Back to all solutions
Retail & Distribution

Web, app, and store - one agent. From discovery to checkout, online and offline without seams.

Conversational shopping, function-calling tools, agent takeover, and native kiosk/signage. The DAG engine runs campaigns, content, delivery, and retargeting as an autonomous loop.

Overview

Retail and commerce face online/offline disconnects, no conversational shopping, no in-store AI guidance, and manual campaign loops - all at once. Existing AI builders stay on web alone, with no kiosk, signage, or STTS voice support, and simple FAQ chatbots cannot actually execute inventory, order, or payment APIs.

AICLUDE combines a Function Calling Tool Registry, agent takeover, STTS, and native kiosk/signage in a single platform. Web widget, Slack, social, kiosk, and signage all connect to the same agent - the customer journey never breaks.

Key Capabilities

Function Calling Tool Registry

Register and invoke inventory, coupon, order, payment, and shipping APIs as agent tools. External APIs run from a natural-language request.

Agent Takeover + Coaching Mode

Human intervention via the agent takeover mode and real-time coaching - human agents step in naturally on top of AI responses.

Native Kiosk & Signage

Unified kiosk and signage operations built in - an area no competitor supports.

STTS Voice Conversation

Real-time speech recognition, AI reply, and voice response in one cycle. Native multilingual (KO/EN/JA).

Persona-Based Hyper-Personalization

Recommendations based on customer preferences and history via the persona & customer context engine.

DAG-Based Autonomous Campaign Loop

The DAG autonomous execution engine runs segment → content → delivery → performance → retargeting end to end.

Omnichannel

Web, widget, Slack, Discord, API, MCP, social, kiosk, and signage.

Case Stories

Self-contained Application Scenarios

Every case is shown in full: Pain, AICLUDE Apply, Scenario, Impact, and Tech: without collapse.

Case 01

Conversational Shopping Agent

Customer Pain

  • Existing AI chatbots only handle simple FAQs - no inventory, order, or payment execution.
  • When a customer asks "check red size M stock → apply coupon → checkout" in one go, they must switch between app and web.
  • For complex inquiries, the handoff to a human agent breaks.

AICLUDE Apply

  1. 1Train product catalogs, FAQs, and policies via Graph N-hop Agentic RAG.
  2. 2Register inventory, coupon, order, and payment APIs as Function Calling tools.
  3. 3The 8-stage Fortress Pipeline autonomously runs "understand intent → execution plan → execute → verify".
  4. 4Complex inquiries seamlessly hand off via agent takeover.
  5. 59 AI security guards defend against prompt injection and data leakage.

Scenario

Show Me
Input · Customer
"Got red size M? Apply the coupon and check out."
  1. 01Step
    Intent capture:product, size, color, coupon, and payment recognized at once
  2. 02Step
    Pre-check:coupon validated
  3. 03Retrieve
    Stock lookup:inventory API queried
  4. 04Step
    Plan:order
  5. 05Step
    apply coupon
  6. 06Step
    checkout
  7. 07Execute
    Run:APIs called in sequence
  8. 08Deliver
    Reply:order number and shipping info delivered

Impact

  • Reduced average handle time.
  • Higher AI auto-resolution rate.
  • Lower agent labor cost.

Tech

  • Function Calling Tool Registry
  • Graph N-hop Agentic RAG
  • Fortress Pipeline 8 stages
  • Takeover
  • 9 Security Guards
Case 02

Autonomous Marketing Campaign Loop

Customer Pain

  • Plan → content → delivery → performance → retargeting are siloed across departments.
  • Rule-based MA tools require re-designing the workflow for every new campaign.
  • Per-segment variant frequency is too low - coverage gaps.

AICLUDE Apply

  1. 1Auto-build execution plans via the DAG autonomous execution engine.
  2. 2Train on personas and CRM history with RAG → auto-generate segments.
  3. 34 image providers + multi-LLM router generate per-channel content in parallel.
  4. 4Multichannel delivery via email, web, and app tools.
  5. 59 AI security guards + AES-256-GCM PII encryption protect customer data.

Scenario

Show Me
Input · Marketer
"Fall promo for women aged 20-30"
  1. 01Execute
    DAG engine builds an execution plan
  2. 02Generate
    Persona RAG auto-generates the 20-30 female segment
  3. 03Generate
    3 image variants generated in parallel
  4. 04Deliver
    Per-channel copy tones split (per social, email)
  5. 05Deliver
    3 social channels + email scheduled delivery
  6. 06Step
    Auto-applied to the next campaign

Impact

  • Shorter campaign lead time.
  • Multichannel consistency maintained.
  • Performance learnings auto-reused.

Tech

  • DAG Engine
  • Persona RAG
  • 4 Image Providers
Case 03

In-Store AI Concierge

Customer Pain

  • Demand for in-store AI guidance and content automation is high, but existing AI builders only support web.
  • No kiosk, signage, or voice concierge.
  • Insufficient multilingual (KO/EN/JA) store staff.

AICLUDE Apply

  1. 1Native kiosk and signage with unified operations built in.
  2. 2STTS voice conversation (real-time recognition, AI reply, voice response) native in KO/EN/JA.
  3. 3Train product, event, and store info via RAG + execute booking and loyalty APIs through Function Calling.
  4. 4Audit Log records all customer data access.

Scenario

Show Me
Input
Japanese tourist (in Japanese): "Any stores with events today?"
  1. 01Step
    Real-time Japanese speech recognition
  2. 02Retrieve
    Event RAG search
  3. 03Step
    Voice reply in Japanese + map and items on the kiosk screen
  4. 04Step
    Customer:"Can I book it?"
  5. 05Execute
    Function Calling invokes the booking API
  6. 06Step
    Reservation confirmed

Impact

  • Multilingual service starts instantly.
  • Reduced burden on store staff.
  • 24/7 guidance available.

Tech

  • STTS
  • Native Kiosk & Signage
  • Graph N-hop Agentic RAG
  • Function Calling
  • Native KO/EN/JA
Case 04

Omnichannel Customer Journey

Customer Pain

  • Online and offline AI run on separate solutions - customer experience breaks.
  • UX friction from re-searching in store for items already viewed online.
  • Membership, points, and purchase history are split per channel.

AICLUDE Apply

  1. 1Web widget, app, Slack, social, kiosk, and signage share the same agent and RAG context.
  2. 2Physical multi-tenancy isolates brands and stores + PII encryption.
  3. 3Personas apply identically across channels via the persona engine.
  4. 4SSE real-time notifications sync state across channels.
  5. 5Once logged in, the in-store kiosk immediately sees the web history.

Scenario

Show Me
Input · Web
customer shows interest in 3 fall coats
  1. 01Step
    (same persona context) Enters store (membership recognized)
  2. 02Step
    kiosk:"Want me to check store stock for the 3 you viewed online?"
  3. 03Execute
    Function Calling queries in-store inventory
  4. 04Audit
    Signage plays video for the 2 items in stock
  5. 05Step
    Fitting booked + points earned

Impact

  • Reduced online/offline UX friction.
  • Higher membership return rate.
  • Unified performance tracking across channels.

Tech

  • Omnichannel
  • Persona RAG
  • SSE Real-time Notifications
  • Physical Multi-tenancy
  • Function Calling

Apply AICLUDE to this industry

We shape each PoC around your data, security requirements, and operating flow.