R² AI Labs · AI productivity, measured

How much working time can AI return to your organisation?

We answer it by measuring first: your work, role by role, and the workflows that would return the time. Then we design the Blueprint, the plan that decides where the time comes back and draws the workflows that return it.

  • The BlueprintIt is yours to keep, whoever builds it.
  • The researchThe same method, run across whole industries, is AIPE, our research program on AI, productivity and economics, published for anyone to read.
See the researchor see how we work
A demonstration · one workflowReady
or pick a role instead
What we have deliveredSix results · four engagements
Major insurance provider96.7%answer accuracy, with every citation shown in the source PDF
Digital product agency70%efficiency gain, measured inside two weeks
Specialist eye hospital15,000/mopatients triaged by an expert voice agent
National League Football Club1 weekto a working prototype, against the technology of far bigger clubs
Major insurance provider25%compliance overhead, reduced
National League Football ClubZerowearable sensors: vision alone carries the model
Fig. 01Six results, tagged

Each result was measured inside the engagement it came from. The case files are in selected work.

What we have found

AI fitted to the work could return about 25 of every 100 paid hours in Australia. Tools bought off the shelf return about 6 of them.

Both figures cover all 19 industries, weight each job by what it is paid, and count an hour as returned only once a person has checked AI's output. By industry, bought tools return from about 3 to 9.

Observation sheet

AIPE research · Australia
about 25
of every 100 paid hours AI fitted to the work could return across Australia, with each job weighted by what it is paid
It is a value at today's wages, not a forecast and not any organisation's own figure. Tools bought off the shelf return about 6; that is the least certain of our estimates.
AIPE research · Australia
8,168
task statements across 364 jobs in 19 industries, read one by one from the government's own published task lists
Our reading of each published task, not an observation of AI in any workplace.
AIPE research · United States survey
fewer than half
of their tasks is what surveyed workers wanted automated
It comes from a survey of workers in the United States. We count the work people want to keep as work that stays with them.
Fig. 02From the researchestimates from public data; the working is on the research pages

It is a value at today's wages: not a forecast, not a count of jobs, and not any organisation's own figure. What AI could return to your company or organisation depends on its own mix of roles.

  • ScopePhysical work is counted, as work that stays with people. The time AI could return comes from work done with words, records and decisions.

See the figures by industry

/ noun /

com·mer·cial·i·sa·tion lab

  1. a lab whose research measures how much working time AI could return to whole industries, and whose engagements return it to one organisation at a time.
  2. the step between research and revenue, where a published finding becomes a working system.

see also: consultancy, n. (what this is not)

Fig. 03The term, defined
What we are

Not a consultancy.
A commercialisation lab.

who we are, in full

Selected work

Four problems, four working systems.

Case files

Case IDigital product agency

Agentic engineering ecosystem

“Twelve-hour hackathons, two weeks, zero margin for a broken build.”

open the case file
70%efficiency gain2 wksduration
Fig. 04The case filesfour engagements, one folder each

all four case studies, in detail

How we work

Three ways we work:
Measure, Design, Realise.

Every engagement runs in the same order. Every fee is stated in writing before a stage begins, and what each stage produces is yours to keep: the Reading's findings, the Blueprint's drawings and its quote.

Transformation blueprintSheet 1 / 1
No. 1Readingyour organisation measured, from your own records; the number every plan starts from
No. 2Blueprintthe design: where the time comes back, what each workflow needs, and a quote that holds unless the scope changes
No. 3Buildone way to realise the design, built at the quoted price
Yours to keep

Issued toyour organisation

R² AI Labs · commercialisation lab

Drawn: R² · Checked: your people

Rev.

B

Fig. 05The Blueprint sheetone per engagement, revised with your people

Issued with this drawing

Issued toyour organisationForto build from, by anyone
From earlier engagementsPlannd · LeaderLab · Gate · SalesLab · Beena
  1. No. 1MeasureThe Reading: a measurement of your organisation, from your own records and conversations with your people. It says where the hours are, and it begins with a conversation.
  2. No. 2DesignThe Blueprint: where the value is decided and drawn. We redraw your workflows around AI and price the work, so you decide with the figures in front of you.
  3. No. 3RealiseThe Build: we build what the Blueprint draws, at the price it quoted, if you choose to build with us. Our five products began the same way.
Next steps

Start with the research, or with the work in front of you.

The findings are open to anyone: how much working time AI could return, industry by industry, and what it means for yours. When you want the same measurement taken inside your own organisation, tell us about the work and we will say where we would start.

Read the AIPE research
Fig. 06The researchpublished for every industry