Skip to content

Template · Technology

Choose an AI platform with evidence for your needs.

Explore a fictional comparison of AI delivery approaches. Use the criteria to structure your evaluation, then replace all example scores, fit values and exclusions with evidence for your actual options.

Start a free trial · review example scores, fit values and exclusion rules before using the template

Loading interactive template…

What's inside

11 starting criteria · 5 starting options · a 9-question intake. For: Platform teams, architecture boards, and CTOs standardizing how business units pick AI tooling.

Cost efficiency
Value per dollar spent. Low (0-3): heavy licensing or build cost per outcome. High (8-10): near-zero marginal cost for the value delivered.
Scalability
Growth from pilot to org-wide or production volume. Low (0-3): struggles beyond a small team. High (8-10): proven enterprise/production scale.
Ease of use
How fast a typical user becomes productive. Low (0-3): engineers required for everything. High (8-10): business users self-serve on day one.
Data security & governance
Data protection, residency, access control, auditability. Low (0-3): data leaves your boundary with limited control. High (8-10): in-boundary with full governance and audit.
Integration & connectivity
Breadth and depth of connections to your systems (APIs, MCP, connectors, data platforms). Low (0-3): closed ecosystem. High (8-10): connects to nearly anything you run.
Customization & extensibility
How far you can adapt it beyond out-of-the-box behavior. Low (0-3): fixed product behavior. High (8-10): build almost anything on top of it.
Automation & orchestration
Support for multi-step, multi-agent, scheduled or event-driven workflows. Low (0-3): interactive use only. High (8-10): full orchestration with monitoring and retries.
Time to value
Time from decision to first real value. Low (0-3): months of build before anything useful. High (8-10): useful within days.
Model quality & reasoning
Intelligence of the underlying models, reasoning depth, code quality, accuracy. Low (0-3): basic assistance only. High (8-10): frontier-model performance.
Reliability & support
Uptime, SLAs, and support when things break. Low (0-3): best-effort internal support only. High (8-10): contractual SLAs with 24/7 vendor support.
Vendor lock-in & portability
How easy it is to leave later. Low (0-3): deeply tied to one vendor's ecosystem. High (8-10): portable, open standards, or you own the platform.

Starting options

These options and their ratings are illustrative. Renaming an option does not update its scores, fit values or exclusions. Review those assumptions for every option in your evaluation.

Office assistant (example) Coding assistant (example) Research assistant (example) Managed AI platform (example) Self-hosted models (example)

Build your evaluation.
Review it with your team.