Azure-Native

AI Assistant Hub for Websites & Business Channels

Turn your website into a multilingual digital workforce — with AI assistants for Sales, Support, Customer Service, and HR, plus optional email automation.

This is not a generic chatbot. We start from a proven core platform, then customize the UI, content, workflows, and RAG knowledge architecture to match your business.

What it does

Answers customer questions 24/7 in their language

Uses your approved knowledge (PDF, Word, Excel, internal pages, databases)

Routes requests to the right team (Sales / Support / HR)

Captures leads with smart forms and conversation insights

(Optional) Reads inbound emails and drafts replies for approval

(Optional) Voice: speak & listen on the website

Department-specific AI agents

πŸ“ˆ

Sales Agent: qualifies leads, suggests next steps, drafts proposals and follow-ups

πŸ› οΈ

Support Agent: technical help based on your manuals, FAQs, and ticket history

πŸ‘₯

HR Agent: internal policy Q&A, recruiting intake, candidate pre-screening

πŸ“§

Email Agent: reads inbound emails, drafts replies, routes to owners — human approval before sending

Industry Edition — Hospitality

🏨 AI Guest Concierge for Hotels

A bespoke edition of AI Assistant Hub, built per property

The same platform that powers AI Assistant Hub, tailored for hotels and tourism: a bilingual (Turkish/English) guest concierge that answers from YOUR hotel's own knowledge base — rooms and rates, restaurants, spa, transfers and policies — with source-cited answers, live availability awareness and graceful hand-off to reception. Every deployment is created specifically for one property: your brand, your documents, your rules.

  • βœ“ Answers in the guest's language (Turkish / English), 24/7
  • βœ“ Grounded in your hotel's documents via Azure AI Search RAG — with source citations
  • βœ“ Room types, rates and availability awareness; booking hand-off to reception
  • βœ“ Fully branded per property — colors, tone, policies, knowledge base
  • βœ“ Runs on the same secure Azure-native platform core (Key Vault, content safety, audit)

See it live: our demo property "Grand Bosphorus Palace Hotel" runs this edition end-to-end — text and live voice.

Ask our AI about the Hotel Edition
Industry Edition — Real Estate

🏘️ AI Property Advisor for Estate Agencies

Grounded in your live listings, not in documents

The same platform, for a business whose knowledge is inventory rather than paperwork. A bilingual (Turkish/English) advisor reads your live listing database: "a one-bedroom rental in Kadıköy between 70 and 90 square metres" returns the listings that actually match, each with its reference code — by text or by voice. Agents manage listings from a status board; customers search and browse on your site.

  • βœ“ Understands a request in plain language — district, budget, size, rooms, rent or sale
  • βœ“ Answers only from listings that exist right now, cited by reference code
  • βœ“ Admin panel for listings with a status board (active, reserved, sold, rented, withdrawn)
  • βœ“ The same advisor by voice — the customer can simply describe what they want
  • βœ“ Connects to your existing listing system; the assistant never sees where the data came from

See it live: our demo agency "Marmara Estates" runs this edition end-to-end.

Ask our AI about the Real-Estate Edition
Included in every deployment

🧠 Admin Agent — the assistant that checks its own work

Several AI vendors, one question, one set of your documents

Every YazılımAI deployment includes an internal assistant for your own team — grounded in your internal documents rather than in a model's general knowledge. What makes it different from ChatGPT, Claude or Microsoft Copilot 365 is that it is not tied to a single AI vendor: one deployment offers models from Azure OpenAI, Microsoft Foundry, Anthropic and Google side by side.

Two things a single-vendor assistant cannot do

βš–οΈ

A second opinion from a competitor

An answer written by one vendor's model is checked by another vendor's model against the same documents, and rewritten if the reviewer finds a claim your documents do not support. The reviewer's notes are shown, so you can see what was objected to and why.

πŸ”€

The same question, side by side

Put one question to up to four models at once, all reading the same documents, answers in parallel columns. Because the evidence is shared, a disagreement between columns is the models disagreeing — not one of them having read something else.

Answers you can check, and costs you can see

  • βœ“ Clickable citations — press a number and read the exact passage the sentence came from
  • βœ“ A cited source that does not exist is refused rather than made to look real
  • βœ“ Token and cost reporting per answer, and a 30-day total by model
  • βœ“ Answers stream as they are written, with a stop button that keeps what has arrived
  • βœ“ Attach a PDF, Word, Excel or text file to a question — nothing is stored
  • βœ“ Reads the web when asked, searching in English whatever language you ask in

πŸ”’ Your internal documents live in a separate index no public assistant can reach, vendor keys never leave Key Vault, and a model with no key configured is never called at all.

Multi-Agent & Model-Agnostic Architecture

Optional multi-agent orchestration: stay model-agnostic — use the GPT-5 family (e.g., GPT-5.2 / GPT-5.1) and, where licensed, other models (e.g., Claude) behind an orchestration layer / Microsoft Agent Framework, with the option to add more models over time.

GPT-5.2 GPT-5.1 Claude Model-Agnostic Microsoft Agent Framework

"Grounded" answers (reduced hallucinations)

With RAG (Retrieval-Augmented Generation), the assistant answers based on your company's sources — and can return citations to the original documents.

Enterprise security & privacy

Runs on Microsoft Azure (single-tenant per client if required)

Role-based access for admins and content owners

Encryption at rest and in transit

Optional Private Network / Private Endpoints

Auditing, logging, and monitoring (Application Insights)

For Technical Teams (Architecture Overview)

Core building blocks (typical):

  • Azure OpenAI / Azure AI Foundry for LLMs (model-agnostic, latest GPT family supported)
  • Azure AI Search for retrieval + hybrid search
  • Azure Storage for document management (admin upload & categorization)
  • Azure Document Intelligence (optional) for PDF extraction & structuring
  • Orchestration layer / Agent framework (multi-agent routing, tool use, workflows)
  • .NET / Blazor front-end + Web API backend
  • Azure SQL / SQL Server for business data and app state
  • Azure Key Vault for secrets and keys

Key capabilities:

  • Knowledge segmentation by department, product line, region, or customer type
  • Admin console for upload, taxonomy, re-index, and content lifecycle
  • Conversation memory (safe), summarization, lead scoring, escalation rules
  • Email workflow integration (e.g., Microsoft Graph) with approval gates
  • Observability: trace prompts, citations, latency, and cost controls

RAG Architecture for Website Assistant

Our website AI assistant is powered by a production-grade Multi-Agent system built on Microsoft Agent Framework.

Microsoft Agent Framework β€” Orchestrator
Tool Gateway RBAC / Approval / Audit
πŸ“ˆ Sales
πŸ› οΈ Support
πŸ‘₯ HR
πŸ“§ Email
AI Search
Azure SQL
Blob Storage
Azure OpenAI

Why Microsoft Agent Framework?

  • Native C#/.NET — first-class integration with Blazor and ASP.NET Core
  • Official Microsoft Multi-Agent framework with stable API
  • Built-in Plugin system for Tools (RAG, SQL, APIs)
  • Multi-Agent orchestration with intelligent routing and hand-off
  • Tool Gateway with RBAC, approval gates, and audit logging

Specialized Agents

  • Sales Agent: service inquiries, pricing, lead capture, demo scheduling — with RAG over company knowledge base
  • Support Agent: technical help, documentation search, customer history from Azure SQL, ticket creation
  • HR Agent: resume analysis, candidate comparison, hiring workflow with human-in-the-loop approval
  • Email Agent: inbound email processing, draft replies, routing to owners — human-in-the-loop approval before sending

Shared Infrastructure

πŸ” Azure AI Search — vector + hybrid search for RAG retrieval
πŸ—„οΈ Azure SQL Database — business data, customer history, conversation state
πŸ“‚ Azure Blob Storage — document management with admin upload and categorization
πŸ€– Azure OpenAI — GPT family models for LLM inference
πŸ›‘οΈ Tool Gateway — RBAC enforcement, allowed-tools policy, human approval gates, audit logging

Delivery Approach

1

Discovery

goals, data sources, languages, security constraints

2

MVP

core assistants + RAG + admin upload + reporting

3

Production

hardening, governance, monitoring, scaling, rollout

Ready to transform your website?

Let's discuss how AI Assistant Hub can work for your business.

Get in Touch