v9ai/salescue-score-v1

SalesCue — score

LeadScorer module from the SalesCue sales intelligence library.

Status: untrained — architecture only, random initialization. Use as a starting point for fine-tuning.

Research Contribution

Causal Signal Attribution via Learned Interventions (v2)

Standard counterfactual masking replaces tokens with [MASK], which shifts the input distribution. LeadScorer learns a null intervention embedding per signal type that represents "this signal was never present" rather than "this signal was removed." This is closer to Pearl's do-calculus: do(signal=absent) not observe(signal=masked). The null embeddings are trained adversarially to produce encoder outputs indistinguishable from inputs where the signal genuinely doesn't exist.

v2 enhancements (grounded in competitive analysis of 60 sales AI platforms, 303 features):

• 32 signals across 6 industry categories (intent, engagement, enrichment, analytics, outreach, automation) — up from 15 generic signals
• Multi-scale signal detection: token-level cross-attention, document-level CLS projection, with learned scale fusion
• Signal interaction graph: learned adjacency matrix for cross-signal amplification/inhibition (e.g., urgency + budget = amplified; engagement_low + meeting_request = contradiction)
• Per-category sub-scores: interpretable 0-100 breakdowns for each of the 6 categories
• Deeper scoring network: 512→256 projection with layer norm and dropout

The model produces an uncertainty-weighted score combining classification and regression (Kendall et al., 2018), with causal attribution via learned counterfactuals.

Usage

from salescue import SalesCueModel

model = SalesCueModel.from_pretrained("v9ai/salescue-score-v1")
result = model.predict("your sales text here")
print(result)

Labels

• hot
• warm
• cold
• disqualified

Architecture

• Backbone: microsoft/deberta-v3-base (shared encoder, 768-dim)
• Head: LeadScorer
• Parameters: head only (backbone loaded separately)

Intended Use

• Primary: B2B sales intelligence — lead scoring, email analysis, conversation insights
• Users: Sales teams, RevOps, GTM engineers building sales automation
• Input: English sales text (emails, call transcripts, prospect communications)

Limitations

• Untrained weights: This release contains the architecture only. Weights are randomly initialized and must be fine-tuned on domain-specific data before production use.
• English only: Designed for English sales text. Performance on other languages is untested.
• Domain-specific: Optimized for B2B sales communications. May not generalize to other text domains.
• Shared backbone: Requires microsoft/deberta-v3-base loaded via the SalesCue library.

About SalesCue

SalesCue is a sales intelligence library with 12 ML modules sharing a single DeBERTa-v3-base encoder backbone. Modules can be composed via Unix-style piping:

from salescue import Document
result = Document("interested in pricing") | ai.score | ai.intent | ai.sentiment

All modules: score intent reply triggers icp objection sentiment spam entities call subject emailgen

See the SalesCue documentation for details.