Automate lead workflowsFeed your AI with Google Maps data and let Claude Code do the busywork
Scrape businesses with the free Vonsel extension, export a CSV or JSON, then hand it to Claude Code and ChatGPT to clean, score, and prep every lead automatically.
AI Workflows··7 min read
Key takeaways
The free Vonsel Chrome extension is the means: capture Google Maps businesses and export them as CSV, Excel, or JSON in seconds
Claude Code turns that file into a repeatable pipeline: it reads local data, writes scripts, and reruns cleaning and scoring on demand
ChatGPT handles quick one-off analysis in a chat, while the Vonsel dashboard is the end: mapped CRM, review intelligence, and AI email per business
The idea
Feed your AI with real business data
Large language models are only as useful as the data you give them. Ask ChatGPT to "find me dentists in Chicago" and you get a plausible but hollow list: made-up names, guessed phone numbers, and addresses that do not exist. The fix is simple. Give the AI real data first, then let it work.
That is the whole premise behind this workflow. You use a scraper to pull genuine, current listings from Google Maps, then you point an AI at that file to do the tedious parts: cleaning, deduping, scoring, segmenting, and drafting outreach. The scraper is the means. The automated pipeline is the end.
The Vonsel Chrome extension handles the first half. It sits on top of Google Maps, captures the businesses you search for, and exports them in the format your AI likes best. It is free, with no trial and no card, so you can build the whole pipeline before you spend a cent. If you want the broader picture, our guide on how to feed your AI with business data covers the concept end to end.
Start with real data, not guesses
Add the Vonsel extension, scrape any niche and city from Google Maps, and export the file your AI will actually work with. Free download. No trial, no credit card.
Open Google Maps, search for something like "roofers in Austin", and let the Vonsel pop-up capture the results. When the capture finishes, pick your export format. Each one suits a different AI habit, and the extension writes all three from the same scrape. See the full breakdown in our export options guide.
Format
Best for
Why
CSV
ChatGPT, spreadsheets
Universal, lightweight, opens anywhere and uploads cleanly into a chat
Excel (XLSX)
Manual review, handoff
Keeps formatting and multiple sheets for teammates who live in spreadsheets
JSON
Claude Code, scripts, local LLMs
Structured and nested, ideal when code parses fields like reviews and categories
For most AI pipelines, JSON is the sweet spot. It preserves the structure of every field, including nested data like review counts, ratings, and categories, so a script never has to guess where one column ends and the next begins. If you plan to eyeball the list first, export CSV too. It takes one extra click.
Step two, the quick path
Work a single list with ChatGPT
When you have one list and a specific question, ChatGPT is the fastest tool. Drag your CSV into the chat, and its data analysis tool reads the file, runs Python behind the scenes, and answers in plain English. No setup, no terminal, no scripts to save.
Prompts that work well on a fresh Google Maps export:
Goal
Prompt to paste after the upload
Clean
"Remove duplicate rows, fix phone number formatting, and flag any listing missing a website."
Score
"Rank these businesses by opportunity: low rating plus high review count means an unhappy but busy business."
Segment
"Split this list into three tiers by review volume and give me each as a separate CSV."
Draft
"Write a two-line cold email opener for each business that references its category and city."
This is perfect for a one-time push. The catch is repetition. Next week, with a new city, you upload again and retype every prompt. That is where a second tool earns its keep. Our walkthrough on how to upload a lead CSV to ChatGPT has more prompt patterns to steal.
Step two, the automated path
Build a repeatable pipeline with Claude Code
Claude Code is an AI agent that lives in your terminal and works directly with the files on your machine. Instead of pasting data into a chat, you drop every Vonsel export into a folder and describe the pipeline once in plain English. Claude Code writes the scripts, runs them, and shows you the result. Then it reruns the same steps on any new file with a single command.
A realistic first session looks like this. You save three JSON exports (dentists, gyms, salons) into a leads folder, open Claude Code there, and say:
"Read every JSON file in this folder. Merge them, drop duplicates by phone number, add a column that scores each business as high, medium, or low priority based on rating and review count, then write the result to scored_leads.csv."
Claude Code does the work and saves the script. The magic is next week: you drop in a new export, type "run the scoring pipeline on the new file", and it repeats every step in seconds. You have turned a manual chore into a reusable, one-command workflow. For deeper prompt recipes, see our collection of Claude Code prompts for lead databases and the hands-on guide to analyzing leads with Claude Code.
0
cost to export from the extension, no trial and no credit card
3
formats out of every scrape: CSV, Excel, and JSON
1
command to rerun the whole pipeline on next week's list
Choosing your tool
ChatGPT vs Claude Code for lead work
Reach for ChatGPT when
You have one list and a one-off question. You want an answer in the browser, no terminal, no saved scripts. Fast exploration and quick email drafts.
Reach for Claude Code when
You run the same steps every week across many files. You want a saved, repeatable pipeline that reads local exports and reruns with one command.
Reach for Codex when
You already work inside an OpenAI coding setup and want your lead scripts to sit next to the rest of your project. Same file-based idea as Claude Code.
Reach for a local LLM when
Your lead lists include personal data you want to keep off the cloud. A model running on your own machine processes the JSON without it ever leaving your laptop.
There is no wrong choice. Many people start in ChatGPT to figure out the prompts, then move the winning ones into Claude Code once the routine is clear. Whatever you pick, the input is the same clean export from the extension, so you are never locked in.
Get the data your AI needs
Install the extension, capture a live Google Maps list, and export it straight into your Claude Code folder or ChatGPT window. Free download. No trial, no credit card.
Building your own AI pipeline is powerful, and it is also work. If you would rather skip the scripting, the Vonsel dashboard is where the extension leads. It applies the same kind of intelligence to your captured leads automatically, no prompts required.
Instead of a raw CSV, you get a mapped CRM where every business sits on a map inside a sales area. Review Intelligence reads each business's reviews and surfaces the pain points, so you know who is unhappy and why. Email Intelligence drafts a tailored message per business from those signals. This is the difference between a plain list and a value-added, contextualized database.
You can also reopen an earlier scrape and add more businesses to it, attach fresh reviews to leads you captured last month, and keep everything in a searchable history. The extension pulls the raw material. The dashboard turns it into leads you can actually sell to.
The scraper gets you the data. Your AI, or the Vonsel dashboard, turns it into deals
Feed your AI, then close the loop
Scrape, export, and automate with Claude Code or ChatGPT, then bring the winners into a mapped CRM. Free download. No trial, no credit card. Explore features or read the AI workflows blog.
How do I get Google Maps leads into Claude Code or ChatGPT?
Capture businesses with the free Vonsel Chrome extension, export them as CSV, Excel, or JSON, and hand that file to your AI. Drag it into ChatGPT to analyze it in one chat, or drop it in a folder and point Claude Code at it to run repeatable scripts across every file.
Is Claude Code or ChatGPT better for lead automation?
ChatGPT is best for quick, one-off analysis of a single list inside a chat window. Claude Code is best when you want to automate the same steps every week: it lives in your terminal, reads local files, writes reusable scripts, and runs the whole pipeline again with one command.
Do I need to know how to code to automate lead workflows with AI?
No. Both ChatGPT and Claude Code accept plain English prompts. You describe what you want done to your exported file and the AI writes and runs the code. The Vonsel dashboard also handles scoring, review analysis, and AI email drafting without any code at all.
Add Chrome Extension