LeedAB

The AI operating layer for supply chain logistics. A brain that learns how your business operates and Artificial Beings that execute on the carrier portals, TMS dashboards, ERPs, and inboxes your team already uses.

The Problem

Supply chain and logistics teams spend their days on repetitive computer work that nobody has been able to automate. Checking carrier portals for shipment status, reconciling proof-of-delivery, walking the supplier map for tier-2 risk, screening orders for export controls, chasing supplier commits, and rebuilding the same spreadsheet every Monday morning. It is skilled work, but it is repetitive, and it does not scale with revenue.

Every existing tool assumes the underlying systems will integrate. They will not. Most of the surfaces an operator touches in a day (carrier portals, supplier portals, foundry and OSAT dashboards, customs systems, freight TMS) were never built to be integrated. The work lives in the gaps between them. That is exactly the shape of work that an on-prem AI operator can now do.

The Product

A brain that learns how your team operates and Artificial Beings (ABs) that run the work. The brain ingests Slack, email, sheets, supplier portals, app screens, and docs. It learns vendor names, document templates, formatting preferences, operational rules, and the edge cases that used to live in your dispatcher’s head. The ABs drive the apps your team already uses like ERPs, TMS, carrier portals, inboxes, and spreadsheets, the way a human operator does. No APIs, no connectors, no IT integration project.

Five agentic workflows ship out of the box:

• Allocation control tower. Reconciles supplier commits, carrier capacity, inbound lanes, and customer commitments into one live state. Flags drift the moment a commit slips.
• Multi-tier supplier risk monitor. Walks the supplier map down to tier 2 and tier 3, holds a memory of who-makes-what-where, and watches for natural-disaster, geopolitical, and single-source exposure as the world changes.
• Compliance screener. Screens every part, customer, and end-use against BIS Entity List, OFAC SDN, EAR ECCN classifications, Foreign Direct Product rule, and customs filings. Pre-flights orders before a human ever touches them.
• LTB & EOL command center. Detects upstream Last-Time-Buy and End-of-Life notices, computes coverage against forecast, and proposes NCNR commit sizes that balance carry cost against re-spin risk.
• Supplier follow-up agent. Chases commits, PCNs, lead times, missing documents, and proof-of-delivery across email, supplier portals, and chat. Drafts the next nudge in your voice.

Every action feeds back into the brain. ABs are reachable from wherever your team already works (Slack, Teams, WhatsApp, or Telegram) with one-tap approvals on anything that goes outbound. Runs entirely on customer hardware. Contracts, invoices, vendor rates, payroll data, and operational records never leave your building.

Traction

Open distribution is working. The CLI hit 2,000+ downloads in its first month with no paid acquisition. Our first paying customer was live with their first workflow on day one and four high-touch workflows across teams inside month one; the pilot converted to a standing commercial engagement before month-end. Customers bring their own model key, so AI spend is their line item, not ours, and COGS does not scale with customers because the product runs on their hardware. Pilot ACV runs $50K–$250K and production deployments land in the $300K–$2M+ range.

Why Now

• The 2021 auto chip crisis cost the industry $210B in unbuilt vehicles. The lessons stuck. Supply chain teams across manufacturing, electronics, and freight are now budgeting for resilience tooling that did not exist five years ago.
• AI agents have crossed the threshold for real desktop work. Computer-use models can now drive enterprise software end-to-end, the same surfaces that a dispatcher, buyer, or compliance officer uses every day, reliably enough to land production workflows.
• On-prem deployment has become a competitive moat, not a fallback. Sensitive contracts, supplier rates, payroll data, supplier contracts, and export-controlled records cannot leave the building. Customers are explicitly procuring against local-first tools.
• Incumbents (SAP IBP, Kinaxis, o9, E2open, Coupa) are strong systems of record but weak at the exception execution that eats operations teams’ weeks. That gap is the category a startup can credibly own.

Wedge

Allocation Control Tower. Local, deployable operating layer that reconciles supplier commits, carrier capacity, inbound lanes, and customer commitments into one live state. Ingests commit emails, supplier spreadsheets, PDF letters, and portal exports. Surfaces gaps before they hit the line. Lands as a wedge inside one product line or one constrained supplier lane, then expands to the rest of the operation.

Adjacent wedges as discovery validates them: on-prem compliance copilot (export control + customs + denied-party screening) and multi-tier supplier risk monitor (supplier-walked tier-2/3 graph with persistent memory and event watching).

Market and Buyer

• Beachhead #1. Mid-market freight and last-mile (10–200 trucks). Companies like Traheel (current customer): dispatch, POD reconciliation, deductions, and driver attendance all run on spreadsheets and phone calls. Operator pain is acute, sales cycle is 4–8 weeks, the VP of Operations signs.
• Beachhead #2. Procurement at fabless designers, IDMs, and OSATs. Marvell-, Broadcom-, Amkor-profile companies where suppliers commit through email and PDF and the line goes down when a tier-2 supplier moves. Higher ACV, longer security review, the VP of Supply Chain signs.
• Economic buyer: VP/SVP of Supply Chain or Operations. Reports into a COO or BU operations leader and controls the pilot budget directly. No central IT or procurement gauntlet for the first deployment.
• Adjacent capital signal: Altana $200M Series C (Jul 2024) for multi-tier supply graph; Tive $40M (Jan 2025) for shipment monitoring; Overhaul $105M (Aug 2025) for in-transit risk. Capital is flowing into the broader category. The local-first, computer-use operating layer that sits across all of it remains underbuilt and underfunded.

Why It Wins

• Local-first changes the security conversation. Procurement, legal, and IT move from “policy exception” to “controlled pilot.” 30-day discovery-to-pilot instead of 9-month enterprise sales.
• Computer use beats integrations. Carrier portals, supplier portals, foundry and OSAT dashboards, and most operational tools were never built to be integrated. LeedAB works them like a human would.
• Memory and monitoring beat one-time discovery. Tier-2/3 mapping, supplier follow-up, and allocation reconciliation each take one human pass; then the brain remembers and the ABs watch continuously. The work stops repeating.

Status

• Product validated in production with a last-mile logistics operator. Cycle time on proof-of-delivery reconciliation reduced from 3 days to 20 minutes. Pilot converted to a standing commercial engagement inside month one.
• In active discovery with manufacturing, electronics, and chip-side targets for the next three pilots. Each is scoped to a single procurement workflow with measurable hours-per-week saved.

Team

• Muiez Makkawi (CEO). Founding engineer at Edge (Series A), where he built risk management systems for commodity trading. Compilers at IBM before that.
• Mina Aziz (CTO). Ex-Microsoft. Masters in AI.