TARGET MODEL

Design your target AI stack.

Define the capabilities your org needs across security, observability, gateways, agents, and more. Get ranked vendor recommendations with gap analysis. For CIOs and platform leads building platforms — not pilots.

TARGET MODEL · SAMPLE

AI Platform — 2026 Q2

6 CATEGORIES
Frontier Models
Anthropic
COVERED
AI Gateway
GAP · 3 candidates
GAP
Observability
Arize
COVERED
Evaluations
LangSmith · fills 4/6
PARTIAL
Agent Framework
GAP · 5 candidates
GAP
Vector DB
Pinecone
COVERED
● 3 covered● 1 partial● 2 gaps

A target model is not a tool. It's the platform specification your team argues about, signs off on, and refreshes every quarter.

Enterprise AI platforms aren't built by picking vendors one-by-one. They're built by deciding which capabilities matter, which gaps are tolerable, and which tradeoffs your organization accepts. Target Model structures that decision into a single artifact.

THE WORKFLOW

Baseline. Gaps. Recommendations.

A structured three-step workflow. One artifact the whole team can review.

01

Baseline

Capture what you have today. Pull vendors from your current stack, mark capabilities as in-place, or start from a template aligned to your industry.

02

Gaps

Define what you need — mark each capability required, optional, or not-needed. Coverage gaps surface automatically with severity and impact context.

03

Recommendations

See vendors ranked by how many of your gaps each fills — not by capability count. Add to your target stack in one click, re-run as requirements evolve.

WHO IT'S FOR

Three roles. One shared artifact.

CIO · Head of AI

Standing up an AI platform the whole org will use. Needs a defensible artifact that procurement, legal, and security can sign off on.

Platform Lead

Owns the AI stack across multiple product teams. Needs to consolidate, reduce vendor sprawl, and decide which capabilities are core vs. optional.

AI Program Manager

Running quarterly reviews. Needs to track what changed, surface new gaps, and keep the target model in sync with evolving requirements.

LIVING ARTIFACT

A document that travels with your team.

Export it as an Architecture Decision Record. Share it. Re-run it. Target models aren't frozen spreadsheets — they evolve as your stack, vendors, and requirements change.

ADR.md
# ADR-0042: AI Platform Target Model Q2 2026
Status: Approved · Stakeholders: CIO, CISO, Platform Lead
## Decision
Adopt Anthropic Claude for frontier reasoning. Defer AI Gateway to Q3 evaluation. Maintain Pinecone for vector storage.
## Context
Current gaps:
· AI Gateway — 3 candidates under evaluation
· Agent Framework — 5 candidates shortlisted
· Observability — partially covered (Arize)
MARKDOWNPDFCONFLUENCE
CADENCE · HOW IT EVOLVES

One year in the life of a target model

Q1
Baseline captured
Q2
First gap analysis · CIO reviewNOW
Q3
Gateway vendor selected · re-run
Q4
Migration complete · Q1 2027 planning
Export to Architecture Decision Record (Markdown) or PDF
Share with engineering leads, procurement, and security teams
Re-run quarterly as a living document — track what changed and why
History of decisions — defensible to any stakeholder, any time
PRO FEATURE
Target Model gap analysis and recommendations require a Pro plan — currently in private Beta.
Request access

Frequently asked

Baseline (mapping your current stack) is free. Gap analysis and ranked recommendations are Pro features.