AI Workforce Skills: The Digital Foundation Is What Pays

Six million postings, and a premium that changes nature

The Skills That Pay 2026 report, authored by Indeed and the Asian Development Bank, analysed more than six million job postings across six Asia-Pacific markets between 2019 and 2024.

The method matters as much as the result. The researchers compared postings with the same job title at the same employer, holding other requirements constant, in order to isolate the pay difference attributable to digital skills.

The headline figure shifts the centre of gravity of the debate on AI workforce skills. Advanced capabilities, programming, artificial intelligence and machine learning, are worth premiums of around 26% and 30% in the country-level results published so far. AI use adds roughly 4%, AI development roughly 6%, as reported by HRD on 22 September 2026^[1].

The two sets of figures trace a hierarchy. The premium originates in the digital foundation, and AI rests on top of it as a thin layer.

The digital foundation carries the AI layer

Someone who codes, reads data and builds models brings a scarce, verifiable capability to the market, and the market pays double digits for it. Someone who runs a conversational assistant on top of a process that has stayed exactly the same brings an increment of a few points. The mechanism is straightforward: AI multiplies a capability that is already there, and it multiplies very little where the foundation is weak.

Many training plans launched in 2025 put AI literacy at the top of the list. The intermediate digital tier, data analysis, process logic, basic automation, was left behind.

The evidence from Indeed and the ADB points to the opposite sequence.

An organisation that buys licences and prompting courses for a workforce with a weak digital foundation is buying the 4% layer. One that rebuilds the foundation buys the 26% layer, and gets the 4% as well.

The entry threshold has risen for everyone

Almost nine postings out of ten now ask for some level of digital skill. Just 13% of the texts analysed ignore the digital dimension entirely.

That number describes a threshold rather than a management fashion. A profile with no digital foundation faces a market reduced to little more than a tenth of open positions.

• Basic tier: office tools, collaboration platforms
• Intermediate tier: data analysis, process automation
• Advanced tier: programming, AI and machine learning

The barrier operates on the external market and inside organisations in exactly the same way. A person with twenty years in the trade and a digital foundation frozen in 2015 watches their internal mobility shrink, at identical professional value. The issue is mobility before it is job security.

The cost of that rigidity stays invisible in the accounts, and highly visible in voluntary departures.

The fastest growth is happening outside tech

The most interesting signal in the report concerns traditional sectors: retail, customer service, education and healthcare are moving beyond basic digital literacy, towards data analysis and AI-powered tools.

For these professional populations the jump is a wide one.

A person at a till, behind a counter or on a ward learned digital tools as a support to the main activity. The new tier demands reading data and exercising critical judgement on output generated by a model, cognitive capabilities of an entirely different kind. Treating that transition as a software upgrade explains a great many adoption failures.

Healthcare and education organisations, moreover, have training structures designed for regulatory compliance rather than for developing technical skills. Here the training model has to be rebuilt before the content is.

The burden of closing the gap falls back on employers

Callam Pickering, Indeed's APAC senior economist, places the responsibility explicitly: digital skills have become a baseline requirement for career progression and for earning more.

The second point is harder for those in charge. Finding candidates who tick every box is difficult, and closing the distance between a good candidate and the ideal candidate falls to the employer, through on-the-job training and investment in human capital.

Silvia Garcia Mandico, ADB economist and co-author of the study, adds the distributional dimension. Broadening the opportunities to acquire and update digital skills allows people to adapt to technological change and to share in its benefits.

Two economists reach the same operational conclusion: converting internal talent pays better than hunting for the perfect profile on the external market.

The entry-level market confirms the hierarchy

The signal is coming from other markets too. In India, the trade press describes the shift from mass hiring to tighter selection, with AI-skilled engineers commanding premium pay^[2].

The convergence between an analysis of six million postings and reporting from Indian campuses points to a structural phenomenon rather than a passing cycle.

For those entering the market now the message is both blunt and useful. A qualification on its own opens few doors; a verifiable digital foundation opens many more. Organisations that redesign their graduate programmes around this hierarchy build a recruiting advantage set to last for years.

The reasonable objection

An analysis of job postings measures stated demand, while the wages actually paid remain outside the field of observation.

That is worth keeping in mind. Postings over-represent formalised roles and structured organisations, and the geographic perimeter remains Asia-Pacific.

Two elements remain solid nonetheless. The comparison is made at identical job title and identical employer, which eliminates most of the alternative explanations. The size of the sample, more than six million postings over five years, makes the result hard to attribute to statistical noise.

The correct objection concerns generalisation, never the robustness of the data.

What changes for decision-makers

For a chief executive, the conversation to take to the board is about the sequencing of skills investment, before its volume.

For the HR function, the priority shifts towards an honest mapping of the existing digital foundation. An AI course aimed at a population still at basic literacy produces the smaller premium the report describes.

For the finance function, the hierarchy between 26% and 4% finally offers a defensible allocation criterion. The documented return is concentrated in advanced skills, and those require long pathways, real teaching and time protected from operational urgency.

For the talent and remuneration committee, the metric to monitor comes down to one: the share of people moving from the digital foundation to the advanced tier each year, by function and by seniority band.

The design question

The question people inside the organisation are waiting for has a precise shape: what share of the workforce today holds the digital foundation on which AI produces the 26% premium, and what pathway takes everyone else to that threshold within twelve months?

A numerical answer to that question is worth more than any adoption plan.