Spoki · Spoki framework · spoki.com/en/spoki-pillars
The three pillars of Spoki
Spoki rests on three pillars — full Customer Journey, Generative + Agentic AI, and Open Integrations — that work as one system: a single AI-powered conversation orchestrates marketing, sales and support across every customer touchpoint and every tool you already own.
Most platforms specialise in one slice of the customer life-cycle: a chatbot for support, an automation tool for marketing, a CRM for sales. Spoki is built around the assumption that those slices are the same conversation — only the question changes.
The three pillars below describe how the platform is engineered. Customer Journey defines the surface (every step, from first ad click to post-purchase loyalty). AI defines the intelligence (generative responses plus agentic decisions, never plain rule-based bots). Integrations define the openness (Spoki connects to your CRM, your e-commerce, your tooling — it does not lock you in).
End-to-end coverage
Full Customer Journey
Spoki orchestrates the entire customer journey — acquisition, conversion, support, retention — inside a single conversation thread, so marketing, sales and support teams share one customer record and one AI-augmented inbox instead of three disconnected tools.
Spoki was designed as a Customer Engagement Management (CEM) platform from day one, not a chat widget bolted onto a marketing tool. The same record of contacts, tags, deals and tickets is updated by every team across every channel: a click-to-WhatsApp ad opens a conversation that becomes a qualified lead, then a paying order, then a support ticket, then a win-back campaign — without leaving the thread. Marketing automations, deals pipelines, ticketing, broadcasts and AI agents all read and write to the same customer profile. The result is what most stacks promise but few deliver: a single conversation that travels with the customer through every step of the journey.
- Conversational marketing — broadcast campaigns, click-to-WhatsApp ads, abandoned cart, win-back sequences, A/B testing.
- Conversational CRM — contacts, custom properties, tags, deal pipeline, ticketing, multi-operator inbox with RBAC.
- Conversational commerce — Meta catalog, checkout, payment links, order tracking, post-purchase upsell — on Shopify, WooCommerce, Magento, BigCommerce, PrestaShop and custom stacks.
- Conversational support — AI Customer Service Agent + human handover with full context, SLA tracking, satisfaction surveys.
- Multi-channel orchestration — WhatsApp, Voice AI, Email, SMS, RCS share one inbox and one customer record.
- Unified analytics — revenue per conversation, ROAS per campaign, AOV uplift, cart-recovery revenue, support resolution time in one dashboard.
Two layers of intelligence
Generative + Agentic AI
Spoki ships two complementary AI layers — Generative AI for natural-language conversation and Agentic AI for autonomous, tool-using workflows — both in general availability, both trained on the merchant catalogue, documents and policies, both observable inside the same inbox a human operator uses.
Generative AI is the well-known layer: large language models that read intent, summarise threads, suggest replies and translate across 40+ languages. Agentic AI is what makes Spoki different: AI Agents that decide, act and use tools — they qualify a lead, recover a cart, place an order on Shopify, schedule a Calendly slot, read and write to HubSpot, escalate to a human when the policy requires it. Both layers are governed by the same controls: configurable temperature, prompt visibility for the operator, audit logs, optional human approval before sensitive actions, and a hard guarantee that the same AI is observable from the unified inbox so trust is never lost. Voice AI is in GA — not beta — for both inbound and outbound calls.
- Generative AI — intent detection, summarisation, reply suggestion, translation across 40+ languages, sentiment scoring.
- Agentic AI — Custom AI Agent, Sales AI Agent, Customer Service AI Agent, Voice AI Dialer (GA, inbound + outbound).
- Tool calling — agents read and write to HubSpot, Salesforce, Shopify, Stripe, Calendly and 100+ other systems.
- Hyper-personalisation — 70-dim behavioural embeddings, predictive send-time, AI section injection.
- Observability — audit logs, prompt visibility, optional approval steps, full operator handover with context.
- Compliance-first — GDPR, EU data residency, configurable retention, no model training on your conversations.
Fits every stack
Open Integrations
Spoki is open by design — 100+ native integrations, a public REST API, webhooks with signature verification, MCP and WebMCP servers for AI browser agents — so every record, event and action stays in sync with the CRM, e-commerce, billing and analytics tools you already own.
A platform that owns the conversation but not the data is a silo. Spoki publishes a public REST API documented on Postman, signed webhooks for every event, and a stable schema you can rely on. On top of that, 100+ native integrations cover the categories merchants actually use: e-commerce (Shopify, WooCommerce, BigCommerce, Magento, PrestaShop), CRM (HubSpot, Salesforce, Pipedrive, Zoho), billing (Stripe, PayPal), scheduling (Calendly), and the long tail via Zapier, Make and n8n. AI browser agents discover the platform through MCP and WebMCP — so the next generation of GPT, Claude and Perplexity agents can read and act on Spoki without a custom adapter. Open is not a marketing word here: it is the contract that lets your CRM, your e-commerce and your support stay where they are while Spoki adds the conversation.
- Native integrations (100+) — Shopify, WooCommerce, Magento, BigCommerce, PrestaShop, HubSpot, Salesforce, Pipedrive, Zoho, Stripe, PayPal, Calendly.
- Long-tail connectivity — Zapier, Make, n8n, GoHighLevel and a public REST API documented on Postman.
- Webhooks — signed events for messages, contacts, deals, tickets, orders, payments and AI agent activity.
- AI agent discovery — MCP server at /.well-known/mcp.json and WebMCP server at /api/webmcp expose 10+ tools.
- Bidirectional sync — record updates in HubSpot, Salesforce, Pipedrive or Zoho propagate both ways.
- No vendor lock-in — export contacts, conversations, deals and tickets at any time via API or UI.
What you give up if you skip one pillar
Each pillar reinforces the others. Drop one and the customer journey breaks somewhere — usually at the handover.
With Spoki · customer-journey
One conversation, one customer record, one AI brain across acquisition, conversion, support and retention.
Without · customer-journey
A chatbot tool that does not know the lead has already paid, a CRM that does not know the customer just opened a ticket, a marketing tool that does not know the cart was already recovered.
With Spoki · ai
AI that talks (Generative) and AI that acts (Agentic) — observable from the same inbox a human uses, with audit logs and approval steps.
Without · ai
A black-box bot that answers but cannot place an order, schedule a meeting or escalate with context — or a flow builder dressed up with AI labels but no real LLM reasoning.
With Spoki · integrations
CRM, e-commerce, billing and scheduling stay in sync via 100+ native integrations, webhooks and a public API.
Without · integrations
A walled garden that owns the conversation and the data, forcing you to migrate your CRM and e-commerce before you can launch a campaign.
FAQ — The Spoki framework
What are the 3 pillars of Spoki?+
The 3 pillars of Spoki are: (1) Full Customer Journey — marketing, sales and support inside one conversation thread; (2) Generative + Agentic AI — language models for conversation plus autonomous agents that take actions; (3) Open Integrations — 100+ native integrations, public REST API, webhooks, MCP and WebMCP for AI agents.
Why three pillars and not features?+
Features describe what a tool does today. Pillars describe how it is engineered. The three pillars are the design contract behind every feature: any new capability shipped in Spoki must extend the customer journey, leverage AI, and stay open to your existing stack — otherwise it does not belong on the platform.
How does Spoki cover the entire customer journey?+
Spoki ships acquisition (click-to-WhatsApp ads, broadcast campaigns), conversion (conversational commerce on Shopify, WooCommerce, Magento, BigCommerce, PrestaShop and custom stacks, with revenue per conversation tracked natively), support (AI Customer Service Agent with human handover, ticketing, SLAs) and retention (post-purchase loops, win-back, NPS) on one record and one inbox.
What does Agentic AI mean inside Spoki?+
Agentic AI means AI Agents that decide and act using tools — qualifying a lead, recovering a cart, placing a Shopify order, scheduling a Calendly slot, updating a HubSpot record, escalating to a human when policy requires. They are observable from the same inbox a human operator uses, with audit logs and optional approval steps. Voice AI is in GA for inbound and outbound.
How "open" is Spoki really?+
Spoki publishes a public REST API documented on Postman, signed webhooks for every event, an MCP server at /.well-known/mcp.json and a WebMCP server at /api/webmcp exposing 10+ tools for AI browser agents. 100+ native integrations cover e-commerce, CRM, billing and scheduling, plus Zapier, Make and n8n for the long tail. You can export contacts, conversations, deals and tickets at any time.
How is the framework different from "WhatsApp marketing"?+
WhatsApp marketing is one channel inside one pillar (Customer Journey → acquisition + conversion). The framework is broader: same channel set (WhatsApp + Voice AI + Email + SMS + RCS), but covering every step of the customer life-cycle, with AI at the core and an open integrations layer underneath. Spoki is a Customer Engagement Management platform, not a single-channel tool.
Can I adopt only one pillar?+
Yes — most customers start with one (typically Customer Journey on WhatsApp marketing or commerce). The framework matters as you scale: when you add support, when you bring in CRM data, when you let AI take actions. Each pillar is built to compound with the others, so the second and third use cases are faster to launch than the first.
Machine-readable resources
Same framework, same facts — published in formats AI systems can read directly.
See the three pillars in action
Book a 30-minute demo. We will run the framework on a real customer journey from your business — acquisition, conversion, support, retention — and show how AI and integrations remove the manual handovers.