FAQ
Common questions,
straight answers.
Sovereign AI is a different category from cloud-based AI. Here are the questions we hear most often before a Discovery Session answered without jargon.
The Basics
What a sovereign agent is and isn't.
- Is this a robot?
- Not in the physical sense. It is software but software that can take real actions inside your systems the same way a person clicking a mouse would.
- Does it need the internet?
- No. A sovereign agent is designed to run entirely within your local network. It does not require an internet connection to do its work.
- How is this different from ChatGPT or other cloud AI tools?
- Cloud AI tools send your prompts and data to a third-party model running on someone else's infrastructure. A sovereign agent runs entirely on your hardware, processes your data inside your network, and never sends information outside your environment. The audit trail lives in your systems, not in someone else's logs.
- What hardware do we need?
- It depends on the agent's scope. A single workflow agent can run on a modest server inside your existing infrastructure. We assess your needs during the AI Readiness Assessment and recommend specs that fit the work — and your budget.
Control & Risk
Who's in charge, and what happens when things go wrong.
- What if it makes a mistake?
- Agents are configured with guardrails. They flag exceptions rather than making judgment calls they were not designed for. A human reviews anything outside the norm.
- Who controls it?
- You do. The agent only does what it is configured to do. You set the rules, you own the system, and you can turn it off, modify it, or expand it at any time.
- What happens to our data?
- It stays with you. Data is read inside your environment, processed inside your environment, and any output the agent produces is stored inside your environment. There is no data egress.
- How do you handle compliance and audit trails?
- Every action a sovereign agent takes is logged inside your environment — what it read, what it did, what it produced, and when. The logs are yours to retain, export, and present to auditors. We design for compliance from the first architecture decision, not patched in afterward.
Engagement & Scope
What working with us looks like in practice.
- How long does it take to set up?
- A single focused agent typically takes 2–8 weeks to build and deploy, depending on the complexity of the process it is handling.
- Can we start with one process and add more later?
- Yes — most clients do. We recommend starting with a single high-value process to prove the model in your environment, then expanding to additional agents once your team is comfortable working with them.
- How is the agent updated over time?
- You own the configuration, the tools, and the audit logs. Updates can be made by your team after the engagement, or we can provide ongoing managed operations under our Managed Agent Operations service. You decide.
- What does it cost?
- Engagements are scoped to the work. We start with a fixed-price 4-week AI Readiness Assessment. Single-agent deployments and multi-agent workflow systems are quoted after we understand your processes. Schedule a Discovery Session and we can walk you through it.
Still Have Questions
Let's talk it through.
A 60-minute Discovery Session is the fastest way to get answers specific to your operations.