AI Systems
Three solutions. Done in order.
Most engagements move through all three. Some clients stop after the first or the second and act on the work themselves. Both are fine.
Solution 01
Workflow Analysis
We sit with your team — owners, operators, the people doing the actual work — and document how the operation runs today. We are looking for two things: where time and money go, and where AI changes the math.
The output is concrete. A workflow map of your operation, with friction points marked, an estimate of where AI delivers leverage, and a candid assessment of what is not worth changing yet. If the answer is “your operation needs documentation before it's ready for AI,” we say so.
- What you provide
- Leadership and operator time, access to the systems and tools you currently use, and a willingness to look at how the work really runs (not how the org chart says it runs).
- What you get
- A working understanding of where AI fits in your business, and a proposal for the design phase if there is a strong fit.
- How long
- One to two weeks.
- How priced
- Fixed scope, quoted at the start. Not hourly. Specifics shared in the conversation.
Solution 02
AI System Design
Once the workflow is mapped, we design the AI system that will run on top of it. This is an architecture exercise: which workflow points get an AI agent, what each agent reads, what it produces, how it integrates with the tools you already use, where humans review, and how the whole system behaves under failure.
The Design phase covers software architecture, model strategy (which models, trained on what, hosted where), data and knowledge requirements, human-review checkpoints, change management for your team, build phasing, and a fixed estimate for the build phase.
- What you provide
- Stakeholder participation, sample data, decision authority on architectural choices.
- What you get
- A buildable, decision-ready AI system design. A document and a model. The design is yours to keep — you can act on it with us or with anyone else. We charge for the work, not for keeping the IP at the design stage.
- How long
- Three to six weeks.
- How priced
- Fixed scope, quoted at the start.
Solution 03
AI Build & Deployment
We build what was designed. End to end.
This includes: writing the AI software, training models on your data and rules, integrating with the systems you already run (your ERP, your CRM, your accounting, your custom tools), building human-review interfaces for the people who will work with the system, setting up monitoring, and training your team.
The build is phased. Each phase has a working deliverable — not a slide. Soft launch precedes hard launch. The first month after launch is included in the build engagement, with the team on hand for tuning, edge-case handling, and team support as people adjust to working differently.
What we don't do: replace your CRM, ERP, or accounting system. Build hardware unless it's part of an existing engineering deliverable (like the retail scanning system). Promise “autonomous” AI with no human in the loop. We build human-reviewed AI inside existing operations.
- What you provide
- Stakeholder participation through build phases, integration access, an internal champion who will own the system after launch.
- What you get
- A working AI system inside your operation, your team trained on it, monitoring in place, and the underlying engine maintained on an ongoing basis.
- How long
- Three to nine months, phased.
- How priced
- Phased and milestone-gated. Quoted in the Design phase based on agreed scope. Long-term support and tuning priced separately after deployment.
The full engagement
See how the work runs end to end.
The whole engagement, from first conversation to a working system, is laid out in Process — five steps, each with a clear deliverable.