Integrations with your systems
Ready-made connectors for popular CRMs and services. We connect other systems through the API.
They answer customers from your knowledge base, qualify enquiries and create deals in your CRM, in messengers and on your website. Anything they can't handle goes to a member of staff.
One assistant can work across several channels at once: WhatsApp, Telegram and a website widget. We connect it to your CRM, accounting system, catalogue and loyalty programme. The conversation history is saved in the CRM card.
Ready-made connectors for popular CRMs and services. We connect other systems through the API.
Customers write where it suits them, and conversations are collected in one system.
We configure scenarios, handover rules and access rights around your processes.
Services you can connect the assistant to. We connect other systems through the API.
A first line of support around the clock: order statuses, terms, documents, answers to common questions.
Clarifies the need, volume and timing, identifies the customer type and passes a completed deal card to the CRM.
Selects the product or service, calculates the cost, issues an invoice and agrees the next step.
Points balance, member tier, personalised offers and repeat-purchase reminders, based on data from your loyalty programme.
Sample data
Regulations, reference guides and the company knowledge base in one window. Role-based access; sign-in can be set up through your corporate account.
The assistant in the bottom-right corner is built like the ones we make for clients: it answers from a knowledge base and, if there is no answer, passes the question to the team. Ask your own question or start with one of these:
This opens the assistant window.
Your knowledge base only. Catalogue, prices, terms and regulations are the sole source of answers. If a question falls outside the knowledge base, the assistant asks for clarification or hands it to a member of staff, with no guessing.
Rule-based handover. A complex question, an unhappy customer, a large order, non-standard terms: you set the rules. The conversation goes to a manager with its full history.
A log of every conversation. You see what was asked and what the assistant replied. We use the log to extend the knowledge base and refine the rules.
Limited permissions. The assistant sees only the data it needs and cannot change prices, terms or deal statuses outside the defined scenarios.
Vesta-Polymer supplies polymer raw materials to manufacturing companies. Customers write on WhatsApp with questions about grades, availability, prices and delivery terms, and replies must be fast and carry exact technical specifications.
We built an AI consultant that handles these conversations on its own:
from kickoff to launch
The customer asks which grade suits blow-moulding 10 litre canisters. The assistant recommends a grade, compares it with an analogue in a table and explains the difference in practice.
| Grade A | Analogue B | |
|---|---|---|
| Density, g/cm³ | 0.954 | 0.952 |
| MFR, g/10 min | 0.30 | 0.35 |
| Best for | blow moulding | blow moulding, containers |
The customer asks the price of 3 tonnes of grade A and whether they can collect it. The assistant quotes an illustrative price and gives the pickup and delivery terms.
The assistant asks whether the customer buys for own production or resale, passes the request to a manager and creates a lead card in Bitrix24 with the client type, volume and product.
The customer asks for an invoice for 3 tonnes of grade A with pickup. The assistant sends a PDF invoice valid for 3 business days and says a manager will be in touch once payment arrives.
Recreated in English from real conversations, shared with the client's permission. Product names are generalised and prices are illustrative.
Sample data
Didn't find your answer? Ask the AI assistant on this site or write to hello@oakai.online. We reply quickly.
Free consultationWe choose it to fit the task and your data requirements. Usually it is a commercial language model. If data must not leave your company, we can deploy an open model inside your environment.
It answers only from the knowledge base and hands anything outside it to a member of staff. Every conversation is visible in the log. We correct an inaccurate answer in the knowledge base, and from then on the assistant replies from the corrected version.
From your systems. We connect the catalogue, price list and accounting system, and the data updates on a schedule or in real time.
Yes. The assistant creates leads and deals, fills in fields and assigns an owner. In the Vesta-Polymer case it is Bitrix24, and we connect other CRMs through the API.
A typical launch takes 2–3 weeks. Vesta-Polymer went live in two weeks. We confirm the timeline for your project after the audit, because it depends on channels, integrations and the size of the knowledge base.
We calculate the cost of implementation after the audit and fix it for each stage. Language-model costs depend on the number of conversations, and we estimate them in advance. There is no hourly billing.
A free consultation with an engineer: 30–45 minutes, no commitment.