Pricing for AI outcomes. Not AI tooling.

Because every workflow is different, and any agency that quotes you a flat rate without scoping the work is either guessing or padding. The starting price is the honest floor for each tier. The fixed quote comes after we scope four things: workflow complexity and decision rules, run volume and model usage, the number and condition of your tool integrations, and your data readiness. The 60-day pilot surfaces all four, so the number you eventually sign is grounded in real usage, not a sales-deck assumption.

One agent system covering one defined workflow, deployed into your environment, managed by us, with a monthly ops report. It is built to prove the agent over longer ops cycles. It is not built to be your AI operating layer. Most customers start with Single Agent to remove their most painful recurring task, then decide whether to expand into Operator.

No. It means the work output of the agent systems in Operator is roughly equivalent to what a full-time employee would produce on the same set of repeatable tasks. The point is leverage, not headcount cuts. You still need humans for the judgment calls, the relationships, and the exceptions. The agents handle the volume so your humans can focus on higher-order problems.

Most businesses land in Operator. It is the recommended commercial center of gravity: two to three agent systems running in production, AI Visibility Foundation included, the work output of about one FTE. Growth makes sense once you have multiple business functions that need to coordinate through one AI layer. Command is for businesses that have decided to go all-in on AI and want a managed operator running it alongside them. If you are not sure, that is what the consult is for.

No. After we scope a workflow, Marshal builds, tests, and deploys the agent inside your environment on our dime. You get to observe it executing your work in real time. This proof step exists so you can evaluate agentic AI the way it should be evaluated: by watching it work, not by reading a deck.

The proof step is the build itself. We construct the agent, deploy it into your environment, and run it through enough cycles to demonstrate it can do the job. You observe. You pay nothing. The 60-day paid pilot is the next step if you want to extend that runtime in a formal, low-fee engagement so you and your team can build operational confidence over real cycles. Pilot pricing is small and scoped to your specific build. At the end of 60 days, you either roll into Operator (or one of the higher tiers) or you walk. There is no obligation either direction.

You walk. No subscription. No "but you owe us for the build" conversation. We carry the build cost because we are confident in what we ship, and we accept the consequences when we are wrong. The point of running pilots this way is for both sides to find out before either side is locked in.

Agent-First if the bottleneck is internal: too much manual work, slow follow-up, dropped handoffs, admin tax eating your team. You will feel the result in weeks. GEO-First if the bottleneck is external: buyers cannot find you, are not naming you, or are choosing competitors when they ask answer engines for recommendations. You will see early movement in roughly eight weeks and compounding from there. If both are broken, start with whichever is bleeding the most revenue right now. The other one is included once you are in Operator, Growth, or Command anyway.

Mostly no. AI Visibility is bundled into Operator, Growth, and Command because visibility and operations compound together. Foundation comes with Operator. Expansion comes with Growth. Full Coverage comes with Command. The published GEO-First path is the same set of programs, just sequenced from the GEO side. If you have a specific reason GEO has to be unbundled, raise it on the consult and we will tell you whether we can support it.

Not really. Once you move into Operator and above, GEO comes with the package because at that operating depth, visibility and execution start reinforcing each other and pulling them apart leaves money on the table.

Yes. The ladder is designed to be climbed. Most clients enter at Operator, then move up as more workflows earn their way in. Moving down is also possible if the scope of work changes.

A workflow-level agent that handles a complete operational outcome, not a single task. It combines an AI agent, automation logic, integrations, rules, logging, and human review where judgment is required. Example: a Lead Capture System that takes an inbound form, qualifies it, drafts a first-touch, routes it to the right owner, logs the CRM, and notifies Slack. That would be one agent system, even though it spans five tools and seven steps.

Workflows always change. Each system is built in modular pieces so individual steps can be swapped, expanded, or retired without rebuilding the whole thing. Post-launch optimization is included in the managed service, and changes are scoped as small follow-on work rather than full new builds.

No. There's no hard commitment. However, Operator, Growth, and Command are typically purchased on an annual agreement with monthly billing. Single Agent is a 60-day pilot, not an annual commitment. The annual term exists because the work compounds over operating cycles, and we have found that customers who commit to a year materially outperform competitors.

Yes. We build on the tools you already own. Standard integrations cover the common revenue, communication, and operations platforms (HubSpot, Gmail, Slack, Calendar, Notion, Salesforce, Apollo, Linear, Stripe, Intercom, Airtable, and the usual suspects). Advanced integrations covering deeper API work, custom middleware, or legacy systems are scoped during the pilot. No rip and replace, no forced migration.

Your core systems stay where they live. Your CRM, inbox, calendar, and project tools never move. Agents reach into those systems through credentials you control and can revoke at any time. The workflow logic, the agent runtime, and the orchestration layer run inside a secure, hosted environment with encryption in transit and at rest, scoped least-privilege access, and full audit logging on every run. Your data is not used to train models, ours or anyone else's. You retain ownership of everything that runs through the system.