SphereIQ by Sphere

Search every approved document. Get answers with sources.

SphereIQ runs semantic search across your internal documents, contracts, policies, procedures, and knowledge bases, then returns answers with source citations and confidence scoring on every response. Teams use AI across approved organisational knowledge without falling back on unsupported answers or uncontrolled public tools.

• – Sourced answers from internal documents
• – Confidence scores on every response
• – Multi-turn conversations with image and file attachments
• – Document-level access controls by team and role
• – Semantic search powered by pgvector embeddings
• – OpenAI, Anthropic, Mistral, Llama, and any OpenAI-compatible model

Classify, document, and prove conformity — without starting from a blank page.

The wizard guides teams through AI system classification against Article 5 prohibited practices, Annex I Union harmonisation legislation, Annex III high-risk categories, and GPAI obligations under Articles 51–56. The output is a conformity checklist, an exportable evidence package, and a maintained AI system registry.

• – Risk classification against Article 5, Annex I, Annex III
• – GPAI obligation mapping under Articles 51–56
• – AI system inventory and registry
• – DPO-ready reports and conformity checklists
• – Exportable evidence packages for board, auditor, or regulator
• – Multi-regulation mapping to NIST AI RMF and ISO 42001

Stop dangerous prompts before they reach the model.

Bulwark screens every prompt and response in real time for PII, prompt injection, unsafe content, and policy violations. Every query, every action, every administrative change is recorded in an immutable audit log — giving security and compliance teams full traceability across AI usage.

• – Real-time PII detection
• – Prompt injection scoring and blocking
• – Unsafe content and policy enforcement per team
• – 40+ regulation-aligned policy templates
• – GDPR, HIPAA, SOC 2, MiFID II, DORA, PCI-DSS, NIST AI RMF, NIS2, and more
• – Immutable, tamper-evident audit logs
• – Exportable governance records and admin activity tracking

The AI emissions figure your sustainability report needs.

SphereIQ measures AI-related emissions by model, team, and user — including CO₂g per 1,000 tokens. Sustainability and finance teams export AI usage data for ESRS E1 reporting cycles, replacing manual data collection with structured, defensible reporting.

• – CO₂g per 1,000 tokens, per model, per team, per user
• – Quarterly export workflow aligned to ESRS E1
• – Methodology documentation with cited sources
• – Auditor support package — raw logs, factor sources, grid data
• – CFO and sustainability dashboard with euros and grams side by side
• – Year-over-year reporting and reconciliation

Conversations become structured organisational memory.

Most AI tools forget everything between sessions. Engram captures facts, entities, decisions, events, insights, and preferences as structured memory — so the AI environment becomes more useful every week instead of restarting every Monday. Every memory is visible, manageable, and governed.

• – Six memory types: facts, entities, decisions, events, insights, preferences
• – Persistent memory across sessions, teams, and devices
• – Semantic Recall Lab for administrator inspection and tuning
• – Memory maturity stages from capture to compounding
• – Native connectors to SharePoint, Jira, Gmail, Google Calendar, Outlook
• – User-visible memory dashboard with edit, delete, and export controls