RB2B → TwentyCRM Import Pipeline

Download

Download RB2B → TwentyCRM Import Pipeline — zip containing the script and README.

Scripts for importing an RB2B visitor-identification CSV export into a TwentyCRM workspace. Creates companies, people, scored opportunities, and "Next Steps" checklist notes.

See tools/rb2b.md for full context on RB2B, the data model, and automation opportunities.

---

Prerequisites

Python 3.11+ — no third-party packages required (stdlib only).

Add to the root .env file:

RB2B_TWENTYCRM_API_KEY=your-client-workspace-token
RB2B_TWENTYCRM_API_BASE=https://your-twenty-instance.com/rest

Get the API key from Settings → Playground in the target TwentyCRM workspace.

Important: Use the client workspace token — not the Pacing agency workspace token. Each TwentyCRM workspace has its own isolated API keys.

---

Scripts

1. rb2b-import.py — Import companies and people

Reads the RB2B CSV and creates all companies and people in TwentyCRM. Custom fields are created automatically on the first run.

python3 scripts/resources/rb2b/rb2b-import.py \
  --csv ~/Downloads/filtered_custom_report_XXXXXXXXX.csv

What it does:

1. Reads ProfileType to split rows into Company and Person groups
2. Creates 9 custom fields on both company and person objects (idempotent — skips fields that already exist):
   • industry, estimatedRevenue (Text)
   • allTimeViews, recentPageViews (Number)
   • lastSeenAt, firstSeenAt (DateTime)
   • leadTags, leadReferrer, visitedPages (Text)
3. Creates all Company rows first — maps standard fields (name, domainName, linkedinLink, employees, address) plus custom fields
4. Creates stub companies for any company referenced by a Person row but not present as a Company row
5. Creates all Person rows — maps name, emails, jobTitle, linkedinLink, city, links to company

Field mapping:

CSV	TwentyCRM field	Notes
CompanyName	name	Company and person
Website	domainName.primaryLinkUrl	Company
LinkedInUrl	linkedinLink.primaryLinkUrl	Both
EstimatedEmployeeCount	employees	Converted from range to midpoint integer
City + State + Zipcode	address	Company
WorkEmail	emails.primaryEmail	Person
FirstName + LastName	name.firstName / name.lastName	Person
Title	jobTitle	Person
Industry	industry (custom)	Both
EstimateRevenue	estimatedRevenue (custom)	Both
AllTimePageViews	allTimeViews (custom)	Both
RecentPageCount	recentPageViews (custom)	Both
LastSeenAt	lastSeenAt (custom)	Both
FirstSeenAt	firstSeenAt (custom)	Both
Tags	leadTags (custom)	Both
MostRecentReferrer	leadReferrer (custom)	Both
RecentPageUrls	visitedPages (custom)	Deduplicated, newline-separated

Output: Writes rb2b-import-log.json alongside the script with every created record's CRM UUID.

---

2. rb2b-opportunities.py — Create scored opportunities

Reads the import log and CSV, scores every lead, then creates the top N opportunities with stage, score, and a "Next Steps" checklist note.

python3 scripts/resources/rb2b/rb2b-opportunities.py \
  --csv ~/Downloads/filtered_custom_report_XXXXXXXXX.csv \
  --top 10

Scoring (1-10):

Signal	Points
Base	2
All-time views ≥ 5 / ≥ 15 / ≥ 30 / ≥ 50	+1 / +2 / +3 / +4
Recent views ≥ 3 / ≥ 10	+1 / +2
"Hot page" tag	+2
Has work email	+1

Stages assigned:

Stage	Condition
MEETING	"Hot page" AND (has email OR views ≥ 20)
SCREENING	Has email OR views ≥ 10
NEW	Everything else

What it creates per opportunity:

• Opportunity record linked to company, with score and stage set
• pointOfContact set to linked person (if one exists)
• One "Next Steps" note with a BlockNote checklist attached via noteTargets

Checklist items generated per lead:

• Research [name]: industry, size, product & background
• Find key decision makers (CMO / VP Marketing / Head of Growth)
• Send personalised cold outreach email (if has email)
• Follow up after 3 business days (if has email)
• Review visited pages for intent signals (if has page data)
• HIGH PRIORITY fast-track note (if score ≥ 7)

Output: Writes rb2b-opportunities-log.json with each created opportunity's ID, stage, score, and tasks.

---

3. rb2b-notes.py — Add notes to existing opportunities

Idempotent script that adds a "Next Steps" checklist note to any opportunity that doesn't already have one. Useful after manually creating opportunities in the UI.

python3 scripts/resources/rb2b/rb2b-notes.py \
  --csv ~/Downloads/filtered_custom_report_XXXXXXXXX.csv

Skips opportunities that already have a note linked via noteTargets.

---

Typical Run Order

# 1. Import all companies and people
python3 scripts/resources/rb2b/rb2b-import.py \
  --csv ~/Downloads/your-export.csv

# 2. Review the score ranking (printed to console) and pick top N
# 3. Create top 10 opportunities with notes
python3 scripts/resources/rb2b/rb2b-opportunities.py \
  --csv ~/Downloads/your-export.csv \
  --top 10

# 4. If you created additional opportunities manually in the UI, add notes to them
python3 scripts/resources/rb2b/rb2b-notes.py \
  --csv ~/Downloads/your-export.csv

---

Re-running / Idempotency

Script	Safe to re-run?	Notes
rb2b-import.py	⚠ Partial	Custom field creation is idempotent; company/person creation will create duplicates if run twice. Check CRM before re-running.
rb2b-opportunities.py	⚠ Partial	Will create duplicate opportunities if run twice on the same leads. Delete existing opps first if re-running.
rb2b-notes.py	✅ Yes	Checks for existing noteTargets before creating.

---

Logs

All scripts write JSON logs to their working directory:

• rb2b-import-log.json — company and person CRM IDs
• rb2b-opportunities-log.json — opportunity CRM IDs, stages, scores

Keep these logs alongside the source CSV — you'll need the IDs if you want to update or delete records later.

---

Automation (Future)

The current pipeline is manual. See tools/rb2b.md for a full breakdown of automation options, including:

• RB2B webhook → n8n → TwentyCRM (no CSV step)
• Automated email sequences via Resend for high-scoring leads
• LinkedIn Matched Audiences export for retargeting
• Slack alerts for score-10 leads