Ai Thinks Chatbot Platform
Build every chatbot assistant on one platform, and run them across every channel from that same place.
With the Ai Thinks Chatbot Platform you can build capable assistants with ease, connect them to your systems, publish them on the channels you choose and report on how they perform.
WhatsAppWeb widgetLive support
Platform topology
Every channel, system and scenario on one platform.
The channel your customer arrives on, the AI layer that produces the answer, the integrations into your corporate systems and the live support that steps in when needed are all part of the same architecture.
User channels
Ai Thinks Chatbot Platform
AI layer
Natural language understanding (NLU)Dialogue managementKnowledge base managementIntent & entity managementMachine learning and continuous improvement
Platform services
User & permission managementFlow managementContent managementMulti-language supportIntegration managementAnalytics & reportingSession & history managementSecurity & data protection compliance
Infrastructure
High availabilitySecure infrastructureData backupScalable architecture
Live support
Integrations
Always onCompliance readyHigh securityScalableReal-time analyticsContinuously learning
How assistants are built
A separate assistant for every need.
Companies and sub-companies are defined in the system, and each one manages its own users, assistants, files and messaging setup. When you create an assistant you pick one of three AI approaches.
Approach 1
Flow-based AI
You decide where the conversation starts and how it moves on. Built for controlled, predictable scenarios.
- A conversation flow designed step by step
- Multiple-choice routing
- Controlled, repeatable outcomes
Approach 2
Document-based AI
Upload your organisation’s files and let the assistant answer from that knowledge base. Built for recurring questions.
- Answers generated from uploaded files
- Documents assigned per assistant
- Shared knowledge through global documents
Approach 3
Agent AI
The assistant does not just answer, it acts. It calls external systems and returns the result to the customer.
- Agents driven by intent type
- Global agents or agents scoped to one assistant
- An architecture built to extend
“I want a different bot for sales, for technical support and for dealer support.”
Activation and channel selection are handled per assistant, so every team can run its own bot.
Flows
Design the conversation without writing code.
A flow is the conversation path your assistant follows. Line up the actions to build a controlled conversation, and change it whenever you need to.
Flow actions
Sample flow
Hello 👋
What can I help you with?
Order trackingProduct informationTechnical supportTalk to an agent
Build once, roll it out to every branch
The flows you create can be copied to sub-companies. A flow prepared at head office runs to the same standard in every branch, which makes central management practical for franchises and multi-branch operations.
Head office
Konya branch
Ankara branch
Istanbul branch
Document-based AI
Let the assistant learn from your own documents.
Companies upload their files and assign them to specific assistants. The assistant then answers from your organisation’s own content.
Supported file types
PDFDocumentsCSVImagesAudioVideoOther file types
Documents per assistant
Each assistant answers only from the files assigned to it.
Global documents
Keep the files every assistant may use in a single place.
Change history
When a document was added and last updated is kept on record.
Example: a school
The school uploads its information files, and questions from parents are answered straight from that content.
- How much is the enrolment fee?
- Do you run a school bus?
- What are your opening hours?
The answer comes from the documents the school uploaded.
Agent AI & API
Run your operational flows through the chatbot and go beyond simply answering.
Because an API request action can be added to any flow, the assistant talks to your external systems, runs the query and returns the result to the customer.
API request configuration
For each API action you define the request details and how the returned data will be used.
- URL
- HTTP request type
- Header
- Body
- Data source
- Field mapping
- Transform
Example: cargo tracking
- 1The customer asks“Where is my parcel?”
- 2The assistant collects the detailsIt asks for and validates the order number.
- 3The API is calledA request goes out to the courier service.
- 4The result comes backThe delivery status is passed to the customer.
Agent available on the platformCargo Tracking
An agent architecture built to extend
Order lookupAppointment managementCRM integrationProduct recommendationStock lookupReservationsPayments
These are not agents shipping in the platform today; they are the directions the current architecture allows. We define what you need together.
Channels
Be present on every channel your customer uses.
You choose where each assistant is published. The platform does more than answer inbound messages: it also runs your campaign and notification messaging.
Runs on the Gupshup WhatsApp provider, with WhatsApp Business message templates managed inside the platform.
- Body text, footer and button definitions
- URL, phone, quick-reply and coupon-code buttons
- Static or dynamic URLs
Template status
PendingApprovedRejected
Website widget
Embed the chatbot in your own site with your brand identity. The widget can also collect visitor details as it opens.
- Background, gradient, text and footer colours
- Sent and received message bubble styling
- AI icon and widget icon selection
Lead form fields
First nameLast nameEmailPhoneUser ID
Bulk & push messaging
Message a single user, a group you pick, or everyone, and include sub-company users in the scope.
- User selection and sub-company scope
- Sending, sent and failed states
- Tracking through message history
Delivery status
SENTDELIVEREDREAD
Live support
When the bot cannot resolve it, an agent takes over.
The platform has a dedicated live support module. The moment an agent sends a message into the conversation, the active bot is switched off and the human continues.
Hand-off path
- Customer
- AI bot
- Not resolved
- Live support
- Support agent
The bot is switched off automatically as soon as the agent replies.
Messages and files
Agents can send messages and files into the chat.
Hold and close
Put a conversation on hold, close it or archive it.
Tagging
Tag the conversation and filter by company and tag.
Conversation tags
Create your own tags and set their text and background colours. Conversations can be filtered by company and by tag.
SalesComplaintTechnical supportVIPAwaiting quoteReturn
Conversation states
OngoingOn holdClosedArchived
Teams
Create teams and assign members to mirror your support organisation inside the platform.
- Sales team
- Support team
- Technical team
- Finance
What it is used for
One assistant, working towards three different goals.
A chatbot does more than answer questions: it starts the transaction, announces the campaign and collects the feedback. We plan your setup around these three headings.
Information & support
- Frequently asked questions
- Product / service information
- Order and delivery tracking
- Account / membership operations
- Technical support
Transactions
- Placing an order
- Buying a product or service
- Bookings and appointments
- Returns and exchanges
- Payments & invoicing
Marketing & engagement
- Campaigns & announcements
- Personalised recommendations
- Surveys & feedback
- Customer segmentation
- Loyalty programmes
What it adds up to
When the three run together, this is what it means for the operation.
- An always-on customer experience
- Higher customer satisfaction
- Operational efficiency
- Lower cost to serve
- More sales and stronger loyalty
Dashboard & analytics
The whole operation on one screen.
Track company, sub-company, user, assistant and conversation counts from a single panel, and follow the last seven days of conversations and notifications, the chats in progress and your conversation analytics from the same place.
Overview
Total users
12,480
9.4% ↑
Companies & sub-companies
86
6.2% ↑
AI assistants
54
11.8% ↑
Ongoing conversations
128
7.5% ↑
Last 7 days
4003002001000
MonTueWedThuFriSatSun
ConversationsNotifications
Assistant type mix
46%Document
- Document-based46%
- Flow-based34%
- Agent AI20%
Ongoing conversations
- WhatsApp · +90 533 *** ** 1200:01:24Cargo tracking
- Web widget · Visitor #482100:00:46Product info
- Live support · Agent took over00:03:12Technical support
View all →
Solutions on this platform
Ready-made solutions running on the Chatbot Platform
You do not have to build on the platform from scratch. Start from scenarios already proven in the field and adapt them to your own flow.
Chatbot Platform
Let us design the chatbot assistant that fits you best, together.
We go through your documents, your flows and your integration needs together, then set out a deployment plan built around them.