I would like to pay someone to complete the set-up of my document agent but there is apparently no one to help. I put out a request and recieved 0 responses. Maybe I need to watch a bunch of Youtube videos and do this all myself. I was using Google AI Pro to guide me through but I had to perform a separate Google search for every step of the process. It was taking a very long time. Perhaps there are some good resources out there for this.
Hey @Bill_Foster , I can finish the document agent setup for you and remove the trial-and-error. Instead of piecing it together from videos and searches, I’ll configure the flow properly in Make, connect the document source, parsing, storage, and AI steps, and make sure the agent returns reliable outputs. If anything is unclear in the current setup, I’ll audit what’s already been built and complete the missing pieces quickly. I can also document the final scenario so you’re not stuck maintaining a black box later. If you want, send me a screenshot of your current scenario or share where you’re blocked, and I’ll tell you exactly how I’d finish it. Feel free to book a meeting with me here you can also checkout my profile here for some testimonials.
Hello @Bill_Foster , welcome to make.com community, I have worked and have experience with Make.com and l will love to collaborate with you on this you can schedule a call Here and you can checkout my upwork profile Here, for my pastworks and certifications
Hi Bill, welcome to the Make community 
Document agent setups can honestly become frustrating fast once AI parsing, embeddings, routing logic, and automation orchestration all start connecting together.
I work with Make-based AI workflow systems and document processing automations, so I can probably help you get this fully structured instead of chasing separate tutorials for every step.
Quick questions:
• what stack/tools are you currently using for the document agent?
• is the main goal document retrieval, summarization, extraction, or internal workflow automation?
A couple related projects:
• AI Voice Agent + API Automation
https://www.upwork.com/freelancers/~0122761e4734295f4b?p=2038586338272239616
• Multi-channel CRM + Automation System
https://www.upwork.com/freelancers/~0122761e4734295f4b?p=2039118619839795200
Happy to connect further.
@Bill_Foster received your request. I’m an AI and Automation expert, let me help set this up for you to your requirements. You can share them with me via dm or schedule a quick 15 min call. https://cal.com/tori-tam/intro-call
Hi Bill, I would separate this into two steps so you don’t keep chasing tutorials.
First, I would do a short audit of what already exists:
- where the documents live
- whether the goal is extraction, summarization, retrieval/RAG, or workflow automation
- what Make scenario/modules are already configured
- where the setup is failing
- what output you expect from the agent
Then I’d finish only the smallest working version first, document it, and leave you with a clear handoff so it is not a black box.
I work with Make/webhooks/API workflows and custom backend systems. My bias would be to keep the first version simple and reliable rather than overbuild a vague “document agent”.
A few questions:
- What tool are you using for the document agent?
- Are the documents in Google Drive, Notion, Airtable, or somewhere else?
- Do you need search/retrieval, summarization, extraction, or all three?
- Is there already a Make scenario started?
Bill,
You are at the exact point where a document agent either becomes useful or turns into a half-working science project. I would treat it as a setup, configuration, and reliability problem: figure out what has already been built, confirm what the agent is supposed to do with your documents, then finish the missing pieces in the right order so you are not stuck searching Google for every small step.
I would approach it this way:
-
I would first confirm which document agent platform you are using, because “document agent” can mean a Gemini/Google AI workflow, a Drive-connected assistant, a no-code agent builder, or a custom RAG setup.
-
I would review the setup you already started before changing anything. If pieces are already configured correctly, I would preserve them instead of restarting from scratch.
-
I would check the document source next, especially whether the files are in Google Drive, local folders, PDFs, Google Docs, Word files, Sheets, or a mix. The ingestion setup depends heavily on that.
-
I would confirm the actual use case before configuring more tools. An agent that summarizes documents is different from one that answers questions, extracts fields, searches across files, generates drafts, or triggers workflow actions.
-
I would test whether the documents are actually being indexed or connected correctly. A lot of document agents fail because the model looks connected, but retrieval is weak or incomplete.
-
I would look at chunking, document structure, permissions, and retrieval behavior before judging the AI output. Bad answers often come from bad document preparation, not from the model itself.
-
I would run real test questions against your actual files, not generic examples from a tutorial. That is the only honest way to know whether the agent is working.
-
I would check whether the agent needs citations or source references. If it cannot show where an answer came from, it may be hard to trust for anything important.
-
I would look at access control carefully. If private or business documents are involved, the agent should not expose files or answers to the wrong users.
-
I would simplify the setup wherever possible. If Google AI Pro is enough, I would not overbuild it. If it is not enough, I would tell you plainly and explain what is missing.
-
I would leave you with clear setup notes so you are not dependent on scattered searches or memory the next time you need to add files, change settings, or troubleshoot the agent.
The closest match from my past work is DocuMind.ai. I built a document intelligence platform that allowed users to ingest PDFs and Word files, convert them into searchable content, and query them using OpenAI embeddings, vector search, Supabase, Redis, Docker, and RAG architecture. The relevant part for your project is not just “AI experience.” It is the full document flow: getting files into the system, structuring them correctly, retrieving the right content, and making the answers usable instead of vague.
Another relevant project is efraudservices.com, an AI document analysis platform for financial PDFs. I worked on document upload, OCR processing, structured extraction, human validation, and dashboard-based review. That ties directly to your situation because document agents often need more than a chatbot layer. They need reliable document reading, clean extraction, source handling, and practical testing against real files.
I also built PrimeCareathome.com, a compliance-focused document platform for a healthcare operation. That system handled document storage, OCR, role-based access, status tracking, and structured records. That experience matters here if your document agent needs to work with sensitive documents, multiple file types, permissions, or a repeatable workflow instead of a loose experiment.
I can step in, review the current setup, finish the configuration, test it with your actual documents, and give you a clean explanation of how to maintain it afterward.
A few questions I would want to clarify:
-
What platform or tool are you using to build the document agent?
-
Where are the documents stored right now?
-
What do you want the agent to do with the documents: answer questions, summarize, extract fields, draft content, or perform actions?
-
How many documents are involved, and what file types are they?
-
Is this for your personal use only, or will other users need access?
- Brandon
brandon@bluegrass-media.com
501-733-1465
Hey @Bill_Foster,
Have sent you a DM as well. I can totally understand that looking at youtube videos and using AI pro to set up this agent can be difficult, so you need to bring an expert to set it up. I’d love to help as I run an automation studio called Automation Jinn where we help companies automate their processes and increase efficiency. I am Make advanced certified with a background in computer science so quite comfortable in integrating API and custom code. I have experience working with both SMB’s and large enterprises.
Some of my relevant work-
• For a Finance client, I automated the documented generation process which integrated with Jotform to automate the process from form submission to PDF delivery. It improved efficiency, reduced errors, ensured accurate PDF delivery, and eliminated operational bottlenecks.
• For an company with 500+ employees, I built end to end AI agents using Make for newsletter curation and Pinterest pin creation. It reduced the time for developing marketing assets from hours to minutes. The scenarios were made with robust error handling.
• For a bakery brand with a large social following, I implemented chatbots integrated with a custom CRM in Airtable and automation workflows to handle orders and FAQs saving hours of manual coordination every day.
I work more as an AI transformation partner for my clients rather than just a builder. Would love to learn about your day to day operations 
Happy to also jump on a call if it’s easier- https://cal.com/sparsh-gupta-pnrp6p/30min
Cheers,
Sparsh
Founder
Automation Jinn
Bill, I would handle this as a setup audit first, because “document agent” can mean a few different architectures.
The fastest useful pass is to identify: where the documents live, what file types are involved, whether you need search, summarization, extraction, or drafting, and which parts are already connected. Then I would test with a small set of real documents and confirm whether the agent can cite or point back to the source material. If retrieval is weak, fixing ingestion/chunking/permissions usually matters more than adding more prompts.
TinyOps Studio can map the first working setup as a fixed $149 audit, then quote the completion work once the current tool and document source are clear. If written contact is easier, email support@tinyopsstudio.com with the tool you are using and one example of the answer or output you want.
Hi Bill. I can help complete the document agent setup in a practical way.
A good first paid slice would be: review your current setup, identify the missing steps, configure one working document flow with test files, and send a short handoff guide. I would avoid using sensitive documents until the setup is confirmed with sample data.
If that works, I can start with the smallest working path first.
Hi Bill, setting up a document agent in Make is very doable. The typical setup is: a trigger (new file in Google Drive, form submission, or email attachment) feeding into an AI module (OpenAI or Claude) that processes the document content, then routing the output wherever it needs to go like a database, CRM record, or Slack notification.
Two questions:
- What platform are you using as the AI backbone: OpenAI, Claude, or something else?
- What should happen once the document is processed: extract specific data, generate a summary, answer questions, or trigger a downstream action?
We can scope this as a small fixed build at West Work Studios. DM me if you want to connect.