Frequently asked questions
The questions CIOs, architects, and continuity leaders ask us most.
What exactly is AI Operational Resilience?
AI Operational Resilience (AIR) is the discipline of ensuring AI capabilities remain available, recoverable, governed, and testable as they become mission-critical infrastructure. It applies the methods of business continuity — inventory, impact analysis, recovery objectives, failover design, and testing — to the AI services your organization now depends on: hosted models, copilots, agents, and RAG pipelines.
How is this different from MLOps?
MLOps governs how models are trained, deployed, and versioned — it's a discipline for teams building models. AIR governs what happens when the AI services your business consumes fail: vendor outages, API throttling, model retirement, connectivity loss. Most organizations consuming AI through vendors have no MLOps problem at all, but every one of them has an AIR gap.
How is this different from our existing BC/DR program?
It's an extension of it, not a replacement. Your BC/DR program has the right methodology but was never pointed at AI services — they entered the organization through SaaS subscriptions and product features that bypassed procurement gates and business impact analyses. We bring the AI-specific knowledge (model routing, prompt portability, knowledge synchronization) that lets your existing continuity discipline cover this new asset class.
Are you selling us local LLM servers?
No. We are vendor-neutral and sell no hardware, models, or cloud contracts. Local inference is one of several failover strategies, appropriate only when a service's criticality justifies it. Many services need nothing more than a second cloud provider or a documented degraded mode. The Assessment determines which — before any architecture spend.
What does the Assessment cost and how long does it take?
The AIR Assessment runs $25K–$75K depending on the size of your AI estate and organizational scope, and takes 3–6 weeks. It's fixed-fee with defined deliverables: AI Service Inventory, criticality scores, dependency map, gap report, remediation roadmap, and an executive briefing. The findings are yours regardless of whether you engage us further.
We're technical — why not just build failover ourselves?
Some organizations can, and we'll say so when it's true. What internal teams typically lack isn't engineering skill — it's the continuity methodology: criticality scoring accepted by business owners, recovery objectives your auditors recognize, quarterly drill discipline, and reporting your board can act on. That operating model, not the routing code, is what makes resilience real and provable.
Which budget does this come from?
Most clients fund the Assessment from an existing line: business continuity, cyber resilience, enterprise architecture, or risk and compliance. The Assessment's executive briefing is specifically designed to make the ongoing investment case in the language those budget owners already use.
Do we need this if our AI vendor has an SLA?
Vendor SLAs refund fees; they don't restore your operations. No major AI provider offers availability guarantees matching tier-one enterprise expectations, and an SLA credit doesn't help the support queue that stalled for six hours. Resilience is your architecture's property, not your vendor's promise.
What happens after the Assessment?
Three paths, in any combination: your internal teams execute the roadmap themselves; we implement the failover architecture with you; or we operate the ongoing discipline through Managed AI Resilience — monitoring, quarterly testing, and executive reporting. The Assessment's roadmap sequences the work either way.
Still have questions?
Talk to an AIR architect — a practitioner, not a sales rep.