How to Import Your Customer List and Find Lookalike Prospects with AI
Step-by-step guide to importing a CSV of your current customers into Origami and using AI to find lookalike companies that match your best-fit profile. Includes scoring, enrichment, and scaling.
Founding AI Engineer @ Origami
You already know who your best customers are. The problem is finding more of them.
Most teams try to reverse-engineer their ICP manually — scanning CRM data, building filters in LinkedIn Sales Nav, guessing at what "similar" means. It takes hours and the results are inconsistent.
Quick Answer: Import a CSV of your current customers into Origami, tell the AI agent to analyze them, and it will build an ideal customer profile automatically — then search for thousands of lookalike companies that match. No manual ICP definition, no filter-building. Upload, describe what you want, and the agent does the research, scoring, and list-building for you.
Why Lookalike Prospecting Works
Your current customers are signal-rich data. They tell you:
- What industries buy from you (SaaS, fintech, healthcare)
- What company size converts (50-person startups vs. 500-person mid-market)
- What business models fit (B2B, marketplace, subscription)
- What locations cluster (US-heavy, EU-focused, remote-first)
- What contract values are typical (SMB deals vs. enterprise)
Traditional prospecting ignores this. You pick filters based on gut feel. Lookalike prospecting uses your actual customer data as the input, and AI does the pattern-matching.
How to Find Lookalike Prospects in Origami (Step by Step)
Step 1: Prepare Your Customer CSV
Export a list of your current customers. At minimum, include:
- Company name (or domain)
- Industry or vertical
- Employee count or company size
- Location
Optional but useful: business model, contract value, customer-since date, product tier. The more context you give the AI, the better the lookalike profile it builds.
Step 2: Import the CSV into Origami
Open your Origami workspace and click the Import CSV button in the prompt area. Select your file. Origami will parse the columns and display the data in a table — company names, business models, demographics, contract details, all of it.
Step 3: Tell the Agent What to Do
The AI agent will recognize you've imported a customer list. It might ask what you'd like to do with it. Be direct:
"These are my current customers. Do an analysis on them and find me new lookalike customers with similar traits."
That's it. Conversational. The agent takes it from there.
Step 4: AI Analyzes Your Customer Base
The agent runs an automated analysis across your entire customer list. It looks at:
- Business model distribution — e.g., "80% of your customers are B2B SaaS"
- Company size patterns — employee count ranges, revenue tiers
- Location clusters — geographic concentration
- Industry breakdown — verticals and sub-verticals
- Contract value ranges — deal size patterns
From this, it builds an ideal lookalike profile — a composite of what your best customers look like. No manual ICP workshops. No guessing.
Step 5: The Agent Searches for Matches
Using the lookalike profile, the agent searches across its data sources. In the demo, it found nearly 5,000 matching companies in total. It then builds a curated list — say 30 top prospects to start — in a new table called something like "Lookalike Prospects."
Each row includes:
- Company name
- Domain
- LinkedIn URL
- Company description
- Employee count
- Location
- Industry
Step 6: AI Scores Each Prospect
Here's where it gets powerful. The agent doesn't just find companies — it scores each one against your customer profile. It adds columns for:
- Fit score (how closely the company matches your existing customers)
- Business model match (B2B SaaS, marketplace, etc.)
- Tech vertical alignment
- Detailed analysis (a written explanation of why this company is a strong or weak match)
You can sort by fit score and immediately see which prospects are worth pursuing first.
Step 7: Refine, Enrich, or Scale
From here, you have options:
- Find decision makers: Ask the agent to "find decision makers at these top-scoring companies." It creates a new table with names, titles, verified emails, and phone numbers.
- Scale the list: If you like the initial 30, tell it to give you 4,800 more companies that match the same profile.
- Narrow further: Add constraints — "Only companies with 100+ employees" or "Only US-based" — and the agent refines.
What Makes This Different from Manual ICP Work
| Approach | Time | Accuracy | Scale |
|---|---|---|---|
| Manual ICP + LinkedIn filters | Hours | Based on assumptions | Dozens of prospects |
| CRM reports + spreadsheet analysis | Hours | Better, but static | Limited by your analysis skills |
| Origami lookalike import | Minutes | Based on actual customer data | Thousands of scored prospects |
The difference is that the AI reads your entire customer list, finds the patterns you'd miss, and then searches at a scale you can't match manually.
Tips for Better Lookalike Results
Include more customer data. The richer your CSV, the better the profile. Contract value, customer tenure, product tier, and business model all help the AI distinguish "good fit" from "sort of similar."
Start with your best customers. If you have 500 customers but 50 of them drive 80% of revenue, import those 50. The lookalike profile will reflect your highest-value segment.
Iterate. After the first batch of lookalikes, review the top-scored prospects. If they're off, tell the agent: "These are too small" or "I need more fintech companies." It adjusts.
Use scoring to prioritize. Don't treat the list as flat. Sort by fit score and work the top 20% first. The scoring is there for a reason.
FAQ
How do I find companies similar to my existing customers? Import a CSV of your current customers into Origami. The AI agent analyzes demographic and firmographic patterns, builds a lookalike profile, and searches for companies that match. It scores each prospect on fit and returns an enriched list.
What data should I include in my customer CSV? At minimum: company name or domain. For better results: industry, employee count, location, business model, contract value, and customer tenure. The more fields, the more precise the lookalike profile.
How many lookalike prospects can I find? Origami can surface thousands of matches. In the demo, the system found nearly 5,000 matching companies from a single customer list import. You can start with a curated batch (e.g., 30) and scale up.
Can I get contact information for lookalike prospects? Yes. After the lookalike list is built, ask the agent to find decision makers at the top-scoring companies. It returns names, titles, verified emails, phone numbers, and LinkedIn URLs in a new table.
How accurate is the fit scoring? The AI scores based on the actual attributes of your customer base — business model, size, industry, location. It's data-driven, not guess-driven. You can review the detailed analysis column for each prospect to see why it scored the way it did.
Is this the same as building an ICP? It's faster. Instead of manually defining your ICP in a doc and then searching for companies that match, you import real customer data and the AI derives the profile and searches simultaneously. The ICP is implicit in your customer list — the agent just extracts it.