×
A larger mandate, delivered by the same team.
Hadaf reaches around six in ten Saudis who enter, seek or move between private-sector jobs each year. The remaining four in ten are not a demand problem — they are a capacity problem. Agentic AI lifts placement capacity without lifting headcount.
“Improve job-matching efficiency via AI powered HRDF platforms”
Six strategic objectives. All equally material.
Hadaf's strategy spans three themes and six objectives, reported to the board through eight KPIs. Agentic AI is relevant to all of them. For the purposes of this financial model, placement has been used as the core — it is the area with sufficient published data to quantify with confidence. The remaining objectives are equally important and will each expand the return.
Placement volumes, addressable market size and one-year retention are all published. The other objectives require Hadaf operating data not yet available to us.
562,000 supported employments a year across a nine-step process. A well-defined workflow can be modelled; a less structured one cannot be, with any honesty.
Contributions to employment is already measured and reported, so the modelled outcome expresses in a metric leadership already tracks.
Modelling one objective and excluding five understates the total return. Every additional stream quantified later adds to it.
This is a modelling choice, not a recommended sequence. Delivery order remains Hadaf's decision, and several objectives can be progressed in parallel.
Placement is the measurable core. Every other objective adds to the return.
The market replenishes itself every single year.
Placement is not a fixed pool that empties. Each year a new cohort enters, a standing population seeks work, and hundreds of thousands move between employers. Three flows are countable and non-overlapping; two are deliberately excluded.
Saudis entering the job market for the first time — school and university leavers.
Unemployed and returning job seekers. 95.9% are willing to accept private-sector roles.
Employed Saudis changing employers within the private sector during the year.
Attrition vacancies. The demand-side view of the same event as MOVE — counting both would double-count.
Freelance and flexible contracts. Overlaps the populations already counted.
2025 estimate. The 2030 projection applies 6% annual growth, in line with observed Saudi private-sector employment growth.
2025 estimate. Of the 562K, 222K are confirmed first-time entrants placed through Hadaf (Vision 2030 Annual Report 2025). The balance is support to people already employed or being re-placed — a different measure, and the reason active placement share sits below total reach.
The mandate is not constrained by demand. Nearly 390,000 people a year are in the market and out of reach — and the market grows to 1.27 million by 2030.
Capacity is capped by officer hours, not demand.
A placement is a sequence of nine steps. Four consume most of the officer's day — and those four are precisely the steps agents handle best. Volume scales with headcount because these steps scale with headcount.
Four steps consume most of the officer day. They are also the four agents handle best.
What the capacity uplift is built on.
The model rests on six inputs. Officer count and cost are the two that move the answer most — both require confirmation from Hadaf before the case is finalised.
Capacity uplift percentages are modelling assumptions to be calibrated against Hadaf cycle-time data, not measured results.
| Year 1 | Year 2 | Year 3 | |
|---|---|---|---|
| Officers involved in placementHadaf to validate | 1,200 | 1,200 | 1,200 |
| Fully-loaded officer costHadaf to validate | $44.8m | $44.8m | $44.8m |
| Baseline supported placements | 562,000 | 562,000 | 562,000 |
| Baseline cost per placement — labour only | $80 | $80 | $80 |
| Capacity upliftModelled | 7% | 14% | 21% |
| Incremental placements at flat headcount | 39,340 | 78,680 | 118,020 |
Officer count and loaded cost are the two inputs that move the answer most. Placement-facing share is modelled at 40% of total headcount; the model carries a 25–60% sensitivity range.
Every point of capacity released converts directly into contributions to employment.
Break-even in year one. Compounding thereafter.
Year 1 carries integration, evaluation and process redesign — capability is built rather than harvested. The return arrives in Years 2 and 3 as adoption matures. All costs are estimates until established through procurement.
| Year 1 | Year 2 | Year 3 | 3-yr total | |
|---|---|---|---|---|
| Value delivered | ||||
| Additional contributions to employment | 39,340 | 78,680 | 118,020 | 236,040 |
| Value of capacity released | $3.14m | $6.27m | $9.41m | $18.82m |
| Investment | ||||
| Agentic AI investment requiredEstimated | $3.10m | $2.67m | $2.45m | $8.22m |
| Return | ||||
| Net position | $0.04m | $3.60m | $6.96m | $10.60m |
| ROI — placement scope only | 1.01× | 2.35× | 3.84× | 2.29× |
| Extended ROI — placement plus other objectivesDirectional | — | — | — | 2.9× – 3.5× |
Agentic AI investment required covers four components: the agentic platform licence, an IT services partner for implementation and integration, LLM inference capacity, and hosting infrastructure on SCCC. Component-level costs are held in the supporting model and will be established through procurement. LLM capacity and hosting also serve Hadaf initiatives beyond placement; charging only the incremental share raises the three-year return to 2.53×.
Extended ROI is directional and has not been researched. Platform, hosting and governance are shared across all objectives, so extending scope adds to the numerator while the denominator grows far more slowly. Sizing it requires Hadaf operating data on training throughput and at-risk caseloads.
Sources, methodology and assumptions+−
Baseline volume. Hadaf's published 2025 supported employment of 562,000 (Arab News, February 2026). Held flat across all three years as the assumed capacity ceiling at today's officer count, so the uplift is isolated from organic growth.
Officer base. Total headcount of 3,000 is an input — Hadaf does not publish headcount and public aggregators conflict by 74%. The 40% placement-facing share reflects that employment, employer engagement and career guidance are the primary mandate. Sensitivity modelled at 25–60%; the return scales directly with this figure.
Officer cost. $37,300 (SAR 140,000) fully loaded is a market benchmark for comparable HR and recruitment roles in Saudi Arabia (SAR 95,000–140,000 base; median ~SAR 118,800 — approximately $25,000–37,000) plus approximately 20% for GOSI employer contribution, end-of-service and allowances. Not a Hadaf figure — government entities do not publish per-entity pay.
Addressable market. ENTER is published (222,000+, Vision 2030 Annual Report 2025) and corroborated by working-age population growth of ~400K per year at 49.5% participation. SEEK is derived from GASTAT Q4 2025 ratios and cross-checks to the published 7.2% unemployment rate. MOVE applies a 12.5% switch rate to the 2.6m private-sector base, set below the 19% implied by Hadaf's own 81% one-year retention.
Capacity uplift. 7 / 14 / 21% reflects adoption and learning-loop maturity, not model capability. Conservative (4/8/12%) and accelerated (10/20/32%) scenarios are modelled. To be calibrated against Hadaf cycle-time data.
Investment. All four components are estimates pending procurement. LLM capacity and hosting serve Hadaf initiatives beyond agentic placement; charging only the incremental share raises the return to 2.53× and leaves roughly $0.79m of reusable AI capacity available to other programmes at no additional cost. The fully-loaded view is shown as the conservative case.
Currency. All figures are presented in USD for consistency across components. Officer costs are incurred in SAR and converted at the pegged rate of SAR 3.75 to USD 1.00. A final version for Hadaf leadership will be presented in SAR with USD shown alongside.
How value is measured. The primary metric is additional contributions to employment — a board-reported KPI. The financial return is the officer cost Hadaf would otherwise have added to deliver the same volume: a cost avoided, not cash released. Additional placements continue to consume programme budget through wage subsidies and training; this is an operating-capacity model, not a fiscal one.
Scope. Placement has been used as the modelling core because it has sufficient published data to quantify. This reflects data availability, not a recommended delivery sequence or a view on relative priority. The five objectives outside the modelled scope are equally material and several can be progressed in parallel.
Deliberately excluded: REPLACE and FLEX flows from market sizing; Hadaf's own +23% growth rate as a forward projection; programme spend as a cost baseline; attribution of the national unemployment decline to Hadaf alone; and cost-efficiency as a headline metric, since no board KPI measures it.
The model covers one objective. Five more will expand it.
The return shown is the floor, not the total. Each objective outside the modelled scope adds a further capacity pool while the platform, hosting and governance layer are already accounted for. These have not been researched and carry no committed figures.
Capacity uplift at flat headcount, delivering contributions to employment, coverage of new labour-market entrants, and Jadarat placement conversion. Quantified in this case.
Training pathway design matched to employer demand, and at-risk occupation reskilling. Adds a second capacity pool without a proportionate cost increase.
End-to-end career progression support. Hadaf's roadmap activates this KPI in 2029.
Capacity released from sourcing, matching, coordination and compliance converts into additional placements at flat headcount.
Better matching quality at the point of hire. The current sustainability index is 81% — each retained placement is a re-placement avoided.
ENTER is the largest single group Hadaf places. Added capacity extends coverage of the annual cohort.
Agentic matching raises conversion from the Jadarat pool, already used by 58% of Saudi job seekers.
Matching can be targeted at youth and women. Female participation is 36% and female unemployment 10.3%.
Employer demand signals captured at intake identify in-demand roles earlier and route candidates accordingly.
One objective returns 2.3×. The mandate has six.
The market grows every year. The organisation does not have to.
A working model accompanies this summary. Every assumption is visible and editable — the next step is replacing the estimated inputs with Hadaf's own figures.
Confidential · Prepared for Hadaf leadership · 2026