Workflow Fit
Does the option operate inside how your teams already work, or require a new process to be learned and adopted?
The AI Layer
The method behind every recommendation we make.
Native or independent, the question is never which vendor. It is what clears the bar for your systems, your data, and your team.
The method
The Five-Factor Framework. The same five questions, every time, for every system.
Every AI option we evaluate—native platform intelligence, an independent specialist, or a custom agent—is scored against the same five factors before a recommendation is made. This is what turns “vendor-agnostic” from a claim into a repeatable method.
Data Boundary
Does the option read and write within your existing governance perimeter, or require data to leave it?
Security Posture
Does it meet access control, audit, and residency requirements without exception handling?
Licensing Economics
What does it cost at the scale you will actually run it at? Demo and production pricing are treated as different numbers.
Operating Ownership
Who runs it after go-live, under what SLA, and at what three-year commitment rather than pilot cost?
Score interpretation
A recommendation only reaches the client when an option clears all five factors at a defined threshold. An option that wins on capability but fails on data boundary or operating ownership does not get recommended, regardless of how strong the demo was.
Updated quarterly · Review date to confirm
The Landscape Index.
A working index of the AI options evaluated most recently against the Five-Factor Framework. This is a snapshot, not a permanent ranking: entries are added, removed, and re-scored as the market moves.
SAP
Joule and Joule Studio evaluated against custom agents on
Databricks or Snowflake.
Salesforce
Agentforce and Einstein evaluated against DevRev and
Sybill.
ServiceNow
Now Assist evaluated against Serval and Rezolve.ai.
Intelligent Automation
UiPath Autopilot and Maestro, and Automation Anywhere
Co-Pilot.
Data Platforms
Databricks Genie and Agent Bricks, and Snowflake Cortex and
Intelligence.
Engineering
GitHub Copilot, Claude Code, Cursor, and Devin, scored by task
type rather than by platform.
Request the full scored index for your systems.
Request the index →Maturity model
Three stages. Most enterprises can place themselves in under a minute.
The maturity stage is not a judgment of ambition. It shows whether AI decisions are still isolated purchases or part of a governed, replaceable enterprise layer.
Reactive
AI adoption is scattered across teams, chosen ad hoc, with no shared evaluation criteria. Each function defends its own tool choice. No one owns the AI layer as a whole.
Directed
A central team or mandate exists, but evaluation still happens vendor by vendor as requests arrive. There is a preferred list, but no repeatable scoring method behind it.
Architected
Options are evaluated against a consistent framework before adoption. Governance, ownership, and exit paths are defined before go-live, not after an incident.
Most enterprises we begin working with are between Stage 1 and Stage 2. The structured AI-readiness review identifies exactly where your organization sits and what moving one stage forward requires.
What ships with every engagement
Governance is a set of documents, not a sentiment.
Every AI layer engagement produces the following before go-live, giving teams a practical control system rather than a policy statement.
Decision Record
Five-Factor scoring for every option considered, including the ones not selected and why.
RACI Matrix
Named ownership for approval, override, notification, and accountability if an AI action fails.
Escalation Path
A defined route from an AI-driven decision to a human reviewer, with response-time commitments.
Audit Log Specification
What gets logged, for how long, and who can query it—agreed before the system goes live.
Exit Plan
What it takes to remove or replace the AI option without disrupting the core system underneath it.
Questions we get from CIOs before they commit
The practical questions behind a replaceable AI layer.
The architecture is built so the AI layer is replaceable without touching the core systems underneath it. When a vendor's position changes, we re-run the Five-Factor evaluation and migrate the edge, not the estate. This is what the Exit Plan is for.
The underlying platforms change rarely. The AI options layered on top change every few months. The Landscape Index is reviewed quarterly, and material shifts are flagged as part of the ongoing engagement—not as a separate re-sale conversation.
Both, depending on the Five-Factor outcome. Native and independent platform AI covers most use cases well. Custom agents are built when a use case genuinely requires reasoning across systems that no platform vendor connects natively.
See where your enterprise sits, and what moving forward requires.
The structured AI-readiness review applies the Five-Factor Framework to your systems. You receive your maturity stage, a scored evaluation, and governance artifacts drafted for your environment.