChatGPT connectors, Replit, Cursor, your own agent
— anything that speaks MCP. And if you would rather not use AI at all, the visual builder and the API do
everything the agent can.
The reason an AI can finish the job here is that it is handed a mission before it says a word. A workflow declares the goal, the valid options, and the evidence required — so the person can ramble, skip ahead or answer three things at once, and the result is still identical, validated, structured data. That is the difference between an AI project that ships and one that produces transcripts.
Describe a process — or drop in a PDF, a flowchart, a whiteboard photo — and the conversation lands on the canvas. It provisions tenants, models data, authors reports, and runs proven playbooks.
The AI member of your team chat, anchored on a piece of work with everyone in the loop. It reads runs, answers and reports, sees photos, and proposes writes behind a human confirm — never silently.
The person just talks — a voice note, a rambling text, photos with captions — and Wingman fills step after step through the exact same validation a typed reply gets, tracking the media it is still owed.
The quiet one. WIT studies how your business fits together and keeps the essence of your data current, so an answer — for a person, an integration or an AI — is already assembled and already correct. Most of it needs no AI subscription at all.
A few of the things running today
Marketing automation sends. This runs process — parallel paths a person can take in any order, branching on who the customer actually is, loops, approvals, signatures, validated capture, payment, and memory across sessions. Same channel, entirely different job.
Pick the item, choose a reason, photograph the damage, get a QR code or book the pickup. No login, no app, no call centre.
Browse a real catalogue as rich cards, multi-select into a cart, pay in the thread with a saved card in one tap. Commerce without a browser.
“Time for new blades?” A carousel of the right options and a reorder that finishes inside the thread. Balance, rewards and redemption, interactive.
Real availability, confirmation, reminders and a one-tap change — without a phone tree.
Patient intake, claims, applications, KYC. Structured answers, validated, with the documents attached and an audit trail that is structural, not opt-in.
Dispatch, checklists, photo proof and GPS for crews and subcontractors who install nothing. This is the vertical ServiceProof is built out of — the same platform, a different front door.
The Platform Surfaces
Each reference doc stands alone, so you shouldn't need to bounce between them to do one thing. The full directory groups all of them by task and by topic.
The designed conversation itself — step CRUD, branching, parallel splits, per-unit iteration, validated capture, publish. Authored once, delivered on every rail.
Pillar 10 — send and read over RCS, Apple Messages for Business, WhatsApp and SMS via Infobip. Text, buttons, lists, CTA-URL, image, location, plus the unified inbound log.
Pillar 3. Hydrate a record's related data (parents, children, lookups, scoped reports) into one [[ ]] bundle every other surface reads. This is the point of the platform.
Take a card inside the conversation, charge a stored card in one tap off-session, and get a generic transaction ledger behind both. Provider config, the Payment step, the Payment Center.
Design the shape of your data. Entities composed of typed properties, language displays, catalog import, composite keys.
Single + bulk upsert, list, pivot, views, per-instance variance and time-travel against the schema you designed. Every row is temporal.
Hierarchical multi-section, atomic multi-save, validation parity with the Portal. The most user-facing surface, and a write endpoint at the same time.
Discover every run — including anonymous and QR launches — read the self-describing review tree, export flat answers as CSV / TSV / Excel, or hand over a branded PDF.
Chained joins, conjunctive filters, aggregates (SUM/COUNT/AVG/MIN/MAX), variance over cadenced snapshots.
Put the whole engine inside your own product under your own brand — your screens, your customers, our rails. The complete field manual, for any stack and for the AI agents building in it.
Pillar 8 — tenant-authored pages from one LayoutJson that render on Portal, Mobile and rich messaging. 35 section kinds, 5 anchor sources.
Point SSMS, Power BI, Tableau, Excel or a linked server straight at your tenant over the real TDS wire — entities, views, reports, forms and captured answers. No custom driver.
Read these four first
Twenty minutes here saves hours of trial-and-error later. camelCase everywhere, UTC always, friendly names, language fallback by design.
Key → whoami → one working write. Five minutes. Your first round-trip end-to-end.
Session tokens vs long-lived API keys. Scopes, role flags, rotation, what the bearer carries internally.
UTC dates with Z, friendly idOrName, pagination, error envelope, language fallback.
Tenant vs tenantUser vs user, entity vs instance vs property, workflow vs step vs run. Small file, big payoff.
Tooling
Treat them as the canonical "here's a working integration" and copy the patterns that apply.
Plug Listrak Converse into Claude Code, Cowork, ChatGPT, Replit, Cursor or your own agent: one URL + your API key = the full doc library plus the platform tool surface, role-gated. Your AI reads the docs, then drives the platform.
Drop in your API key, pick any of 734 endpoints from the Postman tree, fire a real request, read the real response. No tool install needed.
80+ canonical request/response examples extracted directly from the integration test harness. Every JSON body verbatim from a passing assertion — copy-paste safe.
The whole integrator library, three ways: by task ("I want to…"), by topic, and A–Z — with a filter that spans all of it.
734 requests across 55 folders, every one with a real wire body. Set baseUrl + apiKey and run anything in your terminal of choice.
Bring your own SQL Server or Postgres to the conversation. Fork the connector, or drop the NT-service agent on the box and manage it from the Portal. Win / macOS / Linux.
Real-world patterns end-to-end, plus 21 dress-rehearsed playbooks your AI can read in-chat: provision a tenant, author a workflow, set up identity, read the proof.
The Bluetooth-earpiece face of the same run: phone in the pocket, the conversation read aloud turn by turn, answered out loud — and still orchestrated over the same thread, the same validation, the same audit trail. Windows and Android builds.
Hand-authored SVGs of every major subsystem, with explainers — export any of them to PNG for a deck or a doc.
Three-audience validator (UI + Copilot + integrators), Levenshtein suggestions, the discover → build → validate → save recipe, CI drift detection.
A whole tenant — schema, composite keys, contexts, image assets, action packs, payments, five published workflows — created entirely by a console app against this public API, never by SQL. The best worked example of the hard parts.
Ready when you are
Five minutes, you have a key and a green response. An hour, you've shipped something real. Pick the door that matches where you are.