Fixed scope, 3–5 weeks
Local Failover Server + Private AI (On-Prem Resilience Appliance)
One small server at your location that does two jobs: failover and private AI. It doesn't replace the cloud. It's the fallback that keeps you open and keeps your team informed when the cloud has a bad day.
- 1. Failover. It keeps online ordering, booking, a basic site, or internal tools running when Cloudflare, AWS, or your main internet line goes down, then hands control back when they recover.
- 2. Private AI. It runs an AI assistant on your own data, policies, and runbooks, so staff can ask "what do we do right now?" or everyday questions without sending anything to the cloud. It keeps working even when the internet is down or flaky.
Want the short version? Own your Super Intelligence →
Interactive · try it
Pull the plug
Knock out a piece of the cloud and watch traffic reroute to the glowing box on site.
● Primary path online
- Simulated orders kept flowing
- 0
- Revenue kept (simulated)
- $0
Simulation, illustrative numbers
- Approach
- Hybrid, not rip-and-replace
- Scope
- Fixed-scope pilots
- Handover
- Tested runbooks
- Lock-in
- None
Demo, scripted answers
Ask the box
Cut the internet, then ask the private assistant a staff question. These answers are scripted for the demo; no real AI is called.
- Static demo. Example: a staff member asks "What do we do right now?" while the internet is down, and the box answers from your runbook.
What's included
- On-site assessment: which functions must stay up, your internet and power setup, what an hour down costs you, and the questions staff ask most.
- Hardware recommendation and sourcing: a compact server sized to run a local AI model, a battery backup (UPS), and a backup internet line (LTE/5G) if needed.
- Failover setup: a local copy of the critical services, data syncing from the cloud, secure remote access, and health checks plus DNS or routing rules that switch over and back automatically.
- Local AI setup: a private open-source model and assistant installed on the box, with no customer or patient data leaving your location.
- Data scoping: we agree on exactly what the AI can see (menus, policies, SOPs, runbooks, FAQs) and what stays off-limits, with access limited to your staff.
- Plain-English runbooks for outages, plus a short guide and a training session on how staff use the assistant.
- Basic monitoring: alerts if the box, the sync, the backup line, or the assistant stops working.
- A live test: we pull the plug on the cloud path with your team watching and confirm ordering or booking keeps running and the assistant still answers.
Illustrative starting range
$10k–$30k setup + about $2k–$6k hardware at cost
November Founders Price (5 Spots)
$5k–$15k setup (50% off) + hardware at cost
About $2k–$6k in hardware (at cost, with a bit more memory for the AI) plus $10k–$30k for setup, depending on how many functions and documents we cover.
Optional care plan: $750–$2.5k/month for patching, sync checks, quarterly failover tests, assistant updates when your policies or menus change, model upgrades, and hardware support. This can be bundled into the Resilience Retainer.
November Founding-Client Offer: 50% off listed setup ranges for 5 founding partners in exchange for a brief case study. Final pricing is scoped on a discovery call.
Best fit
- Restaurants and multi-location food businesses that depend on online ordering and need staff to answer "what now?" quickly when systems go down.
- Clinics and practices with online booking and check-in, especially those that would rather keep patient information off public AI tools.
- E-commerce and DTC brands with a warehouse or office that need order intake and fulfillment tools to keep running, plus an assistant for SOPs and returns policy.
- Any location with steady staff on site, a known cost per hour of downtime, and policies people look up every day.
How it pairs with the other packages
Audit first
The Audit shows whether a local box is worth it, or whether cloud-only failover is enough.
Core Failover
This covers the cloud side (secondary DNS, origin bypass). The local server adds a physical fallback at your location, and the two share the same health checks and runbooks.
Multi-Rail Payments
The local box can host a backup checkout path when your main processor or site is unreachable.
Retainer
Covers the care plan for both failover and the assistant, and includes the box in the quarterly tabletop exercises.
Example scenarios
Outages like these
Example scenarios, not client stories. These are the kinds of outages this work is meant to soften.
Example scenario
The cloud goes down during the lunch rush
Online ordering runs in the cloud. When it's unreachable, the box in the back office takes over the ordering page so the kitchen keeps getting tickets, then hands back when the cloud recovers.
Example scenario
Staff need answers while the internet is out
The internet line is down and a new hire asks what to do with walk-in orders. They ask the assistant on the local box, which answers from your own runbooks without anything leaving the building.
Questions
Do I have to leave Cloudflare, AWS, or my current tools?
No. Most work is hybrid. We add secondary paths and redundancy behind or alongside what you already use.
How do you use AI in this?
Plain-English runbooks, monitoring summaries, and assistants that answer "what's broken and what do we do?" from your own procedures and data.
What does good recovery look like for a small or mid-size business?
Depends on your revenue and risk. Common targets are RTO of minutes to a few hours for critical paths and near-zero data loss for key transactions. We define this with you on the discovery call.
Book a free discovery call
Every engagement starts with a free discovery call. 5 November founding-client spots available at 50% off listed prices.