How engagements are priced — a framework
PRICEDBYTHESYSTEM.
Pricing
Price follows the system you build.
We don't publish an off-the-shelf price list. The cost of automation isn't one line of model — it's the whole of server, logic, guardrails and LLM, plus the data and the volume you run. Instead, we're transparent about the axes that move the price.
ENGAGEMENT BANDS
Two bands: PoC and full build.
01
PoC — proof band (2–4 weeks)
A short, fixed-scope band that takes your single riskiest workflow and actually proves it can run unattended. Fixed scope, fixed timeline.
- · Number and complexity of the target workflows
- · Number of systems and data sources to integrate
- · Depth of the guardrails and approval gates required
02
Full build — build band (3–6 months)
We scale the proven line into a real operating system — building the whole of server, logic, guardrails and LLM, and handing it over as a custom SaaS the client owns.
- · Breadth of workflows automated and the volume processed
- · Deployment model (cloud / on-prem) and availability requirements
- · Operations and handover scope — co-operate or full transfer
COST DRIVERS
What moves the price.
Two jobs both called "automation" can price very differently along the axes below. We map them with you up front.
- · Scope — the number of workflows automated and the complexity of the steps
- · Data — sensitivity, regulation, how clean it is, and how many sources to integrate
- · Volume — number of items and frequency, and the infrastructure that implies
- · Deployment — cloud or on-prem. On-prem raises unit cost with in-house hardware and isolated inference, but your data stays in the building.
- · Guardrails — the depth of controls, approval gates and audit trail required
- · Operations — build and hand over, or run it together
SAAS vs FOUNDRY
Self-serve and custom build differ.
SaaS — QuelSuite · Vid.QuelSuite
Use a finished product right away. Predictable, usage-based cost; start in the browser with no build. The fast path when a standard workflow fits.
AI Creative Foundry — custom build
We build a dedicated line inside your team. Project-based cost — proven first by PoC, then scaled in a full build. The deliverable is a system the client owns.
FAQ
Frequently asked.
How much does AI automation cost?
There's no off-the-shelf price list. Cost is set not by the model but by scope, data, volume, deployment model, guardrail depth and operations scope. We prove your single riskiest workflow with a fixed-scope PoC (2–4 weeks), then price the full build (3–6 months) on the same basis.
What's the difference between a PoC and a full build?
A PoC is a 2–4 week, fixed-scope band that actually proves one workflow can run unattended. A full build is a 3–6 month build band that scales the proven line across server, logic, guardrails and LLM, handed over as a custom SaaS the client owns.
How does cloud pricing differ from on-prem?
On-prem requires in-house hardware and isolated inference, so its up-front unit cost runs higher than cloud — but data and model inference never leave the building. Private cloud starts faster and scales on usage. You choose based on your sensitivity and regulatory requirements.
NEXT STEP
A quote comes out of scope.
Tell us the workflows, the data and the deployment model you're thinking about, and we'll map the band and the cost drivers with you.