Step 2: Connect HubSpot

Fetch the target contacts from HubSpot using a search filter that focuses on decision makers.

1. Select the Retrieve HubSpot Contacts node.
2. Set Operation to search.
3. Set Authentication to oAuth2.
4. In Filter Groups, add a filter where Property Name is hs_buying_role|enumeration and Value is DECISION_MAKER.
5. Credential Required: Connect your hubspotOAuth2Api credentials.

Flow note: Retrieve HubSpot Contacts outputs to Iterate Contacts Batch to process contacts one at a time.

Step 3: Set Up Contact Batching and Field Mapping

Batch processing ensures one contact is handled at a time, while the field mapping standardizes data for later steps.

1. Open Iterate Contacts Batch and keep default batch options (no changes required).
2. Ensure Iterate Contacts Batch outputs to Contact Anchor, then to Map Contact Fields.
3. In Map Contact Fields, add assignments:
4. Set firstname to {{ $json.properties.firstname }}.
5. Set lastname to {{ $json.properties.lastname }}.
6. Set email to {{ $json.properties.email }}.
7. Set product_to_sell to AI partnerships: a consulting package of AI development and services. We help customers find a strong foothold on AI initiatives bringing them to life cost effectively and always with results..

Contact Anchor is a placeholder node used to anchor the batch flow; it requires no configuration.

Step 4: Set Up Email Retrieval and Persona Extraction

Pull recent email history for each contact and use it to build a persona profile with AI.

1. Configure Fetch Customer Emails with Operation set to getAll.
2. Set Limit to 20 and Simple to false.
3. Set the search query in Filters → q to from:{{ $json.email }}.
4. Credential Required: Connect your Gmail credentials to Fetch Customer Emails (none are configured yet).
5. In Build Persona Profile, set Text to the full expression:
   {{ $input.all() .map(item => `subject: ${item.json.subject} date: ${$json.headers.date} message: ${item.json.text.substr(0, item.json.text.indexOf('> wrote:') ?? item.json.text.length).replace(/^On[\w\W]+$/im, '')}`) .join('\n---\n') }}
6. In Build Persona Profile, set the System Prompt Template to Your task is to build a persona of a customer or potential customer so that we may better serve them for our business. Analyse the recent correspondence of the user, {{ $('Map Contact Fields').item.json.email }}, and extract the required attributes..
7. Ensure Gemini Chat Model A is connected as the language model for Build Persona Profile and set Model Name to models/gemini-2.0-flash.
8. Credential Required: Connect your googlePalmApi credentials to Gemini Chat Model A (AI credentials are set on the model node, not on Build Persona Profile).

⚠️ Common Pitfall: The Gmail nodes require credentials even if the workflow imported correctly. Make sure both Gmail nodes are authenticated before testing.

Step 5: Configure Sales Drafting and Gmail Draft Output

Use the persona profile to draft a tailored sales message, then save it as a Gmail draft addressed to the contact.

1. In Draft Sales Message, set Text to # Profile of {{ $('Map Contact Fields').first().json.firstname }} {{ $('Map Contact Fields').first().json.lastname }} {{ Object.keys($json.output).map(key => `## ${key}\n${$json.output[key]}`).join('\n') }}.
2. Set the System Prompt Template to You are a sales representative drafting an email to close a potential customer on the following product: <product>{{ $('Map Contact Fields').first().json.product_to_sell }}</product> Use the provided profile to draft the a suitable email which reflects similar communication style and addresses their values, ultimately convinces the customer to inquire about and/or buy this product. Provide only the subject and body of the message as this text will go into a template. Omit the subject and signature..
3. Ensure Gemini Chat Model B is connected as the language model for Draft Sales Message and set Model Name to models/gemini-2.0-flash.
4. Credential Required: Connect your googlePalmApi credentials to Gemini Chat Model B (AI credentials are set on the model node, not on Draft Sales Message).
5. Configure Create Gmail Draft with Resource set to draft and Email Type set to html.
6. Set Message to {{ $json.output.body }} and Subject to {{ $json.output.subject }}.
7. Set Options → Send To to {{ $('Map Contact Fields').first().json.email }}.
8. Credential Required: Connect your Gmail credentials to Create Gmail Draft (none are configured yet).

Flow note: Create Gmail Draft outputs back to Iterate Contacts Batch to continue processing the next contact.

Step 6: Test and Activate Your Workflow

Run a manual test to confirm contact retrieval, persona building, and draft creation are working end-to-end.

1. Click Execute Workflow on Manual Start Trigger to run a test.
2. Verify Retrieve HubSpot Contacts returns contacts tagged with DECISION_MAKER.
3. Check that Fetch Customer Emails returns email history for each contact.
4. Confirm Build Persona Profile outputs a structured persona object and Draft Sales Message outputs subject and body.
5. Open Gmail and confirm a draft exists for each processed contact.
6. When satisfied, toggle the workflow to Active for production use.