AI Signal Detection
Finds observable strategic, technology, hiring, product, and organisational change signals from public company sources.
AI Application · Built by Santosh Pania
GTM IntelligenceAI Account Intent Radar
An evidence-led AI application that converts observable public company signals into explainable account intelligence for enterprise GTM teams.
Public sources onlyEvidence for every insightExplainable scoringNo confirmed-intent claims
The GTM problem
Account research is abundant. Useful evidence is not.
Sellers can find thousands of public data points about an account, but most workflows still leave the hard work to a human: deciding what changed, whether it matters to the solution, which evidence is repetitive, and what action should follow.
GTM Intelligence was designed to make that reasoning visible. It does not pretend that public evidence proves buying intent. Instead, it identifies observable change and shows exactly why a signal may matter.
Core output
AI-Derived Account Propensity Score
0–100
Weighted across solution relevance, signal strength, recency, category diversity, and source confidence, with diminishing contribution from repetitive evidence.
What it does
AI applied to a real revenue workflow
Evidence-Based Scoring
Produces an explainable 0–100 account propensity score using solution relevance, signal strength, recency, diversity, and source confidence.
Persona & Outreach Intelligence
Translates evidence into relevant stakeholder personas, outreach angles, and next-best actions for enterprise GTM teams.
Automated Account Research
Inspects public company pages, classifies evidence, removes duplicate signals, and keeps every recommendation traceable to a source.
How it works
From public evidence to GTM action
01
Discover
Inspect public pages on the target company website within strict crawl and safety boundaries.
02
Classify
Identify observable account signals and classify them into a GTM-relevant taxonomy.
03
Score
Measure solution relevance and calculate an explainable propensity score rather than claiming hidden buying intent.
04
Act
Recommend personas, outreach angles, and a next-best action tied back to the supporting evidence.
Architecture
Built as a real application, not a prompt demo
The prototype separates the user experience, analysis service, evidence processing, model adapters, and persistence layer so the AI logic can evolve without rewriting the product surface.
Next.js + TypeScript
Account research interface and evidence dashboard
FastAPI
Analysis orchestration and REST API
Safe crawler
robots.txt-aware public web discovery with SSRF protection
Replaceable AI adapters
Local or transformer-based classification and semantic relevance
SQLite → PostgreSQL-ready
Normalized evidence and analysis storage
Traceable recommendations
Sources, excerpts, relevance, and GTM guidance retained together
Why this matters
AI should make GTM reasoning more useful — and more accountable.
GTM Intelligence is one example of how I approach AI: start with a commercial decision, preserve the evidence, and make the system explain its recommendation.