Methodology

How We Score AI Vendors

Transparent, independent evaluation of AI solutions. Every score is backed by research, and you can customize weights to match your priorities.

Three-Level Capability Scoring

Each solution is evaluated against capability dimensions specific to its category. Every dimension receives a score of 0, 1, or 2:

FULL

Core, production-grade capability. Vendor actively markets and supports this feature with documented evidence.

PARTIAL

Capability exists but is limited, emerging, requires partner integration, or lacks maturity compared to leaders.

NONE

No evidence of this capability in the vendor's current product offering or public documentation.

What We Evaluate

Solutions are assessed on two complementary dimensions: capability scoring and AI due diligence.

Capability Dimensions

Each category has 6–13 specific capability dimensions, grouped into clusters that represent functional areas. For example, Fraud Detection & Prevention has three clusters:

Detection Engine (5 dims)Identity & Device (3 dims)Operations & Coverage (5 dims)

AI Due Diligence Framework

Beyond capability scoring, every vendor company is profiled across 6 strategic dimensions. This framework helps buyers assess the depth and sustainability of a vendor's AI investment — not just what they do today, but how defensible their position is.

1. AI Classification

How central is AI to the vendor's product? This is the most important strategic signal — it determines whether AI is the product or just a feature.

NAI-Native

AI is the core product. The company was founded to solve problems with AI, and the product wouldn't exist without it. Proprietary models, unique training data, and continuous learning are typical.

e.g., Featurespace, Recursion, Viz.ai

AAI-Augmented

An established product significantly enhanced with AI. AI adds meaningful differentiation but the core product predates it. May use a mix of proprietary and third-party models.

e.g., BlackLine, Sage, nCino

OAI-Overlay

AI layered on top via third-party APIs (e.g., OpenAI, Anthropic). Low switching cost for the vendor, minimal proprietary model investment. The underlying product could function without AI.

e.g., Generic SaaS + ChatGPT wrapper

2. Dependency Risk

How reliant is the vendor on external AI providers? High dependency means the vendor's AI capabilities could be disrupted by pricing changes, API deprecation, or policy shifts from their upstream provider.

LLow Risk

Proprietary models trained on proprietary data. No critical dependency on external LLM providers. Full control over model updates, latency, and cost.

MMedium Risk

Mix of proprietary and third-party models. May use external APIs for some features but has fallback options or is actively building in-house capabilities.

HHigh Risk

Core AI functionality depends on external providers (OpenAI, Anthropic, etc.). Vendor has limited control over model quality, pricing, or availability.

3Model Ownership

Does the vendor train and own their models, fine-tune open-source models, or call third-party APIs? Ownership affects differentiation, data privacy, and long-term defensibility.

Proprietary → Fine-tuned → Hybrid → API-dependent

4Data Moat

Does the vendor have unique, proprietary data assets that create a competitive advantage? Vendors with strong data moats improve over time as they accumulate more domain-specific data.

Very High → High → Medium → Low

5Continuous Learning

Do the vendor's models improve from real-world usage? Continuous learning creates a flywheel where the product gets better as more customers use it, widening the gap with competitors.

Yes → Limited → No

Data Sources

Scores are based on multiple independent sources evaluated by Outpace analysts:

Vendor Documentation

Product pages, API docs, whitepapers, and release notes

Analyst Reports

Gartner, Forrester, IDC, CB Insights, and domain-specific analysts

Technical Evaluation

Hands-on testing via free tiers, demos, and sandbox environments

Customer Evidence

Published case studies, G2/Gartner Peer Insights reviews

Market Signals

Funding, acquisitions, partnerships, team composition, patents

Customise Your Scoring

Default scores represent our independent assessment, but every organisation has different priorities:

Capability Weighting

Adjust the weight of each capability dimension from 0× to 5× in the Matrix. Set critical capabilities high and irrelevant ones to zero — rankings recalculate in real-time.

Score Solutions

Create your own evaluation criteria, score vendors against them with your team, and produce weighted comparison reports tailored to your procurement process.

Update Cadence & Independence

Quarterly Reviews

Scores are reviewed quarterly and updated when vendors ship significant product changes. Freshness indicators show whether category data is current.

Independent & Transparent

We do not accept payment for scores or rankings. All evaluations are independent. If you believe a score is inaccurate, contact us to request a review.

Ready to evaluate AI vendors?

Start comparing solutions across 66 categories with full capability scoring.

Explore Categories Sign Up Free