How we work
From opportunity to a running system.
Every engagement runs through the same five stages, from a paid audit that finds the leverage to a retainer that keeps the result running. Where each workload runs — public, hybrid, or private — is decided per workload rather than once for the whole company.
The value ladder
Assess → Architect → Build → Operate → Enable.
Assess
Assess
Find the leverage.
Find the leverage.
A paid Opportunity Audit maps your processes, quantifies where time and money leak, checks your data-privacy posture and AI readiness, and returns a roadmap ranked by ROI and effort.
A scored opportunity map + a 90-day plan.
Architect
Architect
Design the right system.
Design the right system.
What to automate, which model, where it runs, how data is governed. We place each workload on the public → hybrid → private path and design topology, integrations, security, and evaluation up front.
An architecture doc, implementation plan, and cost model.
Build
Build
Ship it for real.
Ship it for real.
Workflow automation, public and hybrid AI systems, private deployments, knowledge assistants, and agents — built, documented, and handed over. Demoed early and often.
Working, documented, adopted systems.
Operate
Operate
Keep it running and improving.
Keep it running and improving.
We monitor, maintain, and evolve what we built — updates, model swaps, new workflows on request. Predictable, on a monthly retainer.
Uptime, monthly improvement, one number to call.
Enable
Enable
Hand off the capability.
Hand off the capability.
Workshops and enablement so your team can run and extend the systems themselves. No lock-in of knowledge — you own what we build together.
A trained team + runbooks.
What we build
The range, not one flavor of AI.
Workflow automation, knowledge assistants, agents with human oversight, private deployment, hybrid routing, evaluation, and enablement. Each one is on the solutions hub with a screenshot of it running.
Browse the capabilities →The Marain Method
Six steps, because adoption is the hard part.
The failure mode of automation is not technical — it is that nobody uses the thing. We treat the human and political shape of a workflow as an engineering constraint, from the first conversation.
01
Listen
Map the real workflow — formal, informal, political. Find who must adopt this, and what would make them resist.
02
Locate
Pinpoint the highest-ROI, lowest-adoption-risk opportunities. Quantify them.
03
Design
Architecture, data governance, eval plan. Privacy and human-in-the-loop by default.
04
Build
Iterative, demoed early. Ship a tracer-bullet slice first, then expand.
05
Land
Roll out with the humans: training, docs, champions, feedback. Adoption is a deliverable.
06
Sustain
Transition to managed. Measure realized ROI. Find the next opportunity.
The menu
Productized offers.
Opportunity Audit
1–2 weeks, fixed
$3,500
Automation Sprint
2–4 weeks, fixed scope
Tailored
AI System Build — public / hybrid
2–5 weeks
Tailored
Private AI Deployment
3–6 weeks
Tailored
Knowledge Assistant (RAG)
2–4 weeks
Tailored
Managed Automations
Monthly retainer
Tailored
Enablement Workshop
1–3 days
Tailored
Fractional AI Lead
Monthly retainer
Tailored
One concrete entry point; builds and retainers are fixed-fee, quoted after your audit — so you pay for scope, not guesswork.
Managed retainers.
The recurring layer — three tiers, from keeping it running to owning the roadmap with you.
Care
Monitoring, maintenance, and minor tweaks. Keep it running.
Grow
Care, plus a monthly allocation for new automations and improvements.
Partner
Grow, plus roadmap ownership, priority response, and quarterly strategy.
Not sure which you need? Start with the audit.
Two weeks, low risk, and its fee credits toward whatever you build next.