Budget: 250 USD Deadline: 5 days
Adi Yancher, good day!
I have 3+ years of experience working with Make.com, AI API (Claude/ChatGPT), Zoho CRM. I am ready to improve your scenario: add double-checking of PDF reports in Hebrew, generate Excel with three sheets, and create tasks in Zoho CRM with accuracy assessment.
Can we discuss the details?
Budget: 200 USD Deadline: 5 days
Hello! My name is Vladislav. I have been automating complex business processes in Make.com for over 2 years. I have experience working with financial data and multi-level verification through AI (GPT-4o / Claude 3.5 Sonnet).
My vision for solving your task:
Double verification (Claude + GPT): I will implement the "arbitration" logic. Claude performs the main data extraction, while GPT checks them against logical rules (for example, the sum of payments cannot exceed the principal of the loan). If the data diverges, the system marks the report for manual verification.
Working with Hebrew (RTL): I know how to work with Hebrew encoding in JSON requests to AI to avoid "mirroring" of numbers and text.
Excel generation: I will use special Make modules or Microsoft Graph API to create a complex Excel file with three sheets. This will preserve the structure of "Details / Data / Summary" without formatting errors.
Zoho CRM: I will set up automatic task creation with full context: I will attach files (PDF + Excel) and output the "Accuracy Score" directly in the task description for the manager.
Resilience: I will add error handling blocks so that in case of a failure of one API, the script does not stop but sends a notification.
Message me privately, I will review your instructions and immediately propose the final script scheme. I am ready to implement a reliable solution that minimizes manual work.
Budget: 250 USD Deadline: 3 days
Hello, I specialize in automations in make.com and am ready to improve your existing workflow. I have enough experience working with tools like Claude, ChatGPT, and integrating them into make.com. I am waiting for you to provide a PDF example so I can suggest the optimal option for completing this project.
Budget: 225 USD Deadline: 7 days
Hello! I am working on automations through the service Make.com. I have reviewed your task. Considering the specifics of Israeli bank reports (RTL, Hebrew, complex tables), I propose the following solution:
Receiving and preparation: The Email module receives the PDF. Next, we use PDF.co to render each page of the PDF into a high-resolution image (JPG/PNG), using a vision approach instead of regular text parsing. This is critical for Hebrew, as regular text extraction from PDFs often disrupts the order of words and numbers.
Initial analysis (Claude 3.5 Sonnet): We pass the images to Claude. This model is currently a leader in visual document analysis and working with Hebrew. It will extract the data and structure it in JSON.
Double-check (GPT-4o): The obtained JSON and text are sent to ChatGPT. The second neural network acts as a "controller": it checks the mathematical logic (sum of parts = total) and compliance with your instructions. The output is a "Confidence Score."
Excel generation: To avoid using expensive third-party converters, we use the Google Sheets API as an intermediate engine. We fill in 3 sheets in the Google Sheets template. We download the file through the Google Drive module in .xlsx format. We delete the temporary file.
Integration with Zoho CRM: Creating a task through the standard Zoho module. Attaching the generated Excel file.
Can you provide 1-2 reports for testing?
Budget: 150 USD Deadline: 7 days
Good day!
I can improve the existing script in Make.com for analyzing PDF reports and automatically creating Excel files.
I would be happy to discuss the details to propose the optimal solution.
Budget: 350 USD Deadline: 7 days
I fully understand your problem: it is necessary to make the process of analyzing mortgage reports more reliable through double-checking and Excel automation. My approach: to use Make.com as the main platform for coordinating the process, integrate the Claude API for initial text analysis (with support for Hebrew and RTL), and the ChatGPT API for verifying the accuracy of the results. For text processing, I will use scripts that support right-to-left text and Unicode encoding. The Excel file will be created through Make.com or an additional script that generates three sheets: a detailed table, cleaned data, and a summary. Tasks in Zoho CRM will be automatically created with file attachments and an accuracy assessment in the description. After implementation, I will provide the source code and documentation and conduct training.
Budget: 300 USD Deadline: 5 days
Hello!
I have successful experience in building analytics systems in various directions in Make/n8n.
Regarding the stack:
For double-checking, I would use Claude 3.5 Sonnet as the main one, and GPT-4.1 mini for verification.
For PDF - if the text is regular, a standard parser with configuration is sufficient; if it's a scan, then OCR support for Hebrew.
I would also add error handling and logging to Google Sheets in the script.
Budget: 250 USD Deadline: 7 days
Hello, Adi!
I am interested in your project. I fully understand the task: the approach of Cross-Validation, where one model analyzes and the other acts as an auditor, is the best solution for minimizing AI hallucinations in fintech.
I specialize in creating complex scenarios in Make.com and integrating LLM. I have experience in setting up document parsing with complex layouts (including working with RTL/Hebrew, which is often a "bottleneck" for standard parsers).
My proposal for implementation (Technical Workflow):
RTL OCR & Pre-processing: I will set up correct text extraction from PDF to maintain the right-to-left reading logic before sending it to AI.
Structured Analysis:
Claude 3.5 Sonnet: Receives instructions and returns data strictly in JSON format.
GPT-4o (Auditor): Compares JSON with the original text. The output provides a Confidence Score and a comment on accuracy.
Advanced Excel: Generation of a file with three sheets through Make.com iterators to maintain data integrity and formatting.
Zoho CRM: Creation of a task with conditional logic (Routers): if the accuracy score is below the threshold (for example, 90%) — we tag it as "Attention/Manual Review."
Why me:
Although I am actively developing my profile on this platform, I have a solid background. My profile includes a personal website where you can see examples of my completed integrations. I work for my reputation, so I guarantee a deep dive into your instructions.
