The 42% Problem: Why Half of AI Initiatives Never Make It Past the ROI Wall

Forty-two percent of companies are stuck.

That is not a feeling. It is a number.

According to PwC’s latest Global CEO Survey (March 2026), 42% of organizations have yet to realize meaningful returns from their AI investments. At the same time, 43% report measurable gains in revenue or cost efficiency from AI.

Those numbers sit almost perfectly side by side.

Which means, today, success in AI is nearly a coin flip.

One group is turning AI into measurable business outcomes. The other is accumulating licenses, pilots, and dashboards while leadership keeps asking the same question:

Where is the return?

For CFOs, this is the only number that matters.

Because the problem is no longer AI access.

The problem is measurement.

Why AI Programs Get Stuck

Most stalled AI initiatives follow the same pattern. The technology is active. Employees are using it. Budgets are expanding.

But nobody can clearly prove value.

Three problems tend to appear at the same time.

1. Signal Sprawl Creates Visibility Without Insight

Every platform reports activity.

Microsoft shows Copilot adoption.
Google shows Gemini usage.
Slack shows engagement.
GitHub shows commits.
Salesforce shows CRM activity.

Each dashboard tells a story.

None tell the financial story.

Because these systems rarely connect in a way that produces a defensible, per-seat ROI measurement.

So every quarter, finance teams do what finance teams always do:

Export. Combine. Rebuild. Defend.

Usually in spreadsheets.

2. Companies Mistake AI Volume for AI Value

There is a growing tendency to celebrate AI scale.

More agents. More tools. More pilots.

But quantity is not performance.

PwC highlighted this directly: organizations often measure AI progress by deployment counts instead of business outcomes.

A company running 40 AI agents with no financial attribution is not ahead.

It is behind a company running four AI systems that produce documented returns.

Without measurement, expansion becomes expensive theater.

3. Nobody Retires AI Spend

Traditional software budgets had governance.

AI budgets often do not.

Once an AI tool enters the stack, it rarely leaves.

Why?

Because no one has enough evidence to kill it.

Finance teams cannot confidently compare:

Copilot vs. Gemini
Commercial tools vs. internal deployments
Seat cost vs. actual contribution

So organizations default to the safest decision:

Keep paying.

That is how a 1,500-person company ends up with seven AI subscriptions and no accountability.

What It Actually Takes to Break Through

Moving from the 42% cohort to the 43% cohort does not require more AI.

It requires better instrumentation.

Three capabilities matter.

1. A Unified Signal Layer

Every productivity system. Every AI assistant. Every source of workforce activity.

Collected.

Normalized.

Measured.

Not separate vendor dashboards.

One intelligence layer.

2. AI Impact Measurement

Adoption metrics are no longer enough.

Leadership needs to know:

What did AI actually contribute?
How much output changed?
What happened beyond baseline behavior?
What financial impact was created?

Contribution—not usage—becomes the KPI.

And the math needs to be visible.

3. A Retirement Workflow

The most overlooked capability in enterprise AI:

Knowing what to stop paying for.

The right system should answer:

Which subscriptions are underperforming?
Which seats should be reassigned?
Which tools should be retired?

Not with charts.

With recommendations.

Board-ready.

With dollar figures attached.

The Mid-Market Opportunity

Large enterprise platforms are unlikely to solve this problem fast enough for mid-market companies.

The buying cycles are too long.

The implementation timelines are too heavy.

The economics assume dedicated analytics teams and thousands of employees.

But mid-market organizations face the exact same challenge:

They need AI accountability now—not after a three-year transformation program.

That creates a significant gap in the market.

Introducing Levos

Levos was built for organizations that need a defensible answer to a simple question:

Is our AI investment creating measurable value?

Levos acts as a Human Capital Operating System designed for CFOs and CHROs who need:

Unified workforce intelligence
AI contribution measurement
Subscription optimization
Board-ready ROI visibility

Without rebuilding their entire HR or analytics stack.

Early Access for CFOs and Chief AI Officers

Levos is opening early access to a select group of mid-market CFOs and Chief AI Officers who want to move beyond AI activity and toward measurable outcomes.

If you are evaluating: