BID Knowledgebase & Architecture Guide

Platform Documentation

Comprehensive guide to the Business Intelligence Desk (BID) autonomous engine: from the continuous 8-phase intelligence loop and real-time competitor monitors to the ACT decision framework and zero-hallucination synthesis.

Current Version: v2.4 (Enterprise Engine) Last Updated: September 2026 Eventro Technologies Limited
01

System Architecture & Design Philosophy

The Business Intelligence Desk (BID) is engineered to solve the intelligence bottleneck experienced by commercial firms and SMEs that lack an in-house enterprise research department. Unlike traditional dashboard tools that simply report internal static numbers, BID is an autonomous intelligence operating system that continually searches, validates, contextualizes, and converts external market movements into prioritized executive decisions.

Client & Interface Layer

Responsive mobile-first web app with Progressive Web App (PWA) offline capabilities, thumb-friendly navigation, and interactive Apache ECharts canvas visualizations.

Rails 8 Core Application

Strict multi-tenant isolation via Current.organization, transactional background workers on GoodJob, and robust RESTful workflows.

Autonomous Intelligence Engine

Multi-engine web harvesting (SerpApi, Tavily, Brave Search) paired with anti-injection PromptDefense™ and multi-provider LLM task routers (Gemini, OpenAI, Anthropic).

// Architectural Hierarchy & Data Flow
[External Market Data] ──► [Continuous Ingestion / Crawlers]
                                       │
                                       ▼
                       [Deduplication & Fact Extraction]
                                       │
                                       ▼
                        [Cross-Verification Engine]
                                       │
                    ┌──────────────────┴──────────────────┐
                    ▼                                     ▼
        [Verified Fact Library]               [Business Pulse Calculator]
                    │                                     │
                    └──────────────────┬──────────────────┘
                                       ▼
                     [ACT Framework: Assess, Choose, Trigger]
                                       │
                                       ▼
                    [Executive Decision & Operational Playbook]
            
02

The 8-Phase Intelligence Loop

Every piece of data ingested into BID undergoes a rigorous eight-phase verification and operationalization pipeline before appearing on an executive’s desk:

1

Automated Continuous Ingestion

Autonomous background workers execute scheduled queries across verified Tier-1 news outlets, regulatory registries, court dockets, tenders, and competitor digital surfaces.

2

Normalization & Deduplication

Raw HTML and syndicated press releases are stripped of noise and fingerprinted to prevent duplicate reports across multiple wire services.

3

Entity Resolution & Context Matching

Extracted organizations, executive names, product lines, and regulations are linked directly to your organization’s tracked Subjects and Business Profile.

4

Fact Extraction & Triangulation

Unstructured text is parsed into discrete claims: what happened, monetary values, date of occurrence, parties involved, and operational impact.

5

Multi-Source Cross-Verification

Claims must be substantiated by authoritative primary references. Items receive a calculated Confidence Score (High, Medium, Low).

6

Strategic Synthesis & Categorization

Verified items are mapped into six strategic categories: Competitor, Regulatory, Market Movement, Internal Operational, Customer Sentiment, or Emerging Technology.

7

Business Pulse Computation

The event is evaluated against your balance of risk and opportunity, instantly updating your 0-100 real-time composite Business Pulse score.

8

Executive Action Triggering (ACT)

Critical alerts automatically generate recommended mitigation options, draft decision matrices, and create assignable execution tasks for team leaders.

03

Configuring Continuous Monitors

Continuous Monitors are autonomous web sentinels configured by analysts or executives. Each monitor runs on an automated cadence, scouring authoritative web endpoints for developments concerning your competitors, suppliers, regulators, or market sectors.

Field Parameter Permitted Values Operational Impact
subject_id Entity ID (Competitor, Regulator, etc.) Scopes the monitor to an enrolled organization or subject profile.
frequency hourly, daily, weekly Sets background execution schedule in GoodJob advisory locks.
importance_threshold low, medium, critical Only items meeting or exceeding this threshold trigger email/in-app executive alerts.
query Freeform search string / operators Passed to multi-source search providers with automated site restrictions.
Monitoring Best Practice

Avoid overly generic single-word queries like "Energy". Instead, configure targeted intent queries such as "SolarMax commercial tariffs Lagos 2026" or "NERC mini-grid licensing changes" to minimize false positives and maximize signal density.

04

Verified Intelligence Library & Facts

The Intelligence Library serves as your desk’s immutable corporate memory. Every item represents an independently verified fact or strategic event, structured into explicit sections:

What Happened (The Fact)

Objective, concise statement of the occurrence, including quantitative changes (e.g. price drops, acquisition prices, regulatory deadlines).

Why It Matters (Operational Impact)

Contextual impact assessment specifically evaluated against your business profile, target markets, and competitive positioning.

Primary Evidence & Citations

Direct URL citations to Tier-1 publications, court filings, official gazettes, or verified corporate disclosures.

Confidence & Importance Matrix

Dual scoring: Confidence (empirical certainty of the evidence) paired with Importance (severity of business consequence).

05

Business Pulse Calculator (0–100 Index)

The Business Pulse is a real-time composite health index calculated by the Analytics::BusinessPulseCalculator service. It aggregates positive strategic momentum against pending risks and market friction.

80–100
High Momentum

Clear market advantage, low regulatory headwinds.

60–79
Stable Position

Steady performance with manageable competitive friction.

40–59
Moderate Pressure

Emerging competitor shifts or compliance deadlines require review.

0–39
Urgent Alert

Critical threats active. Immediate ACT playbook execution advised.

06

The ACT Decision Engine (Assess, Choose, Trigger)

Information without action is overhead. The ACT framework ensures that every high-impact intelligence discovery leads to a documented executive decision:

A

Assess

The engine evaluates internal vulnerability, commercial exposure, and competitor intent. It outlines potential scenarios and financial downsides.

C

Choose

BID formulates a weighted Decision Matrix comparing 2–3 viable paths (e.g. price match vs. value bundling vs. regulatory appeal) with cost-benefit analysis.

T

Trigger

Once the executive approves an option, BID generates an operational playbook with assigned tasks, deadlines, and milestone checkpoints for your operational team.

07

Ask BID (Executive Synthesis & Anti-Hallucination)

Ask BID allows executives and analysts to converse with their intelligence library in natural language. Powered by our proprietary PromptDefense™ architecture, responses are strictly grounded in your tenant’s verified facts.

PromptDefense™ Security Guarantees:
  • Zero Cross-Tenant Leakage: Injected context is exclusively extracted from current_organization.intelligence_items.
  • Prompt Injection Hardening: Raw web crawler text is quarantined and sanitized before being presented to the synthesis model.
  • Strict Citation Enforcement: If an answer cannot be grounded in verified library records, the engine explicitly declines to fabricate assumptions.
08

Intelligence Credits & Resource Economics

Every subscription plan includes a monthly allocation of Intelligence Credits. These credits power the compute and external search harvesting necessary for automated research:

Operation Credit Cost Description
Automated Monitor Run 1 Credit Multi-source crawler scan, deduplication, and fact ingestion.
Ask BID Query Synthesis 1 Credit Semantic retrieval across tenant library and citation-backed answer.
ACT Playbook Generation 2 Credits Assessment formulation, decision matrix, and action task breakdown.
Deep Research Dossier 5 Credits Comprehensive multi-chapter market research investigation with PDF export.

Unused credits rollover according to your active tier terms. Top-up bundles are available in increments of 100, 250, and 500 credits directly from the Billing tab.

09

Security, Data Isolation & Compliance

Enterprise trust is central to BID's engineering. We implement defence-in-depth protocols across every layer of the application stack:

Strict Database Row-Level Tenancy

All queries resolve through current_organization associations. Controllers never permit unqualified primary key lookups, systematically eliminating IDOR vulnerabilities.

AES-256 & TLS 1.3 Cryptography

All data at rest is encrypted with AES-256. In-flight transmission requires modern TLS 1.3 encryption with strict HTTP Strict Transport Security (HSTS).

Zero AI Model Training

Under no circumstances is your tenant data, search queries, or internal intelligence records used to train or fine-tune public foundation models.

Instant Data Export & Purge

Tenant owners have complete data autonomy: export your entire intelligence database at any time, or execute a permanent cryptographic purge upon cancellation.