teradata.
·
AI Strategy
QS v0.1
3-minute briefer · Print-ready · Board-forwardable
From human analytics to agentic intelligence.
Every enterprise AI journey climbs the same six levels — and hits the same Wall between Level 3 and 4 where the LLM-with-search must become an enterprise agent. This is the briefer that maps your customer's position on the journey before you print the strategy deck.
The one-sentence version
Grounding is the layer nobody costed for. 77% of enterprise data is not yet contextualized for agents to use reliably.
— The teachable moment from The Ascent
4-layer reference architecture · 27 requirements
Every AI stack has the same four layers.
Bottom-up. Knowledge Layer as foundation. 27 functional requirements from Forrester, Gartner, IDC.
Requirements
27
L4
Policy LayerTrust · governance · agent identity
CISO · DPO
7 reqs
L3
Agentic LayerAgents reason · act · orchestrate
Chief AI Officer
7 reqs
L2
Translation LayerSemantic grounding · anti-hallucination
CDO · Head of Data
6 reqs
L1
Knowledge LayerFoundation · vectors · lineage · schema
CDO · Enterprise Architect
7 reqs
◤ Data flows upward
4 layers · 27 requirements · 6 levels
Reasoning at apex ▸
The 6-level journey · from analytics to agentsLevel 0 → 5
⚠ The Wall
L0
Human
L1
Assisted
L2
Search-first
L3
Grounded
L4
Agentic
L5
Autonomous
The framework is analyst-derived
Not a Teradata invention. Every requirement traces to a published analyst framework.
Forrester Data Fabric Q4 2025
Forrester Data Governance Solutions Q3 2025
Gartner D&A Governance MQ 2026
Gartner Semantic Layer for AI 2025
IDC MarketScape Unified AI Governance
Gartner AI Engineering 2025
Analyst-grounded framework. Not a marketing deck. Procurement will accept it.
The Teradata answer1 platform · 4 layers
The Teradata Autonomous Knowledge Platform maps directly onto the 4-layer architecture — with Knowledge Layer as the foundation where nobody else invested.
L1 Enterprise Vector Store · CDF · Context Engine
L2 Teradata Fabric · semantic layer · in-DB
L3 Teradata AI Studio · Tera · ModelOps
L4 RBAC · lineage · audit · agent identity
Two hours. One scored roadmap. No further commitment.
The AI Maturity Assessment Workshop: 30-min Self-Assessment · 30-min Use-Case Studio · 30-min Implementation Roadmap. Run interactively from the Maturity Model in this pack. You keep the export.
