Integration of an AI agent in Manychat for processing incoming messages

Budget: 8500 UAH Deadline: 8 days

Good day! I connected ManyChat + GPT through External Request: a custom webhook maintains the dialogue context and decides when to provide the phone number, while ManyChat remains on subscriptions and the catalog. Bitrix24 was connected via an incoming webhook, and the lead with the phone number and all correspondence goes directly into the funnel. The main question: should the car selection agent take from your catalog in ManyChat or is it possible to browse auction lots? This affects almost the entire volume. One week for a working version, then we refine it based on live dialogues.

Budget: 13300 UAH Deadline: 7 days

Good day, I have over 5 years of experience in development, often implementing AI automation for chat processing and integration with CRM.

I will create an agent on the OpenAI API, with the logic of a manager's dialogue and integrations in ManyChat and Bitrix24. Focus on lead transfer and contact acquisition; I have experience with similar tasks for agency services.

I aim to have the first working version ready within a week. I will provide the exact cost after a few clarifications or in private messages.

What functionality in Bitrix24 do you use for lead processing - a simple pipeline or a more complex scenario?

Feel free to reach out!

Budget: 6000 UAH Deadline: 5 days

Good day! I will create an AI agent that will serve as the first line in your chats: it conducts conversations in the logic of a manager, clarifies inquiries about cars, collects phone numbers, subscribes clients to the necessary ManyChat automation, and transfers leads to Bitrix24 along with the information extracted from the conversation.

How it works technically. The agent operates as a separate service on our VPS (Python + OpenAI or Claude), ManyChat calls it via External Request and receives a ready response plus tags for subscription. Bitrix24 — through an incoming webhook: a lead with the source, car model, budget, and phone number. Your current flows in ManyChat remain in place; the agent picks up the free dialogue where the scenario ended and returns the person back to the needed branch.

Based on my experience with your stack: LLM agents, chatbots, API integrations, and CRM hooks — this is what I work with constantly.

I handle the three entry points from the brief differently: catalog, free text, and transition from advertising to a specific car. In the latter case, the agent must immediately see which lot is being referred to; otherwise, the dialogue starts from scratch, and the client drops off.

The question at hand: do you have an up-to-date database of lots and prices that the agent can access (table, feed, auction API)? Or does it initially work with knowledge of the process and guide the client to the manager — this affects how deeply it can calculate the cost.

The price is 6000 UAH, the deadline is 5 days. I am ready to show live bots in the chat so you can see how the dialogue is maintained.

Petro Pankov, BotCraft Group Expand

Budget: 10000 UAH Deadline: 3 days

I have experience in creating AI agents for call handling, I can do it quickly and efficiently.

Budget: 25000 UAH Deadline: 18 days

Good day!

In such projects, the challenge is not connecting AI to Manychat — that’s a one-day job. The challenge is ensuring that the agent does not make up numbers.

Calculating the cost of a car from the States or Korea involves the auction price, delivery, customs duties, excise, VAT, and exchange rate. If you allow the language model to calculate this "from memory," it will confidently provide the client with a figure that may differ from the actual amount by thousands of dollars. The client will remember that specific number, while the manager will sort it out.

Therefore, the architecture is as follows: Manychat remains the top layer — channels, subscribers, scenarios. All the logic resides in a separate service that Manychat calls via webhook. The cost calculation is done not by the model, but by your formula in code. The model only understands the request, retrieves the necessary data, and formulates a response nicely.

Another aspect that is often overlooked is the boundary for handing over to the manager. The agent must clearly know when to stop: atypical questions, negotiations, specific lots, complaints — immediately to a person, with the gathered context of the conversation. This is much more important than beautiful answers: an agent who timely hands over complex issues to a person saves time. An agent who does not hand over costs clients.

What I propose in the first stage: client qualification (budget, country, type of car, deadlines), responses to recurring questions based on your materials, cost calculation according to your formula, and handover to the manager according to the rules. Searching for lots at auctions is a separate stage, as it all depends on what access you have to the data.

Two questions to avoid guessing: do you have a ready-made cost calculation formula in any form — table, calculator, CRM? And do you have official access to auction data?

I work with Python (FastAPI) and React/TypeScript, creating bots and integrations with external APIs — this is exactly the type of task. Expand

Budget: 10000 UAH Deadline: 7 days

This is a specialized task for us; we are gathering exactly such AI agents for processing incoming requests (tyaplyapai.com).

There is an important point. ManyChat itself cannot conduct a free conversation; it operates based on predefined scenarios and buttons. To enable the AI agent to respond in a human-like manner and understand the request, it needs to be separated out to n8n, while ManyChat simply calls it and inserts the response into the chat. Otherwise, a live dialogue will not be possible.

For your three entry options (catalog, manual request, advertisement for a specific car), the AI agent will:
1) conduct a conversation like a selection manager, clarifying what exactly is being sought and helping to choose;
2) take the phone number and sign up for the necessary automation in ManyChat;
transfer the lead to Bitrix24 along with the contact, request, and the source from which it came;
answer questions regarding selection, delivery, and customs clearance.

The timeline is approximately 7-10 days.

Please let me know: where are the conversations currently taking place? Expand

Budget: 8000 UAH Deadline: 7 days

Good day, Oleksandr.

Key for you: I have already built a bot that pulls auto lots from the Korean Encar through its private API, filters by 10 criteria, and automatically checks for hidden accidents, flooding, and rental history, then delivers a ready car card to the client. You import from Korea, the USA, and China, so "smart car selection through API" is not a theory for me, but already completed work. This is something that no one has in the bidding feed.

Regarding ManyChat, I will open the cards right away: it does not allow LLM to freely conduct a dialogue on its own; its flow goes step by step. Therefore, the agent lives on an external webhook (my service + OpenAI/Claude), and ManyChat calls it for a response while maintaining the subscription and catalog. This way, the agent conducts a natural dialogue, does not lose the thread on any of the three inputs (catalog / free text / transition from advertising), collects the number, and cleanly passes a qualified lead to Bitrix24.

In addition, I have ready developments for automation: parsing and structured car cards, deduplication of lots, data collection for CRM. This can be rolled out step by step, from a basic consultant to selection from several markets and retraining on real dialogues of managers, so you do not pay for the maximum all at once.

The price is 8000 UAH, the term is 7 days. I am ready to show in the chat how the dialogue is maintained and how the car card comes in. One clarification below in the discussion.

Budget: 14000 UAH Deadline: 7 days

I can implement an AI agent for processing incoming messages in ManyChat with subsequent lead transfer to Bitrix24.
What I will do:
• AI that conducts dialogue as closely as possible to the logic of a manager
• Collection of phone numbers and client qualification
• Subscription to the necessary ManyChat automation
• Transfer of the lead to Bitrix24 with all data
• Context support (catalog / specific car / free request)
Stack: Python (or Node) + OpenAI/Claude + ManyChat API + Bitrix24 REST API.

Before starting, I will clarify the integration capabilities specifically with your ManyChat account and propose the optimal architecture (via External Request or through an intermediate server).
I am ready to discuss the details. Expand

Budget: 999 UAH Deadline: 1 day

Good day, Oleksandr.

Automating lead qualification and customer retention in dialogue involves building a scenario that replicates the logic of a live employee: from clarifying needs to obtaining contact information. Your funnel with a catalog, manual inquiries, and transitions from advertising requires a single contextual agent that does not lose the thread of conversation through any of the paths. Such integrations with telephony and chat channels have proven that structured dialogue collects contacts more effectively than template responses, and the transfer of leads to CRM occurs without losses.

Stages:

Each stage is documented, and the code is handed over to you.
You can stop after any stage.

≈6 days: 5500–7500 UAH
Creation of a basic AI consultant for lead qualification and contact collection.
- Automatic collection of phone numbers and basic customer needs
- Integration with ManyChat for responses to common questions
- Transfer of qualified leads to Bitrix24
- Setting up subscriptions for basic ManyChat automations

≈7 days: 6000–8000 UAH
Full-fledged AI agent with integration of car data and calculations.
- Intelligent car selection based on customer requests via API
- Automatic calculation of the estimated cost of the car
- Natural dialogue for customer retention using LLM
- Full cycle: from the first message to the customer card in Bitrix24
- Flexible logic for transitioning from advertising to specific lots

≈7 days: 6000–8000 UAH
Maximum automation with deep analytics and support for multiple markets.
- Advanced car selection from three markets (USA, Korea, China) with comparison
- Complex multi-stage qualification with segmentation in Bitrix24
- Continuous AI retraining system based on real dialogues of managers
- Priority support and conversion optimization during the first month
- Integration with the ManyChat catalog for instant synchronization of stocks/lots

Required experience:
- Vibe Coding Platform — Your requirement for business process automation and AI integration is covered: this project uses n8n and Claude for rapid creation of automated workflows and MVPs.
- Rule-Based Telegram ChatBot — Your need for processing incoming messages and lead qualification is covered: development of bots for lead generation and business automation using messenger APIs.
- KeepinCRM Automation — Your requirement for data transfer to CRM (Bitrix24) is covered: this project implements full automation of the sales funnel in CRM based on external APIs and order statuses.
- More here: https://valpere.github.io/uk/portfolio/ Expand

Budget: 1000 UAH Deadline: 7 days

Hello!

I have practical experience in creating AI solutions for business and automating customer interactions. I work with OpenAI, Claude, Gemini, ManyChat, Bitrix24, and API integrations, so I can implement an AI agent that will take on a significant part of communication with clients.

The AI agent will be able to conduct natural dialogue, clarify needs, answer questions, assist with car selection, keep the client engaged in conversation, obtain phone numbers, trigger necessary automation in ManyChat, and transfer leads to Bitrix24. Integration with catalogs and auctions via API can also be implemented so that the agent provides up-to-date information and assists with preliminary cost calculations.

Before starting work, I will analyze your current process, propose the optimal logic for the AI's operation, and implement a solution that closely resembles the work of a live manager.

I would be happy to discuss the project details and suggest the best implementation option.

Budget: 15000 UAH Deadline: 6 days

ManyChat does not allow the AI agent to conduct a dialogue on its own — its logic follows the steps of the flow. Therefore, the agent operates on my side: ManyChat sends it messages via External Request and retrieves the response plus a tag to sign the client.

FastAPI + Claude: the agent reads the conversation history, maintains the manager's line, extracts the phone number and the need for a car, and then creates a lead in Bitrix24 with the country, model, and budget. The catalog, free text, and transition from advertising come in through one entry with different starting data. You will rewrite the working agent in the test flow of ManyChat by the fifth day.

I created the AI assistant for the CHARACTER park after the unsuccessful bot from previous performers — I specifically rewrote the dialogue logic so that it wouldn't resort to general phrases. I have no reviews on the marketplace, so the money is held in escrow until your acceptance.

15,000 UAH, working version in 6 days, revisions based on your comments — up to 8 days. In Bitrix24, check whether you accept leads into the funnel or directly into deals. Once I see a response, I will compile a map of the fields that the agent fills out automatically.

Budget: 27000 UAH Deadline: 15 days

Your team is spending managers' time on manual processing of inquiries in chats, selecting cars, and qualifying clients. I can create an AI agent that will conduct conversations close to the logic of a live manager. It will collect phone numbers, subscribe clients to the necessary automation in ManyChat, transfer leads to Bitrix24, and answer questions. I will connect the agent to your catalog and scripts so that it correctly processes inquiries from the catalog, manual requests, and transitions from ads to specific cars. I will set up API integrations between ManyChat and Bitrix24 and describe the logic of qualifying and retaining clients in the dialogue. I am ready to discuss the details of the scripts and start implementation.

Budget: 25000 UAH Deadline: 5 days

Hello, Oleksandra.
I have done similar projects — I know the nuances and will do exactly what is described. There are a few clarifications before starting.
The final price/timeframe will be determined after all clarifications.

Profile: Freelancehunt
Reviews: Freelancehunt