Enzwa

The Enzwa AI pricing benchmark · September 2026

What AI enablement work costs.

Published prices from workshops to production builds. Separated by the work being bought, because “AI consulting” is not one comparable service.

Last reviewed 8 Sep 2026Next review Dec 2026Cadence QuarterlyDisplay currency HKD
90advertised offers
43providers
13countries
8offer types
Advertised prices, not winning bids.

This is a transparent sample of prices a buyer can find without a sales call. It does not include negotiated discounts, private proposals, software licences or pass-through cloud costs.

The market pattern

Compare like with like.

Each line shows the observed range and median midpoint for directly published, engagement-priced packages. Per-person, daily and monthly rates are kept in the collection below but excluded here.

Loading benchmark…

The long ranges are the point. A low-priced readiness product may be a questionnaire. A high-priced assessment may include interviews, process mapping, data review, governance and a working prototype.

How to read it

The label is not the scope.

  1. Workshop and training builds awareness or capability. It does not necessarily diagnose a workflow.
  2. Readiness assessment tests whether the organisation, data and controls are ready.
  3. Workflow diagnostic maps one repeated task and defines where AI may help.
  4. Pilot or proof of concept tests a bounded solution with representative work.
  5. Production build connects the solution to real systems, users and support.
  6. Adoption and governance covers behaviour, ownership, policy and assurance around the tool.

The collection

See every advertised offer.

Filter the evidence by market, work type or industry. Every row links to its source and retains the supplier's original currency and buying unit.

Loading collection…

ProviderMarketWork typeOffer and scopeAdvertised priceApprox. HKDSource

Method

A benchmark with its workings visible.

Prices are captured in the supplier's stated currency, then converted to HKD using cross-rates derived from the European Central Bank reference rates published for 7 September 2026. The benchmark uses the midpoint where a supplier publishes a range.

Supplier-written market guides remain visible but are excluded from the core category medians. So are per-person, per-day and per-month prices. Scope still varies, and currency conversion does not adjust for local purchasing power.

Coverage: publicly advertised prices located in the sources reviewed. This is not a census of every supplier and it under-represents larger firms that publish no price.

Start with one workflow.

The benchmark gives context. The AI Workflow Reset finds the repeated task, the usable sources and the review controls before a pilot is scoped.