Building the intelligent workforce engine.
A strategic perspective on AI-enabled transformation for HRDF — the Human Resources Development Fund of Saudi Arabia.
HRDF has transformed Saudi Arabia's workforce landscape. The harder question is what comes next. The data to answer it already exists inside HRDF's systems.
HRDF built the right machine. Now the machine needs to think.
Saudi Arabia revised its unemployment target to 5% by 2030. That revision is not a celebration of what has been achieved — it is a harder problem. Volume and throughput got the Kingdom to 6.4%. The next 1.4 points require something different.
The machinery that placed 562,000 citizens and cut unemployment to 6.4% was built for throughput. It delivered. That is not the 5% problem.
See the data
Youth unemployment sits at 13.8% for males and 20.4% for females ages 15–24 — more than double the national headline. 40% of Saudi CEOs name skills gaps as their primary challenge. The 5% target is a matching problem — not a volume one.
See the data
The 5% target is not a scale problem. It is a connection problem.
Three stakeholders. One untapped asset. The connection is what's missing — not the data.
Three stakeholders. One untapped asset. The same shift — unlocked.
Citizens, companies, and the country each experience the same structural gap: HRDF holds the data to serve them better — and it isn't connected to the decision yet.
From navigating eight programmes independently — to a platform that navigates for them
Two million citizens navigate eight disconnected programmes. No thread connects their skills to the right placement — before the mismatch becomes a dropout.
Read the full shift
From reacting to Nitaqat announcements — to receiving intelligence 14 months in advance
A significant share of establishments enter Nitaqat non-compliance every cycle — detected after the window closes. The signal that would flag it earlier is in HRDF's own systems — a lead time to validate in the POC.
Read the full shift
From quarterly outcome reports — to a forward model of what the labour market will need
HRDF reports what happened. The National Observatory mandate requires modelling what will. That shift cannot happen on annual reporting cycles.
Read the full shift
What the shift looks like — for each stakeholder
Citizens
Companies
Country
The data already flows through HRDF's systems. Lyzr puts it to work.
The connection HRDF needs. Two engines. One platform. Already in the data.
Lyzr doesn't bring new data to HRDF. It connects what HRDF already holds to the decisions that data should be driving.
Make every operation smarter. Before problems surface, not after.
Nitaqat risk flagged well before the window closes. Training spend aligned to this quarter's market demand — not last year's. The data for both already flows through HRDF's systems. Lyzr puts it to work.
See what this closes
Turn HRDF's data into a national intelligence capability no one else can build.
HRDF holds the only complete, live picture of how Saudi Arabia's private-sector workforce is built, trained, placed, and retained. No other institution in the Kingdom has this. Activated, it becomes a genuine workforce intelligence capability — sector demand modelled forward, employer quality scored, policy modelled before activation.
See what this closes
One engine sharpens what HRDF already does. The other builds something that didn't exist before.
Every interaction makes the next decision sharper.The shifts are clear. The architecture is ready. This is running in 90 days.
From first production agent in 90 days — to national intelligence platform in 12 months.
Each stage closes a measurable gap. Each creates the data foundation the next one builds on.
Stage 01 — deployed, not demoed.
The Grant and Disbursement Intake Agent is the recommended POC anchor. Measurable, scoped, deployable in 90 days. Fully inside HRDF's perimeter.
Grant and Disbursement Intake Agent
An establishment submits a wage subsidy or training support claim. The agent validates completeness against HRDF's eligibility criteria, cross-references employer Nitaqat status, prior claim history, and GOSI registration — with no manual lookup and no system switching. Exceptions route to a human reviewer with a structured brief: what the issue is, what context is relevant, what decision options exist. Every case closes with a structured record that feeds Stage 02's compliance intelligence layer. Stage 01 does not just save time. It creates the data foundation that closes Gap 01.
Built inside HRDF's perimeter. On HRDF's infrastructure. With HRDF's models.
Lyzr sits inside what HRDF already has. The platform provides the orchestration, governance, and multi-model control plane — deployed within HRDF's own infrastructure. Nothing leaves the perimeter.
On-prem, inside your perimeter
Kubernetes deployment within HRDF's own infrastructure. Works with Alibaba/SCCC, Grok, self-hosted models, or any provider. No cloud dependency. No data egress. Azure's Saudi Arabia East region extends this when available (Q4 2026) — no architecture changes required.
Built to SDAIA standards
Governance isn't a layer added later. The SDAIA AI Adoption Framework, NCA ECC/CCC controls, and PDPL requirements are built into how the platform operates — audit trails, entitlement policies, and human oversight are defaults, not configurations.
Your agents. Your IP. Your platform.
Every agent HRDF's team builds belongs to HRDF. Works with any model HRDF chooses to run — no lock-in to any single provider. Designed from day one to hand over. The goal from day one is for HRDF to run this without us.
The organisations that trust Lyzr look like HRDF.
Deploying AI inside a government perimeter — where data cannot leave, compliance is non-negotiable, and the stakes of a failed rollout are institutional — is a specific test. These are the organisations that have run it.
Lyzr's largest customer by deployment scale. Data-private agents running on local models, zero external LLM data sharing — the identical deployment model HRDF requires.
Lyzr serves as the agentic control plane — the orchestration layer connecting the institution's data to its operational decisions.
A regulated enterprise in workforce analytics and advisory. In production on Lyzr's platform.
$8M Series A (October 2025, Accenture Ventures) and $14.5M Series A+ led by Accenture (March 2026, valuation to $250M). The world's largest professional services firm has invested in Lyzr twice in five months. Accenture does not make back-to-back investments in infrastructure companies unless it is deploying them at scale with enterprise clients.
Trust at this level isn't claimed. It's earned in environments where the constraints are identical — and the consequences of getting it wrong are real.
Four steps to first production agent. The only commitment this asks for is the conversation.
One conversation. One scoping document. That is all this asks for right now.
HRDF holds the data behind Saudi Arabia's workforce.
Lyzr makes it compound.