B2B Procurement Marketplace for Industrial Goods in India

In an Indian SME with 20–200 employees — say, a sheet metal fabrication unit in Manesar, a chemical trader in Ankleshwar, or a packaging unit in Bhiwandi — procurement of industrial goods and raw materials looks like this:

A purchase manager (often the owner or a one-person function) opens WhatsApp and sends the same message to four to six saved contacts: "Need 500 kg MS flat bar 50×6mm, grade IS 2062, delivery by Thursday, Rajkot." Suppliers respond with a rate per kg, minimum order quantity, delivery timeline, and payment terms — all in a mix of text, voice notes, and photos of rate cards. The manager screenshots the best quotes into a WhatsApp chat with the accounts person. Accounts approves. The purchase manager calls the supplier to confirm. A delivery timeline is agreed verbally.

The same manager then chases the supplier three days later on WhatsApp: "Sir, what is the status?" Then again. Then calls.

The full loop — one purchase order for one SKU — consumes 45 to 90 minutes of manager time and involves at minimum two phone calls and six to ten WhatsApp messages. For a factory running 40 active SKUs, that is 30 to 60 hours of pure procurement time per month, done by someone whose actual job is production or sales.

Where money leaks:

• Broker commissions: For commodity materials — steel, aluminium, polymers, chemicals — a broker network sits between buyers and large mills. Brokers take 1.5% to 4% of order value. A ₹10 lakh monthly steel order costs ₹15,000–₹40,000 in broker fees, invisible in the P&L as "logistics" or "handling."
• Price opacity: No buyer knows the landed cost versus the actual mill rate. Suppliers charge a processing fee on top of the LME (London Metal Exchange) rate. A buyer paying ₹72/kg for aluminium when the LME-adjusted mill rate translates to ₹68/kg is losing ₹4/kg without knowing it.
• Working capital lock-up: Suppliers — especially new ones — demand advance or cash on delivery. Repeat buyers get 15-day credit. A ₹5 lakh order at 50% advance means ₹2.5 lakh locked for two weeks.
• Quality disputes post-delivery: Material arrives, inspection happens on the shop floor, and non-conformance is discovered after payment is released. Resolution is a phone call, a threatening WhatsApp, and sometimes no resolution.
• Order drop: The verbal purchase confirmation does not become a purchase order. The supplier "forgets" the colour or gauge. The buyer gets the wrong material. Rework or reorder at cost.

---

Who profits from it staying manual:

• Commission brokers: The broker's entire business model depends on information asymmetry. They know which mill has excess stock, which factory is closing, and when a price will move. Their margin evaporates if a buyer can get real-time quotes from five suppliers without the broker in the middle. A steel broker in Mumbai's Cotton Green area told a researcher in 2023 that their "data advantage" — knowing which re-rolling mill in Raigarh had cheaper seconds — was their only moat.
• Established suppliers: If a buyer has been ordering from Supplier A for five years, Supplier A has no incentive to share their best price proactively. The buyer learns Supplier B is ₹2/kg cheaper only when Supplier B cold-calls the factory.
• Purchase managers: In family-run SMEs, the purchase function is often held by a relative or a senior employee. Their value is partly relational — they know who to call. A tool that makes procurement fully transparent also makes the purchase manager replaceable.

Who is hurt:

• SME owners and factory managers: They are overpaying for materials by an estimated 5%–12% versus market rate, according to industry estimates (IBEF, 2023 — unverified at granular level), simply because they lack price discovery infrastructure. On a ₹1 crore monthly material bill, that is ₹5–12 lakh per month.
• CFOs of growth-stage companies: Working capital tied up in advance payments to new suppliers, and bad debts from quality disputes, are invisible costs that appear in cash flow statements but are hard to attribute to procurement failure.
• Junior purchase executives: They spend 60% of their day on coordination (chasing quotes, chasing delivery) and 40% on actual negotiation. They would be more effective — and more loyal — if they had tools that made them look smart.

Who would pay to change it:

• An SME owner spending ₹5+ lakh per month on raw materials who has ever asked "am I paying the right price?" — this is a person who has felt the pain but had no way to act on it.
• A purchase head at a ₹20–200 crore company who cannot demonstrate procurement savings to the CFO because there is no data.
• A startup founder building a product business who wants procurement off their plate so they can focus on design and sales.

The right paying persona is the SME owner or factory manager — someone with direct P&L responsibility who has felt the pain of overpaying but has no one to call except their existing broker.

---

The narrow thing to start with:

A WhatsApp-based AI procurement agent for one industrial material category in one Indian state.

Day one functionality: The buyer sends one WhatsApp message — "Need 500 kg MS flat bar 50×6mm IS 2062, delivery Bhiwandi by Friday" — and receives, within 90 minutes, a ranked shortlist of three suppliers with their best price, delivery date, minimum order quantity, and payment terms, as a clean message on WhatsApp. The agent has already called or messaged suppliers to collect this information. The buyer selects one. The agent sends a purchase order summary. Done.

This is not a marketplace. No supplier listing page. No app to download. No onboarding flow.

Who pays and how:

• SHAPE: Per-order outcome fee

- ₹1,499–₹3,999 per fulfilled order, charged to the buyer, when the buyer places an order from the shortlist - Free to send the query; the fee fires only when an order is placed - Volume discount: ₹999 per order for buyers committing to 10+ orders per month

• Alternate SHAPE: Monthly subscription (for repeat buyers)

- ₹4,999/month for unlimited queries in one material category - Good for factories with predictable, recurring purchase needs

The fee is positioned not as a "software subscription" but as a "procurement service fee" — like paying a commission, but the AI agent is cheaper and faster than the broker it replaces.

Why this shape works: The buyer pays for a confirmed order, not for software. The buyer's frame of reference is the broker's commission (1.5%–4% of order value). At a ₹1 lakh order, ₹1,499 is 0.15% — a fraction of what a broker would charge. The buyer compares the AI agent not to software but to the broker. This is a favourable comparison.

---

Verified players (known to operate in India as of 2024–2025):

• IndiaMART InterMesh: The dominant B2B listings platform in India. Suppliers maintain product listings; buyers send enquiries. Not a procurement workflow tool — the transaction does not happen on the platform. The enquiry-to-order conversion happens outside, on phone or WhatsApp.
• Moglix: B2B platform focused on MRO (maintenance, repair, operations), industrial tools, safety equipment, and raw materials including steel and polymers. Operates warehouses in key industrial corridors. Targets manufacturing SMEs. Raised funding through multiple rounds. Processes purchase orders on its platform.
• OfBusiness: B2B procurement platform for raw materials (steel, polymer, chemicals) combined with embedded financing for SMEs. Order value and credit facility offered together. Operates in multiple Indian cities.
• Zetwerk: B2B manufacturing and procurement platform connecting buyers with manufacturers for custom and standard parts. Strong in precision machining, castings, forgings.
• Udaan: B2B trade platform covering electronics, apparel, food, and general industrial goods. Not specialised in heavy industrial materials. Operates pan-India.

What is missing (no credible player occupies this space):

• A WhatsApp-native procurement agent that calls and messages suppliers on the buyer's behalf, collects quotes, and returns a ranked shortlist within hours — for industrial goods, not FMCG.
• A per-order procurement service for SMEs in a single material category (steel, chemicals, polymers) that replaces the broker's intelligence function without requiring the buyer to change their workflow.

The gap is not "no B2B platform exists." The gap is "no platform has successfully replaced the broker's informational role in commodity industrial materials for the SME buyer who will not install an app or migrate their ERP."

---

Kill fact 1: Indian SME purchase managers will not switch from their broker relationship, even for a free trial.

The broker is not just a price-discovery mechanism. The broker is a relationship, a credit arranger, and a problem-solver who handles quality disputes and delivery failures. If a ₹5 lakh steel order arrives with the wrong grade, the broker fixes it. An AI agent that only shortlists quotes does not fix anything.

How to check cheaply: Spend two days in an industrial area (Bhiwandi, Manesar, Vatva, or Rajkot) and conduct 15 structured conversations with purchase managers at SMEs. Ask: "When you need to source a new material, what do you do first?" If more than 10 of 15 say "call my broker" or "call my regular supplier," the wedge is too thin to start with that buyer segment. Budget: ₹15,000 travel + time. Pass mark: fewer than 8 of 15 cite broker or regular supplier as default for new sourcing.

Kill fact 2: Suppliers will not share real prices with an AI agent and will quote worse rates than they give their regular buyers.

Suppliers in India give different prices to different buyers. A factory buying 500 kg per month gets a different rate than a buyer ordering 5 tonnes once. If suppliers learn the agent aggregates and shares quotes publicly, they will quote nominal rates. If they fear the agent is creating price competition, they will refuse to respond.

How to check cheaply: Before building anything, set up a test WhatsApp Business account. Send quote requests to 30 suppliers (found on IndiaMART or trade directories) for one specific material with specific specs. Ask for a formal written quote via WhatsApp. Track: how many respond, how many give a formal written rate vs. a vague "call me." If fewer than 15 of 30 respond, or if responses take more than 24 hours, the supplier-side activation cost is too high to make the per-order economics work. Budget: ₹5,000. Pass mark: 15+ formal responses within 24 hours for a repeatable query.

Kill fact 3: The working capital problem cannot be separated from the procurement problem, so buyers will not pay for quote aggregation alone.

If the real pain is not "I don't know the right price" but "I don't have credit to buy at the right price," then an AI agent that saves ₹2/kg on steel does not solve the buyer's actual problem. The buyer will take the cheapest material they can get credit for, not the cheapest material available.

How to check cheaply: In the 15 SME conversations above, ask: "What stopped you from buying at the best price you found?" If more than half cite credit, payment terms, or cash flow — not information — the wedge must include financing to be compelling. Budget: embedded in ₹15,000 above. Pass mark: fewer than 8 of 15 cite credit as the primary blocker.

---

What to test: Can a human-piloted AI procurement agent, using WhatsApp, phone, and a simple spreadsheet, acquire and retain paying buyers for steel flat bar procurement in Maharashtra for a total spend of ₹1,00,000?

The test setup (human-piloted, simulating the AI agent):

• One material category: MS (mild steel) flat bar, all standard sizes
• One geography: Maharashtra (targeting Bhiwandi, Thane, Navi Mumbai industrial areas)
• One human operator running the procurement workflow: sending WhatsApp messages to suppliers, collecting quotes, shortlisting, sending the shortlist to the buyer
• 20 target buyers: SME factories in sheet metal, fabrication, or light engineering, with 30–200 employees, currently spending ₹3–15 lakh per month on steel

Week 1–2 (₹20,000):

• Build supplier list: 25 MS flat bar distributors and mills from IndiaMART, trade directories, and local references. Verify they stock standard sizes.
• Draft and test the WhatsApp query message. Test with 5 buyers informally — no charge, just to see if they respond and what they ask.
• Set up a simple Google Sheet to track: buyer, query, quotes received, quotes shortlist, order placed or not, fee charged.

Week 3–6 (₹30,000):

• Formally onboard 20 buyers: approach via cold WhatsApp, LinkedIn, or a local industry association meet. Offer the first order free if they place an order.
• Run the agent workflow for each buyer: collect quote, shortlist three, send shortlist, follow up on order.
• Charge ₹1,999 per fulfilled order (discounted test price).

Week 7–12 (₹50,000):

• Analyse conversion: how many queries became orders. Identify why non-buyers did not buy (price, credit, preferred existing supplier).
• Iterate the script: better quote collection, faster turnaround, clearer shortlist format.
• Measure NPS from five happy buyers: ask for a referral. One referral from a happy customer indicates product-market resonance.

Pass mark:

• 8 or more paying orders at ₹1,999 each = ₹15,920 revenue against ₹1,00,000 spend
• Revenue is not the point. The point is:

- At least 3 buyers ask to expand to another material category (proves wedge is not a one-time use) - At least 2 buyers refer a colleague (proves word-of-mouth signal) - Quote collection time is under 2 hours per order (proves the workflow is automatable)

• If all three signals fire, the wedge is real and the next step is building the actual AI agent to replace the human operator.

Fail mark:

• Fewer than 5 orders placed despite 20+ queries → buyers not converting
• Zero referrals after 8 orders → no organic growth signal
• Quote collection taking 4+ hours per order → human operator is not a proxy for an AI; the workflow is not compressible

---

AGENCIFY, with a clear path to AI-FY.

The initial wedge should be a human-run procurement agent — not software, not a marketplace — because the buyer already trusts a broker and a phone call, not an app. A service that does the broker's job better, faster, and cheaper is the only thing that will make a purchase manager switch. A software dashboard will be logged into once and abandoned. A marketplace will face the same chicken-and-egg problem every marketplace faces. Only a service — first human-piloted, then AI-replaced — can overcome the trust barrier that makes Indian SME procurement sticky. The 90-day test is designed to validate the service wedge before investing a single line of product code, because the hardest problem here is not building the AI; it is proving the buyer will pay for the outcome.

Availability confirmed against the .in registry (RDAP) on 2026-09-23. Prices and ownership read from our own intelligence tables. Nothing here is estimated.

Single-word, available now

• cave.co.in — available
• caves.co.in — available
• procurements.in — available
• procurements.co.in — available

Also available (compound)

• mycave.in
• cavehub.in
• cavemart.in
• cavemandi.in
• cavebazaar.in
• cavedirect.in
• cavesupply.in
• caveconnect.in

Taken and developed — do not chase

• myprocurement.in · entropy 4.85
• industrialkart.in · entropy 5.43
• industrialdirect.in · entropy 5.28
• industrialconnect.in · entropy 4.59
• industrialsupply.in · entropy 4.93

---

Generated 2026-09-23 14:45 UTC. Topic from our research queue; no market-size figure appears here unless a source is named. The domain block above is read from our own intelligence tables and confirmed at the .in registry (RDAP); the model wrote the analysis, not the domain facts.