OpenAI + Gmail, product research summaries in your inbox
You start with a simple product idea. Then the tab explosion begins: Google results, “best of” lists, Reddit threads, half-baked notes, and a dozen “maybe” links you’ll never find again.
This OpenAI Gmail automation hits marketing leads who need quick vendor options, but founders and ops managers feel it too when purchasing decisions stall. You want something you can forward in one click, not a messy browser history.
This workflow turns your description into search queries, pulls results, summarizes what matters, and emails you a clean report with links. You’ll see how it works, what you need, and where teams usually trip up.
How This Automation Works
The full n8n workflow, from trigger to final output:
n8n Workflow Template: OpenAI + Gmail, product research summaries in your inbox
flowchart LR
subgraph sg0["Manual Flow"]
direction LR
n0@{ icon: "mdi:play-circle", form: "rounded", label: "Manual Trigger", pos: "b", h: 48 }
n1@{ icon: "mdi:swap-vertical", form: "rounded", label: "Set Product Description", pos: "b", h: 48 }
n2["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/openAi.dark.svg' width='40' height='40' /></div><br/>AI: Generate Search Queries"]
n3@{ icon: "mdi:code-braces", form: "rounded", label: "Split Queries", pos: "b", h: 48 }
n4@{ icon: "mdi:cog", form: "rounded", label: "Google Custom Search (CSE)", pos: "b", h: 48 }
n5@{ icon: "mdi:swap-vertical", form: "rounded", label: "Combine Search Results", pos: "b", h: 48 }
n6["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/openAi.dark.svg' width='40' height='40' /></div><br/>AI: Summarize Products & Pri.."]
n7@{ icon: "mdi:message-outline", form: "rounded", label: "Send Report Email", pos: "b", h: 48 }
n3 --> n4
n0 --> n1
n5 --> n6
n1 --> n2
n4 --> n5
n2 --> n3
n6 --> n7
end
%% Styling
classDef trigger fill:#e8f5e9,stroke:#388e3c,stroke-width:2px
classDef ai fill:#e3f2fd,stroke:#1976d2,stroke-width:2px
classDef aiModel fill:#e8eaf6,stroke:#3f51b5,stroke-width:2px
classDef decision fill:#fff8e1,stroke:#f9a825,stroke-width:2px
classDef database fill:#fce4ec,stroke:#c2185b,stroke-width:2px
classDef api fill:#fff3e0,stroke:#e65100,stroke-width:2px
classDef code fill:#f3e5f5,stroke:#7b1fa2,stroke-width:2px
classDef disabled stroke-dasharray: 5 5,opacity: 0.5
class n0 trigger
class n3 code
classDef customIcon fill:none,stroke:none
class n2,n6 customIcon
The Problem: Product Research Turns Into Tab Chaos
Manual product research feels easy for the first five minutes. Then it gets slippery. You search, open a few results, skim titles, jump between “reviews” and retailer pages, and try to remember which link mentioned the one feature you actually care about. By the time you’re ready to decide, you’ve got scattered notes, duplicate tabs, and a vague feeling you missed a better option. And if someone asks, “Can you share what you found?” you end up recreating the whole trail.
The friction compounds. Here’s where it breaks down in real life.
- Each new product idea means another research session that can easily burn an hour before you even have a shortlist.
- Search results are inconsistent, so you keep rewriting queries and chasing slightly different angles.
- You can’t compare options cleanly because the “insights” live across snippets, blogs, and retailer pages.
- Sharing findings is annoying, which means decisions get delayed or made on gut feel.
The Solution: OpenAI-Powered Research Summaries Sent to Gmail
This workflow acts like a lightweight research assistant you can actually use. You start by giving it a natural description of what you’re trying to buy (the more specific, the better). OpenAI turns that description into a set of focused search phrases, then n8n runs those queries through a Google Custom Search Engine so you’re pulling from Google’s index without manual clicking. The workflow collects the top titles and snippets, feeds that structured input back into OpenAI, and produces a decision-ready summary: common product types, key features that keep showing up, and places to buy. Finally, it sends the whole report to your Gmail inbox with direct links you can open or forward.
The workflow starts with your product description and a quick “generate queries” pass. Next, it runs multiple searches and aggregates the results into a clean bundle. Then the summary is written and emailed to you as a single report.
What You Get: Automation vs. Results
| What This Workflow Automates | Results You’ll Get |
|---|---|
|
|
Example: What This Looks Like
Say you’re sourcing one new tool each week and you usually run about 8 searches, opening maybe 6 results per search. If you spend only 2 minutes per page, that’s roughly 1.5 hours of skimming before you even summarize anything. With this workflow, you write one description (about 3 minutes), wait for searches and analysis (about 5 minutes), and you get a Gmail report you can forward immediately. That’s about an hour back, every week.
What You’ll Need
- n8n instance (try n8n Cloud free)
- Self-hosting option if you prefer (Hostinger works well)
- OpenAI to generate queries and summaries
- Gmail to deliver the report to your inbox
- Google Custom Search API key + CSE ID (get it from Google Cloud Console and your CSE settings)
Skill level: Intermediate. You’ll be connecting accounts and setting up Google CSE credentials, which is straightforward but a little fiddly.
Don’t want to set this up yourself? Talk to an automation expert (free 15-minute consultation).
How It Works
You provide the product description. The workflow starts from a manual run in n8n, where you paste a plain-English request like “lightweight hiking backpack for multi-day trips with good ventilation.”
OpenAI turns your request into search phrases. Instead of you guessing keywords, the AI generates multiple query lines designed to surface relevant product categories, features, and sellers.
Google CSE retrieves results at scale. n8n splits the query lines, runs each one in Google Custom Search, and collects the top titles, snippets, and URLs so you’re working from a broad scan, not one lucky search.
OpenAI writes the report and Gmail delivers it. The workflow aggregates the results, summarizes patterns and options, then emails you a clean report with links you can open, share, or drop into your purchasing process.
You can easily modify the search depth to change how many results get analyzed based on your needs. See the full implementation guide below for customization options.
Step-by-Step Implementation Guide
Step 1: Configure the Manual Trigger
This workflow starts when you manually execute it, then passes into the product description setup.
- Add the Manual Execution Start node as your trigger.
- Confirm the connection from Manual Execution Start to Assign Product Summary in the canvas.
- Keep Flowpast Branding as an optional sticky note for documentation only.
Step 2: Connect Google Custom Search
This step generates search queries and runs them through Google Custom Search to gather product data.
- In Google CSE Lookup, set Operation to
search. - Set Query to
={{ $json.query }}. - Set Search Engine ID to
[YOUR_ID]. - Credential Required: Connect your googleApi credentials in Google CSE Lookup.
⚠️ Common Pitfall: Leaving [YOUR_ID] empty will cause the Google CSE request to fail. Replace it with your Custom Search Engine ID.
Step 3: Set Up Search Phrase Generation and Query Splitting
These nodes convert a product description into multiple search queries and split them into individual items.
- In Assign Product Summary, set productDescription to
Lightweight, durable hiking backpack for multi-day trips (30-40L), good ventilation, comfortable for long hikes.. - In AI Create Search Phrases, set Model to
gpt-3.5-turbo. - Ensure the user message content in AI Create Search Phrases includes
{{ $json.productDescription }}. - In Divide Query Lines, keep the Function code that splits
items[0].json.choices[0].message.contentinto separate queries. - Credential Required: Connect your openAiApi credentials in AI Create Search Phrases.
⚠️ Common Pitfall: AI Create Search Phrases has no credentials configured in the workflow. Add your openAiApi credentials to avoid authentication errors.
Step 4: Aggregate Search Results and Generate the Report
This stage collects all search results and turns them into a summary report using the AI model.
- In Aggregate Search Results, set Mode to
jsonand Value to={{ $('Google CSE Lookup').all() }}. - In AI Summarize Findings, set Model to
gpt-3.5-turbo. - Ensure the user message includes
{{ $node["Assign Product Summary"].json.productDescription }}and the results mapping expression for titles, snippets, and links. - Credential Required: Connect your openAiApi credentials in AI Summarize Findings.
Step 5: Configure the Email Output
This step sends the summarized report to an email recipient using Gmail.
- In Dispatch Report Email, set To Email to
[YOUR_EMAIL]and From Email to[YOUR_EMAIL]. - Set Subject to
AI Shopping Assistant: Product Research Report for "{{ $node["Assign Product Summary"].json.productDescription }}". - Set Text to the full template, including
{{ $node["AI Summarize Findings"].json.choices[0].message.content }}. - Credential Required: Connect your gmailApi credentials in Dispatch Report Email.
⚠️ Common Pitfall: Using placeholder values like [YOUR_EMAIL] will prevent delivery. Replace both fields with valid email addresses.
Step 6: Test and Activate Your Workflow
Run a manual test to verify the full execution path from search creation to email delivery.
- Click Execute Workflow starting from Manual Execution Start.
- Confirm that Google CSE Lookup runs once per query produced by Divide Query Lines.
- Verify that AI Summarize Findings outputs a concise report and that Dispatch Report Email sends an email with the summary.
- Once verified, toggle the workflow to Active for production use.
Common Gotchas
- Google CSE credentials can expire or be missing permissions. If things break, check your Google Cloud Console API key restrictions and the Custom Search API enablement first.
- If you’re using Wait nodes or external rendering, processing times vary. Bump up the wait duration if downstream nodes fail on empty responses.
- Default prompts in AI nodes are generic. Add your brand voice early or you’ll be editing outputs forever.
Frequently Asked Questions
About 30-60 minutes, mainly because Google CSE setup takes a bit.
No. You’ll connect accounts, paste in API keys, and adjust a couple of prompts.
Yes. n8n has a free self-hosted option and a free trial on n8n Cloud. Cloud plans start at $20/month for higher volume. You’ll also need to factor in OpenAI API usage and Google Custom Search requests, which usually stays low for occasional research.
Two options: n8n Cloud (managed, easiest setup) or self-hosting on a VPS. For self-hosting, Hostinger VPS is affordable and handles n8n well. Self-hosting gives you unlimited executions but requires basic server management.
Yes, and it’s one of the best reasons to run this in n8n. You can adjust the “AI Summarize Findings” prompt to output a shortlist table, a pros/cons breakdown, or a simple “top 5 picks” section. If you want the email to include richer structure, edit the fields used by “Dispatch Report Email” so the message includes headings and cleaner link grouping. Common tweaks include limiting retailers, focusing on one feature (like “battery life”), or writing the summary in your brand voice.
Usually it’s an expired OAuth connection in n8n. Reconnect the Gmail credential, then confirm the Gmail account has permission to send mail (and that you didn’t change passwords or security settings). If the node runs but no email arrives, check spam and verify the “To” field is being set correctly in the email node.
Plenty for normal weekly research.
Often, yes, because this workflow benefits from branching, looping through multiple queries, and shaping data before summarizing. Zapier and Make can do it, but multi-search flows get awkward fast, and you may end up paying more as tasks grow. n8n also gives you the option to self-host, which is useful if you want to run lots of research without worrying about execution caps. If you’re keeping it extremely simple (one input, one summary, one email), Zapier might feel quicker. Talk to an automation expert if you want help picking the right setup.
Once this is set up, you stop doing research the hard way. The workflow does the scanning and summarizing, and you get a clean email you can act on.
Need Help Setting This Up?
Our automation experts can build and customize this workflow for your specific needs. Free 15-minute consultation—no commitment required.
