AI Strategy5 November 20268 min read

The AI Productivity Dividend — Part 5 of 6

The talent pool we designed out of work

Five-day roles quietly exclude capable people at predictable points in their lives. Part 5 of The AI Productivity Dividend argues for reduced-time roles as access to scarce capability, sets six tests that separate access from precarity, and explains why the offer must never be a discount.

#Operating Model#Leadership#Work Design#Responsible AI
A single stream of warm light enters a brushed-metal manifold and divides into five channels of different widths, one fitted with a small amber valve.

After the second cloud cost analyst was appointed, the finance director of Hantam Service Cooperative, a fictional shared-services cooperative in the Northern Cape, did something nobody had asked her to do. She went back through ten years of exit records.

Most said "resigned: personal reasons". When she spoke to the managers involved, a pattern appeared. A senior accountant had left to care for a parent. An experienced network engineer had retired two years earlier than he wanted to, because the choice was five days or none. A systems administrator had left because a health condition made full weeks unsustainable, and she was, in her manager's words, "the best troubleshooter we ever had". A project manager had left to run a small business and would have stayed two days a week if anyone had asked. All of these people are illustrative.

None of them was offered a smaller role, because none existed. The cooperative had not decided to lose them. It had designed every job in a way that made losing them the default.

Where five-day design loses people

People do not leave work at random points. They leave, or reduce, at predictable moments in a working life.

Where five-day design loses peopleA WORKING LIFECaringresponsibilitiesHealth andvariable capacityApproachingretirementPortfolioworkStudy andretrainingOutcome-designedreduced-time role on fair termsOtherwise: an exit, recorded as personal
People reduce or leave work at predictable points in a working life. When every role requires five fixed days, each of these points becomes an exit regardless of capability. An outcome-designed reduced-time role, on fair terms, is a door back in rather than a lesser job.
  • Caring for children, parents, partners or others who need support.
  • Health and variable capacity: for some people with disabilities or long-term conditions, full weeks are harder to sustain than the work itself.
  • Approaching retirement, when experienced people want to keep contributing but not at full time.
  • Portfolio lives: founders, educators, community leaders and volunteers who want to do serious professional work alongside something else.
  • Study and retraining, including as roles change around AI.

For each of these groups, a role that requires five fixed days removes them from the pool regardless of their capability. The organisation then reports a skills shortage. Sometimes it is one. Often it is partly a design shortage.

Part 3 showed the mechanics of hiring for the outcome. This article is about why that matters beyond one role, and how to do it without creating a lesser class of work.

Access, not discount

There is a tempting and corrosive version of this argument: these people want fewer days, so they will accept less. Less pay per hour, fewer benefits, no progression, short contracts. That version treats carers, parents, older workers and people with disabilities as a discount labour pool, and it is wrong in principle and in practice.

The proposition is different. Some of the most capable people available to an organisation cannot or will not sell five days a week. Reduced-time roles, designed from outcomes and paid fairly, give the organisation access to that capability. The gain is the capability, not a lower rate. An experienced engineer at three days is not a cheaper engineer. She is an engineer the organisation would otherwise not have at all.

This is also where the productivity dividend in Part 1 becomes concrete. When AI and process redesign reduce the hours a set of outcomes needs, one destination for that gain is roles that fit more people's lives. That is a choice, and it competes with other uses of the same gain. It should be made openly.

Six tests that separate access from precarity

Reduced-time work can be good work or insecure work. The difference is in the terms, not the hours.

Access or precarity: six testsAccess to capabilityPrecarityPay rateSame hourly rateLower rate per hourSecurityPermanent work, permanent termsRolling fixed-termBenefitsPro rata, gaps closedMissing or lapsedProgressionOn outcomes ownedStalled by visibilityInfluenceWritten voice countsDecided in absenceBoundariesOff days are offUnpaid availability
Reduced-time work can be good work or insecure work, and the difference lies in the terms, not the hours. A role that passes all six tests gives the organisation access to capability it would otherwise not have. A role that fails two or three is a cheaper job with a better name.
  1. Pay rate. The same hourly rate as a comparable full-time role, moving on the same cycle. Where section 198C of the Labour Relations Act applies, it requires part-time employees to be treated, taking hours into account, on the whole not less favourably than a comparable full-time employee unless there is a justifiable reason.
  2. Security. Permanent work offered on permanent terms. A permanent need filled with rolling fixed-term contracts is precarity, whatever the hours.
  3. Benefits. Pro rata at least, and gaps closed deliberately. For example, family responsibility leave under section 27 of the Basic Conditions of Employment Act applies only to employees who have worked for the employer for longer than four months and who work at least four days a week. A three-day employee has no statutory entitlement to it, even where they have the same family responsibilities as full-time colleagues. The statutory leave is narrow in any case (a sick child, or a death in the family). An employer can close the gap by policy.
  4. Progression. Promotion judged on outcomes owned, not presence, as set out in Part 3 and, for technical roles, in Engineering roles are defined by outcomes, not tools.
  5. Influence. Reduced-time staff take part in decisions that affect their work, through the written and meeting practices in Part 4.
  6. Boundaries. No unpaid availability on non-working days. A "three-day role" with a phone that must be answered on the other four is not a three-day role.

A role that passes all six is access. A role that fails two or three of them is a cheaper job with a better name. The legal points here are general information, checked against the legislation on 5 November 2026, and not legal advice.

Accommodation is design, not a favour

The Employment Equity Act defines reasonable accommodation as any modification or adjustment to a job or to the working environment that will enable a person from a designated group to have access to or participate or advance in employment. People with disabilities are one of the designated groups. A working pattern designed around outcomes, with clear handovers and fewer meetings, is often the kind of adjustment that makes participation possible.

The practical point for employers is that well-designed reduced-time roles are useful well beyond any one group. When accommodation is designed into the role rather than negotiated case by case, the person who needs it does not have to explain themselves to get it, and colleagues who need it for other reasons benefit too. What any particular employer must do in a particular case is a question for its HR and legal advisers.

Do not ask, do not assume

Two habits undo good design.

Asking for reasons. An employer needs to know whether someone can deliver the outcomes in the pattern offered. It does not need to know why they want that pattern. Reasons often involve health or family, and health information is special personal information under section 26 of the Protection of Personal Information Act. Collecting it without need is a risk, and it makes people justify their lives to get a fair job.

Assuming who wants what. Offering reduced-time roles mainly to women with children, or nudging older employees towards fewer days, turns an access policy into a stereotype. The Employment Equity Act prohibits unfair discrimination on grounds including gender, age, family responsibility and disability. The fair default is to offer the pattern to everyone the role suits and let people choose. The wider duties that apply when assessing applicants in South Africa are outlined in Hire on evidence.

Who gets the benefit, honestly

There is a fair criticism of everything above. Flexible and reduced-time arrangements are easiest to offer in knowledge work, to people who already have bargaining power. A cloud cost analyst can work three days with good handovers. A security guard, a cleaner or a call-centre agent on a fixed roster often cannot, and their work may intensify to cover the flexibility of others. If the productivity dividend flows only to already-advantaged staff, it widens a gap rather than narrowing one. This is CloudNala's analysis rather than a measured finding, but any organisation running a pilot should check it in its own data.

Hantam did not solve this, and does not claim to. It did two things. It extended its family responsibility leave policy to all employees regardless of days worked, and it asked every manager of shift-based teams what a share of the AI gain could look like for their people: in some teams, earlier finishing on a rota; in others, investment in training through learning paths by role. The point is to ask the question for everyone, not only for the roles where the answer is easy.

What Hantam changed

  • Pattern by default. Every vacancy is now assessed for a reduced-time design, using the attendance requirement test from Part 3, before it is advertised. Some roles stay five days. The decision is recorded either way.
  • Alumni invited back. Former employees who left for the reasons above are told when an outcome-designed role opens. Two have applied.
  • Returners supported. People returning after a break get a short, role-specific refresher on how AI tools are used in their work, in the spirit of AI literacy is not a prompting workshop, rather than being expected to have kept up on their own.
  • Exit interviews ask one new question. Would a different working pattern have kept you? The answers go to the people committee each quarter.

Checklist

  • Review exits for patterns at predictable life points, without identifying individuals.
  • Treat reduced-time roles as access to capability, never as a discount.
  • Apply the six tests: pay rate, security, benefits, progression, influence and boundaries.
  • Close statutory gaps that disadvantage reduced-time staff, such as family responsibility leave, by policy.
  • Do not ask for reasons. Do not steer particular groups towards reduced hours.
  • Ask what a share of the gain looks like for roles that cannot easily go part-time.
  • Have HR and employment-law advisers review terms before roles are advertised.

Key points

  • Five-day design removes capable people at predictable life points: caring, health, approaching retirement, portfolio work and retraining.
  • The proposition is access to scarce capability through better job design, with fair pay, progression and influence. It is never a discount.
  • Six tests separate access from precarity: pay rate, security, benefits, progression, influence and boundaries.
  • Do not ask why someone wants reduced hours, and do not assume who does.
  • Check whether the dividend is flowing only to staff who already have the most flexibility, and ask the question for everyone.

Next: Part 6, When shorter work fails. All parts are on the series page.

How CloudNala can help

CloudNala helps organisations use AI-driven changes in work to widen access to scarce skills rather than narrow it. We review roles for reduced-time design, apply the six access tests to proposed terms with your HR and employment-law advisers, and help you design the queue, handover and meeting practices that make reduced-time roles sustainable for the people in them and for the colleagues around them.


Work with CloudNala

CloudNala helps organisations move from technology ambition to practical execution across cloud, AI, data, platform engineering and digital services.

Whether you are exploring AI, modernising your cloud environment, building a public-sector digital service, or turning an idea into a working MVP, we can help you shape the roadmap and deliver the next step.

Talk to us about AI and work design or write to us at consult@cloudnala.co.za