Your CRM now talks: ask it, build it, let it analyze
Type “How many new customers did we get this month?” and get the answer from your data. Build a chatbot with the visual builder and let it talk to your customers across 7 channels. And let AI read every incoming conversation: who is close to buying, who is about to complain — with their scores right in front of you.
Starts at 2.500 ₺ / month
7-day free trial · No setup fee · Upgrade your plan any time
Scanning and interpreting your data by hand also has a cost — and it isn't billed to the software, it's billed to your time.
If your time is worth ₺800 an hour: the monthly cost of your subscription equals 4 hours of work.
Not one AI, but three layers of AI
Most CRMs say “artificial intelligence” but mean a single chat box. In ÖzgürKod, AI works in three separate layers. The assistant layer serves you: you ask questions about your CRM data in natural language; you pull customer counts and sales summaries with date range, segment and source filters. There are four ready-made report templates you can run with one click: Sales summary, Customer analysis, Pipeline status and Sales rep performance.
The chatbot layer talks to your customers: the dialogues you design in the visual flow builder run 24/7 on WhatsApp, Instagram, Facebook, Telegram and web chat. The analysis layer reads for you: it summarizes incoming conversations, produces four separate scores and suggests “should a human take over this conversation?”. All three layers run on the API key you define, in your isolated database — your data never mixes with other accounts.
Design chatbots with 13 step types
Not a simple question-and-answer bot; flows that branch, think and hand off. Drag steps and connect them:
- Basic steps: message, question, condition (if/else), wait, time check
- AI steps: intent-based AI routing, sentiment routing, answer from the knowledge base
- Advanced steps: external API call (http), transfer to another flow, visitor/lead check
- Human handoff: when the bot can't resolve it, the conversation moves to a sales rep; who paused it and why is on record
- Practical tools: import/export the flow as JSON, have AI draft one with “Auto Flow”, watch live sessions
4 scores and 1 recommendation for every conversation
AI reads your WhatsApp conversations and reports each one in numbers — so anyone looking at the dashboard in the morning knows what to spend the day on. The scores are an estimate, not a precise measurement: they are there to guide your priorities, not to replace your judgment.
Sentiment score
Is the customer happy or tense? The sentiment at the start of the conversation and at the end are tracked separately.
Sales potential
How close to buying? High-scoring conversations rise to the top of the list.
Complaint risk
Catch a crisis before it grows: at-risk conversations are flagged before they turn into complaints.
Urgency
A customer who says “I need it today” doesn't get lost in tomorrow's call list.
You're in control: model, cost, channel
Multiple AI providers
Google Gemini, OpenAI, DeepSeek and others: choose the provider; set the model name, creativity (temperature) and response length per account.
Turn on/off per channel
Manage automatic replies channel by channel: on for WhatsApp, off for Instagram. For campaign recipients, the bot goes quiet on its own for 24 hours.
Quota & cost control
It runs on your own API key; you set the daily analysis quota (default 50 conversations). Automatic analysis turns off with a single key.
Frequently asked questions
What questions can it answer?
Questions about your CRM data: customer counts (filtered by date, segment and source), sales summaries, pipeline status, sales rep performance. Four ready-made report templates run with one click; free-form questions are answered with the same data tools.
What languages does the chatbot speak?
You can build flows in Turkish, English and German. Each flow targets a platform: all channels or only the channel you choose (WhatsApp API, WhatsApp QR, Instagram, Facebook, Telegram, web chat).
What if the bot gives a wrong answer?
It can. AI is probabilistic; the measures we take reduce the chance of error but do not eliminate it. The flow includes a “hand off to a human” step — when the bot cannot resolve something, it transfers the conversation to a sales rep and a notification is sent. You can pause the bot manually in any conversation; who paused it and why is logged. Human verification is required on critical matters: disclaimer.
Is my data used to train AI?
Your data stays in your own isolated database and is processed only with the API key you define, only for your queries. There is no model or data sharing with other customers.
Let’s be honest about AI
AI works probabilistically: despite all the measures we take, the bot may produce information that is wrong, incomplete or not actually true. What the bot says is not a statement by ÖzgürKod; human verification is essential on critical matters.
Conversation scores and AI assistant answers are decision support, not decisions. Human verification is required on matters with real consequences such as binding pricing, contracts, payments and medical or legal guidance.
The full disclaimer — Terms of use, section 8Turn on your AI layer today
Enter your API key and get your first ready-made report in 1 minute.
View plans