Sovereign vs Steward: When You Run It vs When We Run It For You

There is a question we get on almost every discovery call, and we want to answer it plainly before you have to ask it.

“I understand what Sovereign is. But I don’t have an AWS architect on staff. What are my options?”

The answer is Steward - the managed tier of the same platform. Same orchestration OS. Same plans-tasks-memory architecture. Same human approval gates. Different deployment model: we run it, you direct it.

This post explains both tiers, when each one fits, and how to move between them.

The Two Deployment Models

Sovereign and Steward share a codebase. They share a philosophy. They share the core belief that business operations should run on plans and durable state, not chat windows and one-shot prompts.

What they differ on is who owns and operates the infrastructure.

Sovereign is a full customer-owned deployment. Your AWS account. Your Secrets Manager. Your ECS cluster, your Aurora database, your EFS state volume. Prospectr deploys the initial runtime, trains the agent on your first one-to-three roles, and provides a retainer-based support tier for ongoing tuning and skill development. After deployment, the infrastructure is yours to operate.

Steward is a managed service on the same technology. You don’t need an AWS account. You don’t manage the infrastructure. Prospectr operates the shared (but tenant-isolated) platform, deploys the skills you select, and keeps the agent running. You get the operational output - the inbox triage, the reporting, the campaign monitoring, the lead routing - without the infrastructure responsibility.

Steward Pricing: Skills as the Unit

Steward is priced by skill count, not by seat or usage volume.

A skill is a defined operational domain. Inbox triage is a skill. AR follow-up is a skill. Lead pipeline management is a skill. Campaign monitoring is a skill. Each skill represents a tested, production-grade capability that the agent executes autonomously within your defined approval gates.

Tier	Skills	Monthly
Starter	1 skill	$500/mo
Growth	Up to 3 skills	$1,500/mo
Pro	Up to 5 skills	$2,500/mo

Custom skill development is available at $250/hr and is permanently added to the Sovereign skill library - every customer who uses it after you gets the benefit, and the skill stays in your deployment regardless of your tier.

Enterprise SLA (99.9% uptime + priority custom-skill queue) adds $1,999/mo.

LLM tokens and server costs pass through at Prospectr’s agency wholesale rate - no markup on compute.

Who Sovereign Is Built For

Sovereign is the right choice when three conditions are true:

You have (or can get) AWS infrastructure experience. You do not need a dedicated DevOps team, but you need someone who can manage an AWS account, review CloudFormation stacks, and handle the occasional infrastructure event. Many Sovereign customers have a technical co-founder, a part-time CTO, or an IT director who handles this.

You want full data ownership and portability. If your business handles sensitive data - healthcare-adjacent records, financial data, client-confidential information - having your agent’s memory and execution logs in your own AWS account may be a compliance requirement or a board-level preference. Sovereign gives you that.

You are planning to extend the agent into custom capabilities. If you know you will eventually want the agent to integrate with proprietary internal systems, build custom skill logic, or develop capabilities that are specific to your business model, starting with Sovereign gives you the infrastructure foundation to do that. Steward is extensible too, but Sovereign gives you the deepest access.

You need the hybrid model architecture. Direct Anthropic API access (for the latest Opus model) requires an API key in your Secrets Manager. That requires your AWS account. Steward uses the same Bedrock-based routing for the managed tier - if the absolute latest Opus is a priority, Sovereign is the path.

Who Steward Is Built For

Steward is the right choice when:

You want the operational output without the infrastructure management. Most businesses are not trying to run an AWS environment. They are trying to get their AR collected, their inbox triaged, their campaigns monitored, their leads routed to the right person. Steward gets you those outcomes without requiring you to become an AWS customer.

You are earlier in the agentic AI adoption curve. If you are deploying AI-powered operations for the first time, starting with a managed tier lets you learn what the agent is capable of, what your team’s approval gate preferences look like, and which skills drive the most value - before you commit to owning the infrastructure.

You want to validate ROI before full deployment. Steward’s Starter tier at $500/mo lets you test one skill in production. If inbox triage alone saves four hours of staff time per week, the ROI case for expanding to three skills is obvious. Sovereign’s deployment cost is a larger upfront commitment - Steward is the lower-friction path to validation.

Your ops complexity fits within a defined skill set. The Sovereign skill library covers the majority of operational needs for the businesses we work with: email triage, lead routing, campaign monitoring, AR follow-up, report generation, client onboarding sequences, and more. If your needs fit within five or fewer defined skill domains, Steward may be all you need.

The Migration Path: Starting Steward, Graduating to Sovereign

We designed the two tiers to be a graduation path, not a fork.

The operational logic - the SOUL files, the approval gate configurations, the skill parameter sets, the memory structures - is portable. A Steward deployment that matures over six months, with a trained agent and established operational patterns, can be migrated to a Sovereign deployment without starting from scratch.

The migration involves:

1. AWS account setup. We provision the target infrastructure stack via CloudFormation.
2. State migration. Agent memory, plan history, and skill configurations export from the Steward platform and import into your new Sovereign Aurora database.
3. Credential migration. You provision new API keys and OAuth tokens into your Secrets Manager. Prospectr provides the IAM templates.
4. Cutover. We run both environments in parallel for two weeks, verifying outputs match, before cutting the DNS and decommissioning the Steward tenant.

The agent you end up with on Sovereign is the same agent you built on Steward - same operational patterns, same approval gate calibrations, same skills - now running in your infrastructure.

The Reverse Path: Sovereign to Steward

Less common, but it happens. A company downsizes its technical team, or acquires a business unit that prefers managed services, or decides that infrastructure management is not where it wants to spend energy.

The same portability works in reverse. Sovereign state exports cleanly into the Steward platform. You move from owning the infrastructure to having us run it. The agent’s capabilities and memory come with it.

Choosing Right Now

If you are reading this and trying to decide, here is the shortest decision path we know:

Start with Steward if: you are doing this for the first time, you do not have a dedicated technical resource, or you want to validate one or two skill domains before committing to a full deployment.

Start with Sovereign if: you have an AWS account you already manage, you have compliance or data sovereignty requirements, or you know you want custom skill development in the near term.

When in doubt, take a discovery call. We have deployed agents across 40+ verticals and 225+ active clients through Prospectr’s own operations. The right answer for your specific situation usually becomes clear in the first conversation.

One More Thing: What You Are Not Buying

We should say this plainly.

You are not buying a chatbot, a no-code workflow tool, or a prompt-chain sequencer with a premium price tag. Both tiers - Steward and Sovereign - run on the same orchestration architecture: explicit plans, durable task queues, agent memory that persists across sessions, and reflection cycles before declaring work done.

The agent that runs your inbox triage on Steward’s Starter tier is architecturally the same agent that runs a 20-skill Sovereign deployment for a 200-person operations team. The capability scales with your skills and your infrastructure footprint. The OS underneath stays the same.

That is what we mean when we say the model is a component. The system is the product.

---

Ready to choose your tier? See full pricing for Steward and Sovereign, or book a discovery call to walk through your specific use case.

You can also read about the Sovereign Agent architecture, explore the Steward managed tier, or contact us directly at info@prospectrdigital.com or (612) 293-0179.

Prospectr Digital - founded 2006, based in Minneapolis, MN.