What's the real return on AI upskilling?
A transparent model — no black boxes. Move the sliders and watch the numbers. Every constant is anchored to U.S. BLS labor data and peer-reviewed productivity research, shown in full below.
Your team
Adjust to match your reality
faster on AI-suitable tasks — the default here is anchored to the Harvard Business School + BCG field experiment. Adjust the slider below to model your own scenario.
Team size
People you'll upskill
Productivity gain on suitable work
How much faster AI-suitable tasks get done
Share of the week that's AI-suitable
Portion of work open to AI-enabled redesign
Hours worked per week
Per person, full-time basis
Loaded hourly cost
BLS employer cost per hour — edit to match your team
Training price per person
Set your per-person budget — custom proposals, engagements from $9,999
The exact math, live
Why 2.5 hrs/wk? A 25.1% gain applied to just 25% of a 40-hour week. The default 25.1% is the Harvard Business School + BCG field-experiment result — consultants worked 25.1% faster on AI-suitable tasks.
Why $46.15/hr? The U.S. BLS all-in employer cost — wages plus benefits — for private-industry workers, so saved time is valued fully loaded, not just base salary. Over 46 working weeks that compounds into $5,328 of recovered capacity per person.
Annualised ROI
5.3×
Every $999 invested returns $5,328 in annual capacity per person.
$5,328
Annual value / person
2.5 hrs
Time saved / person / wk
$63.9K
Annual capacity · team of 12
10 wks
Payback period
Break-even is just 28 minutes saved per person, per week.
* Indicative, for a workflow-mapped AI upskilling program. Pricing is flexible — we run longer or shorter programs, so final scope and pricing are confirmed on your discovery call. Contact us for a proper quote.
We'll turn these numbers into a custom, sourced business case — no commitment.
Worked examples · conservative 15% scenario
The evidence
Peer-reviewed, not vibes
The productivity range in this calculator comes from two of the most-cited field studies on generative AI at work.
Generative AI at Work
NBER · Brynjolfsson, Li & Raymond
A study of 5,000+ customer-support agents found generative-AI assistance increased productivity by 15% on average, with the largest gains among less-experienced workers.
Read the studyNavigating the Jagged Technological Frontier
Harvard Business School + BCG
In a field experiment, consultants using GPT-4 completed 12.2% more tasks and finished 25.1% faster on tasks suited to AI, with quality gains too.
Read the study“We charge about $999 per person. In the Harvard/BCG field experiment, people worked 25.1% faster on AI-suitable tasks. Apply a 25.1% gain to only 25% of the week and that's 2.5 hrs saved per person per week — about $5,328 of annual value, or 5.3× ROI.”
Assumptions
- 46 working weeks / year. Net of typical PTO, holidays and ramp.
- $46.15 / hour loaded cost. BLS employer cost, private industry, Dec 2025.
- 25% of the week AI-suitable. Conservative share open to AI-enabled redesign.
- 25.1% productivity gain (default). Harvard/BCG: faster on AI-suitable tasks.
- Value = recovered capacity. Hours freed for higher-leverage work, valued at loaded cost.
Sources
- U.S. Bureau of Labor Statistics — Employer Costs for Employee CompensationPrivate-industry loaded hourly cost basis ($46.15/hr, Dec 2025 release).
- Brynjolfsson, Li & Raymond — Generative AI at Work (NBER w31161)Peer-reviewed field evidence: ~15% average productivity gain.
- Dell'Acqua et al. — Navigating the Jagged Technological Frontier (HBS 24-013)Harvard/BCG field experiment: +12.2% tasks, 25.1% faster on suitable work.
Estimates are directional and deliberately conservative. Actual results vary by team, workflow and adoption. Figures value recovered time at loaded employer cost, not incremental revenue.
Build the business case in 30 minutes
We'll tailor this model to your team, share a sourced ROI report, and walk through the training plan. No commitment.
Book a free discovery call