Alkali

What do they actually do

Alkali builds ProcessMate, an AI assistant (beta) that turns engineering PDFs and P&IDs into a searchable digital model, runs automated design checks (e.g., line-size and material mismatches), recommends vendor parts to assemble a draft bill of materials, and helps generate RFQs to speed up quoting and procurement. You can try a public demo; the company reports digitization from drawings to a model in under 24 hours and positions the tool to shorten the quote phase rather than replace existing systems (homepage/demo, pilot write‑up, YC page).

The product is in early pilots with chemical engineering and project teams. Alkali highlights a pilot where the system helped select 75 pieces of equipment, with engineers reviewing and approving all final decisions; humans remain the authority for design and purchasing choices (pilot write‑up, HN beta discussion).

Who are their target customer(s)

• Process/project engineers at chemical plants: They spend hours extracting line sizes and checking for mismatches from scanned P&IDs/PDFs, which causes errors and slows handoff to procurement.
• Procurement or purchasing managers: They manually build BOMs and chase uniform vendor quotes, extending the quote phase and delaying orders.
• Fabricators and vendors quoting from customer drawings: They receive incomplete or ambiguous documents, leading to clarifications, slow quotes, and mismatched parts supplied.
• EPC/construction project managers: They discover schematic issues late, creating rework, schedule slips, and extra cost because problems aren’t flagged before procurement.
• Small chemical startups or operations teams scaling up: Limited engineering bandwidth makes long procurement cycles and design errors costly; they need faster, lower‑risk ways to get accurate parts lists and quotes.

How would they acquire their first 10, 50, and 100 customers

• First 10: Run tightly scoped pilots via direct outreach to engineers, procurement leads, and fabricators in existing networks; ingest one project’s P&IDs, deliver a searchable model, design checks, and a draft BOM to show time saved and fewer errors (demo/homepage, pilot write‑up).
• First 50: Expand through partnerships with fabricators, small EPCs, and vendors; host hands‑on workshops/webinars and publish repeatable case studies and RFQ/BOM templates to convert interest into pilots (pilot write‑up, YC page).
• First 100: Add channel and product integrations with major suppliers and procurement platforms so BOMs flow into quoting; standardize pilot→paid conversions with SLAs, supported by field sales and technical content showing improving reliability (APE‑0) (APE‑0 beta, demo/homepage).

What is the rough total addressable market

Top-down context:

Alkali sits across three spend pools: chemical EPC workflows, plant‑design/engineering software, and procurement software. Using 2024 estimates, a conservative sum of chemical EPC (~10% of ~$838B EPC), plant design software (~$4.2B), and procurement software (~$9.3B) puts TAM around $90–100B annually (EPC, plant design software, procurement software).

Bottom-up calculation:

Assume chemicals are ~10% of the $838B EPC market (~$83B), then add plant‑design software (~$4.2B) and procurement software (~$9.3B) → ~$96–97B of annual spend Alkali’s workflow could influence (EPC, plant design software, procurement software).

Assumptions:

• Chemicals represent ~10% of global EPC value (conservative, round share).
• TAM is the portion of EPC/design/procurement spend that a copilot accelerating P&ID→BOM→RFQ can materially influence.
• Estimates use public market reports; exact figures vary by source and definition.

Who are some of their notable competitors

• AVEVA: Full‑scale CAD/PLM suite used by EPCs to manage P&IDs, 3D plant models, clash checks, and material reports; powerful but not an AI‑first tool for ingesting scanned drawings and sourcing checks (docs).
• Autodesk (Plant 3D / Takeoff): Provides P&ID drafting, 3D plant modeling, and automated takeoffs/MTOs, mainly for workflows starting in CAD rather than messy scanned PDFs or procurement‑focused matching (Takeoff overview).
• Iomosaic (SmartPFID): AI tool to digitize legacy P&IDs, auto‑tag equipment, and embed data—overlaps on diagram digitization but not end‑to‑end BOM→RFQ procurement automation.
• iCaptur.ai: AI document/diagram processor for energy/industrial P&IDs and datasheets; strong on extraction and handoff to other systems rather than design checks and vendor matching.
• SymphonyAI (P&ID ingestion): Enterprise ingestion pipelines to parse P&IDs into structured data for asset management; part of a broader stack, not a procurement‑oriented copilot.