The Engine Inside EngageX

Meet CorteX AI

CorteX is the proprietary AI engine that powers EngageX, transforming raw event data into meaningful business opportunities in real time.

It goes far beyond simple matchmaking. By analysing profiles, intent signals, behaviour patterns, and interaction outcomes, CorteX helps the right visitors, exhibitors, and opportunities find each other with precision.

CorteX is not a separate product. It's the AI engine inside EngageX — what makes every match, every score, every voice transcript work.

Intent Analysis

Behavioural Intent Matching

Analyses browsing behaviour, session attendance, engagement patterns, and in-app actions to identify real interest — not just who participants are, but what they actively seek.

SIGNAL Real intent,
not just who
LIVE
Profile Intelligence

Multi-Signal Profile Enrichment

Enriches profiles with company intelligence, professional data, and event activity for a richer, more accurate picture of every participant.

Company Data Event Activity Intent Signals
Why CorteX Is Different

Not Just Matching.
Understanding.

The engine that thinks beyond keywords.

Most event platforms match people based on keywords or static profiles. CorteX goes deeper.

It understands business context, buying intent, engagement signals, and historical patterns to recommend connections that are more relevant, more timely, and more likely to create real value.

The result is a smarter event experience where participants spend less time searching and more time having meaningful conversations.

7×
Higher conversion
Real
-time learning
∞
Signal inputs
Active · Real-Time Intelligence Engine
Adaptive Learning

Real-Time Learning

Continuously improves as the event progresses — learning from meeting outcomes, lead signals, user feedback, and engagement quality to refine recommendations in real time.

Model accuracy↑ improving
Signal coverage98%
Data Security

Privacy-First Architecture

AES-256 encryption, privacy-conscious processing, and GDPR-aligned consent flows. Intelligence never comes at the expense of data security.

AES-256 GDPR Zero-Trust
GDPR-ALIGNED Privacy-first
by design
The CorteX Architecture

Three engines. One intelligence.

CorteX isn't a single model — it's three specialised engines purpose-built for how people actually connect at events. Together they capture voice, score relevance and read the room.

01
Speech

Whisper transcribes meeting audio live. Claude condenses every conversation into a 2-sentence summary, 3–5 key points, and 2–5 action items with owner and deadline.

Near real-timeEnd-to-end pipeline
In-memoryAudio never persisted
02
Matching

A 57-dimensional feature vector per person, L2-normalised and weighted for industry, geography, keyword overlap and company size. BAND qualification applies a multiplier. Cosine similarity ranks candidates in under 300 ms.

<300msPer-match latency
57-dimFeature vector
03
Vision

A business card, a QR code, a booth poster — Vision reads it all. Every capture returns a 0.0–1.0 confidence score so the UI adapts: auto-populate above 90%, pre-populate with review between 70–89%, manual below.

4 tiersConfidence-aware UX
In-memoryImages never persisted
CorteX in the console

Watch the model think.

Every prediction is debuggable. Model status, ICP match scoring, OCR confidence — all surfaced in the operator console.

CorteX show-thinking — model reasoning panel ICP matching — 57-dimension feature vector ranking Model status — latency, queue depth, and health OCR trainer — confidence-aware Vision capture
What CorteX Does

Nine capabilities. Zero black boxes.

Every capability is measurable, debuggable, and explainable. No magic — just well-engineered AI that tells you exactly what it did and why.

01Matching Engine

Behavioural + Profile Matching

57-dimensional vectors weighted by industry (2.0×), keyword hash (1.5×), geography (1.0×), and company size (0.5×). BAND tiers apply a qualification multiplier. Cosine similarity ranks candidates in under 300 ms.

94%Intent accuracy
02Speech Engine

Meeting Intelligence

Whisper transcribes every conversation. Claude distils a 2-sentence summary, 3–5 key points, and 2–5 action items with owner and deadline. Every meeting ends with a shareable debrief — no note-taking required.

5–30sEnd-to-end
03Vision Engine

Business Card OCR, Confidence-Aware

GPT-4o-mini returns a 0.0–1.0 confidence score on every field. The UI adapts to certainty — auto-populate, pre-populate with review, highlight uncertain, or fall back to manual. Transparent machine vision.

≥0.90Auto-populate
0.70–0.89Pre-populate + review
0.50–0.69Highlight uncertain
<0.50Manual entry
04Hybrid Scoring

Two-Layer Lead Scoring

Layer 1 — deterministic rules (seniority, company size, meeting depth, follow-ups, capture method, sessions) return a 0–100 score in under 100 ms. Layer 2 — Claude adds an ICP-fit narrative 3–8 s later. Hybrid AI, properly layered.

HOT WARM COOL COLD
05BAND, Explained

Score + Narrative, Not Just a Score

Every lead comes with a Claude-generated narrative explaining the score. Not a number in a void — a sentence your sales team can act on.

BAND A×1.00

"Director-level decision-maker at a 500-employee SaaS firm, attended three of your target sessions, active budget cycle in Q3 — engage immediately."

06Pre-event

ICP File Matching with 5-Step Reasoning

Upload exhibitor + visitor CSVs. Claude returns a 5-step reasoning chain per pair — Profile Scan, Industry Fit, Need & Solution, Strategic Opportunity, Composite Score. Transparent AI, shown step by step.

  1. 1Profile Scan
  2. 2Industry & Domain
  3. 3Need & Solution Fit
  4. 4Strategic Opportunity
  5. 5Composite Score
07Outreach Automation

Phase-Aware Virtual SDR

Three phases, three cadences. Pre-event: invitation flows. During-event: thank-you triggers. Post-event: no-show reschedules and nurturing. Every message is consent-gated — dispatch is never accidental.

Pre
During
Post
08Personalised Outreach

AI Follow-up Draft

Per-lead messages generated by Claude. Pick a tone (professional, warm, urgent) and a channel (email, LinkedIn). You get a subject, body, and opening line ready to edit and send.

ToneProfessionalWarmUrgent
ChannelEmailLinkedIn
09Market Intel

Session Intelligence & Role-Aware Chatbot

Sessions scored for each exhibitor profile: +30 industry match, +15 per keyword / product / solution match. HIGH_RELEVANCE ≥40, RELEVANT ≥15, TANGENTIAL otherwise. The chatbot adapts to your role — exhibitor, visitor, or organiser — and keeps conversation context up to 50 turns.

HIGH_RELEVANCE RELEVANT TANGENTIAL
Why You Can Trust CorteX

Three proofs, not three promises.

Every AI system claims to be private, reliable, and smart. These three sections show exactly how CorteX actually is.

Privacy-first

Your audio and images never touch our disks.

Speech and Vision process raw media in memory — never persisted. Only structured outputs (transcript, extracted fields) are stored, AES-256 encrypted at rest. GDPR, DIFC Data Protection, and UAE PDPL consent flows gate every dispatch.

  • AES-256 at rest · TLS 1.3 in transit
  • Audio + images never persisted
  • GDPR · DIFC DPL · UAE PDPL
  • 7-year append-only consent audit
Reliable by design

Every AI function has a deterministic fallback.

When a model is slow, offline, or uncertain, CorteX gracefully degrades. Matching falls back to keyword overlap. Speech falls back to raw transcript. Vision returns confidence 0 and hands to manual. Nothing ever hard-fails in front of a customer.

  • 9 AI functions · 9 deterministic fallbacks
  • Offline-first mobile capture
  • Async Claude enrichment — never blocks UX
  • Prompt caching — 30–40% token reduction
Learns, live

Every swipe teaches the next recommendation.

Match feedback — swipe right, swipe left, meeting booked, lead captured — is recorded and the Redis cache is purged. The very next request reflects the new signal. No overnight retrains. No batch jobs. Live learning.

  • Feedback → cache invalidation → fresh match
  • Collaborative filtering signals captured
  • Accuracy measurably improves per event
  • Post-event model tuning per organiser
Where you'll see CorteX

CorteX runs invisibly under every EngageX experience.

See it in action on the role pages:

Visitor — AI matchmaking → Exhibitor — BAND scoring → Organiser — Mission Control AI →
Get In Touch

Ready to transform
your next event?

Get in touch with our team and discover how EngageX connects visitors, exhibitors, and organisers with real purpose.

Phone / WhatsApp
+971 4 000 0000
HQ
Innovation Hub, AI, DIFC Dubai, UAE

Get in Touch with Us