Digits meet rebar
On 14 September 2026, Riyadh-based construction technology startup BRKZ secured $31 million in new capital: a $13 million Series B equity round co-led by Wa’ed Ventures and 500 Global (with BECO Capital and ANB Capital), plus an $18 million growth debt commitment from Stride Ventures, according to Wamda. Founded in 2022 by Ibrahim Manna, BRKZ digitises building-materials procurement for contractors and earlier in 2026 received a strategic investment from SIC, the investment arm of the Saudi Industrial Development Fund.
Why this is an engineering and UX story, not only a funding story
Construction procurement in the Kingdom is a multi-party mess: suppliers, contractors, logistics, financing terms, and quality disputes. BRKZ’s pitch is an AI-powered matching and fulfillment engine plus financing solutions across that network. That means three product surfaces that must stay coherent:
- Search and match that understand Arabic material names, grades, and site constraints.
- Fulfillment tracking dispatchers will actually open (not a PDF emailed at midnight).
- Financing overlays that do not break trust when payment terms stretch.
Build notes for industrial AI products
Ground models in SKUs, not vibes. Generative assistants that invent lead times will get someone fined on a job site. Prefer retrieval over freeform inventiveness; show supplier source and last confirmed stock timestamp.
Design for dusty thumbs. Field users wear gloves, work in glare, and switch between Arabic and English part numbers. Large tap targets, offline-tolerant order drafts, and bilingual SKU search are table stakes.
Debt is a product feature. The $18M growth debt facility exists to fund working capital and flexible payment terms. Surface those terms as first-class UI — schedules, late-fee clarity, and dispute states — not as a back-office spreadsheet.
Vision 2030 alignment is distribution, not decoration. SIDF-linked capital signals industrial policy interest. Your roadmap should still obsess over on-time delivery and invoice accuracy; policy tailwinds do not forgive broken ops UX.
iFynx takeaway
BRKZ shows where Gulf deep-tech capital is going: unsexy vertical workflows with AI as routing intelligence. If you build for industry, ship reliability theatre — timestamps, sources, bilingual SKUs — before you ship chat.
Acceptance checklist for AI construction procurement UX
Digitising building-materials buying for Saudi contractors fails when RFQ chaos, Arabic ops, and delivery reality are treated as afterthoughts.
- RFQ states are explicit: draft, sent, quoting, awarded, PO issued, in transit, delivered, disputed
- Line items carry unit, grade, and delivery window constraints that suppliers cannot silently ignore
- Price and availability quotes expire with visible clocks; expired quotes cannot be awarded without refresh
- Arabic and English labels match what site engineers and procurement clerks actually say
- Audit trail links award decision to quotes compared — not a free-text “chose vendor” note
- Kill switch can pause new awards for a category (e.g. cement) during price shock without freezing deliveries already in transit
If a project manager cannot reconstruct why Vendor B won last Tuesday’s rebar lot, procurement UX is incomplete.
UX pitfalls on job sites and in back offices
- Mobile RFQ creation requires desktop-only fields; site engineers escalate via WhatsApp instead
- Supplier portals ignore RTL and show truncated Arabic SKU names
- Delivery ETA updates never reach the site WhatsApp group that actually waits for trucks
- Dispute flows bury photo evidence of damaged materials behind five taps
- AI “recommended vendor” lacks explainable factors, so PMs ignore it and trust personal networks
Acceptance should include a shadowed day on a live Riyadh project — not only a polished demo RFQ.
Ops playbook: four weeks to awardable digital RFQs
- Week 1: freeze a state machine and owners for RFQ→PO→delivery
- Week 2: ship expiry clocks + compare-quotes table with audit export
- Week 3: bilingual mobile RFQ create/edit for the top five material categories
- Week 4: dispute photo flow + category pause drill
- Success metric: % of awards with complete quote comparison audit ≥ target from baseline
Procurement AI earns trust when awards are boringly explainable.
Saudi construction product constraints
Contractors will ask about Arabic ops, VAT invoices, and who is liable when AI recommendations miss grade or timing.
- Budget offline-tolerant draft RFQs for weak site connectivity
- Clarify data shared with suppliers vs kept private to the contractor tenant
- Map delivery windows to real logistics corridors, not abstract “Riyadh” pins
- Keep human override for awards obvious — never bury it under an AI score
Digits meet rebar only when the site team adopts the tool under heat and dust.
Mapping the procurement journey before designing screens
Construction procurement involves more people than most B2B products. Before designing a single screen, map who does what on a typical order:
- Site engineer: raises the need, often from a phone on site, with quantities taken from drawings or a bill of quantities.
- Procurement officer: compares suppliers, negotiates terms and issues the purchase order.
- Finance: approves budgets, payment terms and any financing.
- Supplier: confirms availability, price validity and delivery slot.
- Driver and site receiver: deliver, inspect and confirm quantities and quality on arrival.
Each role needs a different view of the same order. A common failure is to build one dense desktop screen for procurement and expect site staff to use it on a phone in the sun.
Where AI helps and where it should stay out
AI can add real value in a procurement platform when it works from structured data:
- Normalising requests: turning “50 tons rebar 16mm” or the Arabic equivalent into a standard catalogue item with grade and specification.
- Extracting quantities from bills of quantities or uploaded spreadsheets, with every extracted line shown for confirmation.
- Supplier matching based on catalogue coverage, location, past delivery performance and confirmed stock.
- Exception alerts when a delivery is late, a price changes or a quantity does not match the order.
AI should not invent prices, lead times or stock levels. Every number a buyer acts on should come from a supplier confirmation or a system of record, with a visible timestamp.
Saudi compliance touchpoints
Procurement platforms in the Kingdom touch several regulatory areas. ZATCA’s e-invoicing (Fatoora) programme requires in-scope businesses to issue and, in the integration phase, share invoices with ZATCA in a structured format, rolled out in waves; a platform that invoices on behalf of suppliers or itself must handle this correctly. Personal data of site staff, drivers and contacts falls under the Personal Data Protection Law, enforceable since 14 September 2024. If the platform offers financing, the financing partner’s licensing and disclosure requirements shape the UX. See our ZATCA e-invoicing engineering checklist.
Delivery and proof of receipt
Disputes in construction often come down to what arrived, when and in what condition. A good receipt flow lets the site receiver confirm quantities line by line, take photos, note damage and sign, even with weak connectivity, and syncs when the signal returns. The supplier and buyer then see the same record, which shortens payment cycles and reduces arguments.
Financing inside the workflow
Embedded financing works when it is offered at the moment of need, for example at purchase order approval, with clear terms, and when repayment status is visible next to the orders it funded. Avoid burying financing terms behind a single “pay later” button; contractors need to see cost, tenor and what happens if a project payment is delayed.
Metrics that matter
- Time from site request to confirmed delivery slot.
- Share of orders delivered complete and on time.
- Number of disputes per hundred orders, and time to resolve them.
- Repeat ordering rate per contractor.
Arabic and bilingual catalogue design
Building materials are described differently by suppliers, engineers and drivers, and often in two languages. A good catalogue stores a canonical item with its specification and grade, plus Arabic and English names and the common local synonyms people actually type. Search should tolerate spelling variants, mixed Arabic-English queries and unit differences (tons, kilograms, pieces). Showing a photo and the key specification next to each search result reduces wrong orders more than any clever ranking algorithm.
Supplier onboarding
A marketplace is only as good as its supplier data. Make onboarding simple for suppliers who may not have modern systems: allow catalogue upload from spreadsheets, confirm prices and stock through a lightweight mobile view or messaging, and show suppliers the performance data buyers see, such as on-time delivery and dispute rates, so they have a reason to keep it accurate.
Frequently asked questions
What is construction procurement software?
Software that helps contractors request, compare, order, receive and pay for building materials and services in one workflow.
Can AI set material prices?
It should not. Prices should come from supplier quotes or catalogues, with AI used to match, compare and flag changes.
Does a procurement platform need to support ZATCA e-invoicing?
If it issues or processes tax invoices for in-scope businesses, yes. Plan for it in the architecture, not as a later add-on.
Should the field app work offline?
Yes, at least for drafting requests and confirming deliveries. Construction sites often have weak connectivity.
Related: Embedded SME lending in the Gulf · Saudi PDPL checklist. Building a B2B or field app for the Kingdom? See iFynx app development for Saudi Arabia and AI development for Saudi Arabia.
