Where the money actually goes.

The build, broken into five lines

A single-workflow system splits roughly into scoping and design, integration, the operator interface, the automation itself and verification. In most engagements integration is the largest of the five and the model work is the smallest.

That ordering surprises buyers who have priced against a demo, because a demo skips integration entirely.

Running cost, per unit of work

Ask for the cost per enquiry, per call or per document, not per month. A monthly figure hides the fact that the cost scales with your volume, which is the thing you are hoping will grow.

Ask per enquiry. A monthly number hides the scaling.

Running cost has three parts: model inference, infrastructure, and any third-party services such as telephony or verification. Each should be a separate line.

The four lines usually missing

• Verification against a real baseline, which takes a week of somebody’s attention
• Your team’s time in scoping and review: real hours, worth costing
• Change budget for the first eight weeks, because real traffic finds cases the audit did not
• Documentation and handover, which is work and should be visible in the quote

A quote without these is not cheaper. It is less complete.

Why staged pricing protects you

Pricing each stage separately means you can stop after the audit, after scope or after design, and keep what was produced. It also forces the supplier to define what each stage delivers.

A single all-in number for a project nobody has scoped yet is a guess with a contract wrapped around it.

Comparing against the two alternatives

Off-the-shelf tools are cheaper and fit your workflow approximately; in-house builds fit exactly and cost three to nine months of your own team. A custom build sits between the two, and the honest comparison is against your own volume and the cost of the current manual process.

If the manual process costs less than the system, the correct answer is not to build. Nexvix says so at scope.