The Most Undermeasured Number in Business

Every AI investment is a bridge with two towers. One tower is the technology: what AI does to the work. The other is your people: whether your workforce can actually put it to use. Two towers do not make a bridge. What carries the value across the gap between them is the architecture: the workflow redesign, governance, and adoption that turn a capable tool into a result on the income statement. This series takes each tower in turn, then the span that connects them. This part is about the second tower, your people.

When a quarter gets tight, the people investment is the first line cut. The coaching budget, the development program, the engagement work. They go first because no one has put a dollar on them, and a number nobody can defend is easy to erase.

That is the quiet problem with human capital. It is the largest lever most mid-market companies own, and the one they measure least.

Part 1 ended on a question: what are your people worth to the return on AI? This is the answer, and it runs larger than most leaders expect.

The value that is quietly bleeding out

Start with engagement, because it sets the ceiling on everything else.

Only 20% of employees worldwide are engaged at work, the second straight annual decline, with no region improving.¹ Low engagement costs the global economy around $10 trillion a year, roughly 9% of GDP.² Those are abstractions until you bring them home.

For a 100-person company, the calculator models a 10-point engagement gain as about $180,000 a year in recovered performance, using the 18% productivity premium that engaged teams show.³ Mid-market firms routinely carry 60-plus points of engagement headroom. The money is not missing. It is sitting inside disengagement, waiting.

What the evidence actually supports

Here is where human capital ROI (return on investment) stops being a slogan and starts being a floor.

Workplace coaching has a peer-reviewed effect on performance of about g=0.60 across meta-analyses.⁴ Leadership training moves organizational results at δ=0.72 across 335 samples.⁵ And 70% of the variance in a team's engagement traces to one person: the manager.⁶ These are effects established across hundreds of studies in peer-reviewed journals.

For the 100-person company, the tool turns that floor into dollars. Coaching ten managers returns about $173,000 a year, a figure the model has already halved for attribution honesty, and nets your $60,000 investment to roughly $113,000 a year. Retention is larger still. Five points of lower attrition is worth about $375,000 at a replacement cost of three-quarters of salary.⁷ That single line dwarfs most mid-market software budgets.

Why you can trust these numbers

Notice what the tool does with the figures it is least sure of.

It credits no return at all to generic wellness programs, because two large randomized controlled trials found no significant effect on spending, absence, or performance.⁸ And it holds financial-wellbeing benefits to a small, capped figure it labels directional, because the clean causal proof does not exist yet.

That second one is not a verdict that financial wellbeing fails to pay. It is a measurement problem. Financial habits change slowly, and the longer studies that would capture the return have not been run. This is a long game, and the tool will not credit a long-game return it cannot yet source.

That restraint is the reason to believe the numbers above it. An instrument that inflated the figures it could not prove would leave you no way to separate the solid numbers from the hopeful ones. This one draws the line in the open, which is why the $375,000 in retention and the $113,000 in coaching can be taken at face value.

Why the people tower carries half the load

Return to where Part 1 landed. AI compresses the skill distribution, and the least experienced gain the most. So the return on your technology is really a return on your people's ability to use it, and engaged teams realize that value faster than disengaged ones.

There is a deeper finding underneath. What separates the companies that win on people is not training spend. It is management practices, systems, and workflows.⁹ The dollars follow the architecture, not the intention. That is the whole subject of Part 3.

See it in your own numbers. The Living ROI Calculator at changeadvisor.net models the people tower from your inputs. For a 100-person company it puts retention near $375,000 a year, coaching at about $113,000 net, and a financial-wellbeing line it deliberately holds to a small, directional figure until the long-run evidence is in. Every figure traces to a graded, published source. Nothing is stored on the server.

The undermeasured number, measured

What your people return is the largest number most leaders never put on a slide. It is measurable, it is large, and at least one line of it is honest enough to run at a loss.

Part 3 is where the two towers meet: why AI investment and people investment only pay off when the architecture spans them, and what the distance between "realized" and "structurally supported" is actually costing you.

---

What Actually Returns is a three-part series.
Part 1 — The technology tower: what AI actually returns.
Part 2 — The people tower: the most undermeasured number in business. (you are here)
Part 3 — The span: the architecture that closes the gap.
Run your own numbers at changeadvisor.net. Not sure where your organization stands first? Start with the AI Readiness Assessment.

References

1. Gallup. "State of the Global Workplace 2026." April 2026. [T2 · 20% engaged, 2nd consecutive annual decline]
2. Gallup. "State of the Global Workplace 2026." April 2026. [T2 · ~$10T, ~9% of global GDP]
3. Gallup. Engagement–performance meta-analysis (Q12), 2024. [T2 · top vs bottom quartile: +18% sales productivity, +23% profitability. Correlational, disclosed.]
4. Theeboom, Beersma, and van Vianen. "Does Coaching Work?" Journal of Positive Psychology, 2014. [T1 · g=0.60 performance/skills]
5. Lacerenza, et al. "Leadership Training Design, Delivery, and Implementation." Journal of Applied Psychology, 2017. [T1 · δ=0.72 results, 335 samples]
6. Gallup. "State of the American Manager" (updated 2026). [T2 · 70% of team-engagement variance attributable to the manager]
7. McKinsey Global Institute. "Performance through People." February 2023. [T2 · People+Performance Winners: attrition ~5 pts lower; n=1,800]
8. Song, Zirui, and Katherine Baicker. "Effect of a Workplace Wellness Program on Employee Health and Economic Outcomes." JAMA, 2019. [T1 · no significant effect; supported by the Illinois/QJE 2019 RCT]
9. McKinsey Global Institute. "Performance through People." February 2023. [T2 · management practices/systems/workflows, not training spend alone, distinguish winners — the architecture-not-mindset finding]