OpenAI introduced Frontier in February 2026 as an enterprise platform for building and operating agents. Alongside Agentforce 360 and other ecosystems, it clarified the next architecture question: what belongs in an agent platform?
The model is only one runtime dependency. Enterprise value depends on identity, context, tools, policy, evaluation, deployment and an operating model that improves the system.
Cloud Group point of view
Choose an agent platform for how well it integrates with accountable work—not for a benchmark in isolation. Preserve strategic control of data contracts, business actions and outcome telemetry.
A practical playbook
The strongest next step is narrow enough to govern and useful enough to produce evidence. We would structure the work around these moves:
- Define platform capabilities before comparing vendor feature lists.
- Separate proprietary accelerators from portable enterprise contracts.
- Test identity propagation and action authorization end to end.
- Require evaluation, versioning and incident response as core services.
- Plan coexistence for native Salesforce and cross-enterprise agents.
The architecture and operating implication
Use an enterprise control plane to register agents, tools, models, policies and owners. Let domain platforms such as Salesforce remain authoritative for customer state and actions. A shared observability layer should connect work across platforms without forcing all execution into one stack.
Measure what changes
Model activity is not a business result. Track a small set of indicators that connect behavior to accountable work:
- Time to launch a governed agent
- Portability of actions and evaluations
- Cross-platform trace completeness
- Concentration risk and exit effort
The enterprise agent platform is an operating system for accountable digital work. Models will change; the organization’s ability to govern and improve work must persist.
Primary sources
This field note is grounded in the product and market context available at the time of publication.
