Skip to main content
Fit Score evaluates how well accounts align with your Ideal Customer Profile (ICP). Unlike traditional scoring that relies on static firmographics, Fit Score uses AI agents to research accounts and answer custom questions that aren’t stored in your CRM.

​
How it works

Seam uses a two-agent system to score accounts:
  1. Research agent - Investigates the account across multiple data sources (CRM, LinkedIn, company websites, public data)
  2. Evaluation agent - Reviews findings against your criteria and assigns a grade with explanation
This process takes 30-60 seconds per account and produces a letter grade (A, B, C, or D) with a detailed explanation.

​
Grading system

GradeMeaningAction
APerfect fit - strongly matches all ICP criteriaTop priority for sales and marketing
BGood fit - matches most criteriaStrong candidate for engagement
CPartial fit - matches some criteriaLower priority or nurture track
DPoor fit - does not match ICPExclude from active campaigns
Why letters instead of numbers? Sales teams don’t need to debate whether a 92 is better than an 88. Simple A/B/C/D grades make prioritization clear and actionable.

​
Defining your fit criteria

Configure custom criteria using natural language prompts. Fit Score can answer questions that traditional data providers can’t:
  • Employee count ranges
  • Revenue or funding stage
  • Geographic location
  • Company age
  • Growth rate
  • “Does this company use Snowflake or Databricks?”
  • “Do they have internal and external APIs?”
  • “What is their IT environment?”
  • “Do they use a specific tech stack?”
  • “How big is their Total Addressable Market?”
  • “Do they serve B2B or B2C customers?”
  • “Are they a marketplace or direct seller?”
  • “Do they underwrite merchants?” (fintech)
  • “Do they have SOC 2 compliance?”
  • “Are they in a highly regulated industry?”
  • “Do they handle sensitive data?”
  • “How many languages do they need for localization?”
  • “Do they have distributed teams?”
  • “What’s their content production volume?”
  • “Do they process high-volume transactions?”
  • “Is this company a direct competitor?”
  • “Are they too small for our product?”
  • “Do they serve industries we don’t support?”
Start with 5-10 clear criteria that reflect your closed-won customer patterns. You can refine and expand after back-testing.

​
Multiple fit models

Create different fit models for different segments or use cases:
  • SaaS fit score - Criteria for software companies
  • Manufacturing fit score - Criteria for industrial businesses
  • Enterprise fit score - Criteria for large accounts
  • SMB fit score - Criteria for small/mid-market
  • Product line A - Criteria for specific product vertical
Each model can have completely different criteria and grading logic.

​
Back-testing your fit model

Before deploying, Seam validates your criteria with a 45-account test:
1

Select test accounts

  • 15 known good accounts (closed-won customers)
  • 15 known bad accounts (lost deals, poor fits)
  • 15 random accounts from your TAM
2

Run fit score

Seam scores all 45 accounts using your criteria
3

Validate results

Verify that:
  • Good accounts score A or B
  • Bad accounts score C or D
  • Random accounts distribute appropriately
4

Refine criteria

Adjust criteria if results don’t align with expectations, then retest
Don’t skip back-testing! This ensures your criteria accurately identify the right accounts before you spend credits scoring your entire TAM.

​
How scores are displayed

​
Detailed explanations

Every Fit Score includes a detailed explanation:
“This account scored a B because it is headquartered in the US, operates in a target industry, has 500-2,000 employees, and serves a TAM larger than 75,000 accounts. However, it lacks SOC 2 compliance, which is preferred for enterprise deals.”

​
CRM integration

Fit Scores sync directly to your CRM:
  • Displayed as fields on Account records
  • Explanation text stored in custom fields
  • Available for reporting, filtering, and workflows
  • Visible to sales reps without logging into Seam
Fit Score in Salesforce

​
Using fit score in plays

Fit Score acts as a filter in automated plays to ensure you only engage good-fit accounts. Example play: New Funding Outreach 
Trigger: Account raises Series B funding
  ↓
Filter: Fit Score = A or B?
  ↓ YES
Actions:
  - Research and find 3 buying committee members
  - Enrich contact data
  - Add to congratulatory email sequence
  - Notify assigned SDR

Filter: Fit Score = C or D?
  ↓ NO
Skip - don't waste resources on poor-fit accounts
This prevents wasting:
  • ✅ Marketing budget on wrong targets
  • ✅ SDR time on low-potential accounts
  • ✅ Email sends on non-ICP companies

Build Plays with Fit Filters

Learn how to configure plays using Fit Score

​
Multi-dimensional scoring

Fit Score is one dimension in Seam’s comprehensive scoring model:
Score typeWhat it measuresTimeframe
FitICP alignment (foundational)Static/Stable
IntentBuying signals, topic researchDynamic/Momentary
SignalsTrigger events (hiring, funding)Dynamic/Momentary
ReachContact availability and qualityStatic/Stable
Key distinction: Fit Score is foundational and stable—it measures whether a company is fundamentally a good target. Intent and Signals scores tell you when a good-fit account is heating up.

Intent Scoring

Track buying signals and research activity

Signal Monitoring

Monitor trigger events across accounts

​
Best practices

Analyze your best customers and translate their characteristics into criteria. What do your A+ customers have in common?
“Does this company have a dedicated data engineering team?” is better than “Does this company care about data?”
Define what makes an account NOT a fit. Competitors, wrong industries, or accounts too small/large.
For each criterion, specify what constitutes a strong yes, a partial match, and a no. This helps the evaluation agent grade accurately.
After scoring your first 100-200 accounts, review the grades. Are the right accounts getting A/B scores? Refine criteria if needed.
Combine Fit Score with Intent and Signals. An A-fit account with high intent is your hottest opportunity.

​
Limitations & considerations

Scoring takes 30-60 seconds per account. Not suitable for real-time scoring during inbound form submissions.
Accounts must exist in your CRM (Salesforce/HubSpot) to be scored. Seam doesn’t discover net-new accounts via Fit Score.
Research quality depends on publicly available data. Highly private companies or niche industries may have limited information.
Most valuable for companies with 1,000+ potential accounts. Less critical if you have 50 well-known target accounts.
Fit Score produces a grade + explanation. If you need specific data points (e.g., “number of retail locations”) stored as fields, use enrichment plays instead.

​
Common questions

Most customers re-score quarterly or when significant company changes occur (funding, acquisition, leadership changes). Fit Score is relatively stable over time.
Not directly. Upload accounts to your CRM first, then run Fit Score. Alternatively, use Seam’s TAM building features to discover and score accounts simultaneously.
Scores can change if the company evolves (e.g., grows from 200 to 600 employees, launches APIs). Re-score periodically to keep data fresh.
Yes. Seam shows which data sources informed each criterion’s answer, providing transparency into how grades are determined.
The evaluation agent will note that information is unavailable. You can choose whether unavailable data should impact the grade or be ignored.