B2B Staffing Platform for SMEs in India

Who does it and how:

An SMB with 20–200 employees has no dedicated recruiter. Hiring is done by the founder, a co-founder, or a single HR person who also handles payroll, compliance, and office admin. In companies below 50 people, nobody's full-time job is recruitment.

Blue collar and entry-level (security guards, delivery staff, factory workers, data entry operators):

• Staffing brokers and labour contractors operating under the Contract Labour Act. The broker maintains a roster; the company calls when it needs workers.
• WhatsApp groups managed by a supervisor or HR person. Someone screenshots a roster, forwards a number, and calls directly.
• The broker is paid by the candidate's first salary or by the employer at a flat monthly rate per headcount. The candidate often pays nothing upfront — subsidised by the candidate's desperation.
• Time to fill: 1–7 days for replacement workers; 2–4 weeks for new role types.

White collar, ₹3–15 LPA (sales reps, operations, accounts, admin, customer service):

• Naukri.com: companies post jobs, receive 50–500 resumes per posting, and spend 3–6 hours manually filtering. Naukri's algorithm surfaces profiles of active job-seekers who update regularly — meaning the same candidates appear across every competing posting.
• Indeed India: used but less dominant for Indian SMBs; more relevant for fresher and walk-in roles.
• LinkedIn: limited reach for SMB hiring in India outside metro tech roles; candidates with 3–10 years experience are present but rarely active job-seekers.
• Personal network and employee referrals: the founder calls known contacts, asks for references. Works once or twice, then hits the wall of network exhaustion.
• Staffing brokers and local recruitment consultants: typically charge 8–15% of the candidate's annual salary, payable on joining. Work well for mid-senior roles (₹10 LPA+) but are economically unattractive for SMBs filling roles at ₹3–6 LPA, where the fee of ₹25,000–72,000 feels disproportionate to the role's value.
• Reference checks: rarely done systematically. The employer asks two names from the candidate and calls them informally.

Where time and money leak:

• Founder time: a 50-person company with 6–8 annual hires loses an estimated 300–500 person-hours per year to recruitment — time that ships no product and serves no customer.
• Bad hire cost: a wrong hire at ₹3–6 LPA in India costs ₹2–5 lakhs in training, ramp-up salary, and eventual separation — and typically takes 6–18 months to correct.
• Staffing broker fee: paying 8–15% of annual salary on every hire, every year, instead of building an amortised cost structure.
• Sourcing budget waste: Naukri.com subscription at ₹10,000–25,000 per month for a small company, generating low-quality volume.
• Offer rejection: SMBs make offers that candidates reject because the process took too long or the employer couldn't move fast enough. Speed is a competitive advantage no SMB currently has.

---

Who profits from it staying manual:

• Traditional staffing brokers and recruitment consultants: their entire business model depends on the information asymmetry between employer and candidate. More transparency online erodes their moat. They are not threatened by SMB-specific tools because SMBs are not their primary clients.
• Job boards (Naukri.com, LinkedIn India): they profit from employer subscription spend and candidate resume purchase. They benefit from volume — more postings, more revenue — not from efficiency. A tool that reduces job postings per hire is against their interest.
• HRMS and payroll vendors bundling hiring modules: they sell the suite, not the hiring outcome. The hiring module exists to prevent the customer from leaving the ecosystem, not to solve hiring.
• Informal networks (WhatsApp job groups, local broker circles): the people who maintain these groups derive social capital or small referral fees. They have no incentive to build a better system.

Who is hurt by it staying manual:

• SMB founders: the single biggest losers. Every hour on hiring is an hour not spent on product, sales, or customer delivery. The compounding cost is invisible but real.
• Candidates in Tier 2 and Tier 3 cities: they never see SMB openings because those companies cannot afford broker fees and post inconsistently. The best candidates in smaller cities are invisible to SMBs in metros.
• Companies with 50–200 employees that have an HR person but no recruiting infrastructure: they are caught in the middle — too large to rely on founder-sourcing, too small to afford an RPO or enterprise ATS.
• Mid-level hiring managers who inherit bad hires because they had no structured screening process.

Who would pay to change it:

• SMB founders at companies with 20–150 employees, in non-tech roles, actively hiring 3+ roles per year — directly, if the cost is lower than a broker and the quality is equal.
• HR heads at 100–500-person companies who want to reduce their cost-per-hire from broker rates (₹20,000–80,000 per hire) to something lower.
• Founders who have experienced one bad hire and are motivated to improve screening on the next opening.

---

The narrow starting point: A managed shortlisting service for SMBs hiring non-tech white collar roles in the ₹3–15 LPA range. The agent (human or AI-assisted) handles: job brief capture, candidate sourcing, first-round phone/video screening with structured questions, reference check collection, and interview scheduling — presenting a shortlist of 3–5 candidates to the employer for a final decision.

Day one scope:

• The service takes a single open role from an SMB founder.
• It posts the role on Naukri (with the founder's account credentials) or sources directly via LinkedIn Recruiter Lite and WhatsApp outreach.
• It screens inbound applications with a structured phone conversation: role fit, salary expectations, notice period, two reference contacts.
• It presents a ranked shortlist of 3 candidates with a screening summary and reference call notes.
• The founder interviews and makes the final offer decision.

Pricing SHAPE: Per hire, on outcome. ₹5,000–12,000 per successful hire, due only when the candidate joins. This is:

• 60–80% cheaper than a traditional broker fee (which would be one month's salary for a ₹6 LPA role = ₹50,000).
• Zero risk for the employer — they pay only on result.
• Comparable to the economics of a budget Naukri posting + founder's own time cost.

The per-hire shape aligns incentives: the service is only profitable when it places candidates, forcing quality over volume.

---

Job boards — the dominant channel:

• Naukri.com: the clear leader for Indian white collar hiring at SMB scale. ~90 million resumes as of recent public filings. Revenue from recruiter subscriptions (₹10,000–₹50,000 per month for SMBs). Not a hiring agent — it is a listing and search platform. The employer does the screening.
• Indeed India: a job aggregator and search engine. Free to post. Generates high volume of applications, most from candidates mass-applying to every listing. Not targeted at SMBs specifically.
• LinkedIn: significant for tech roles and candidates with 5+ years experience in metros. Weak for SMB non-tech roles. Recruiter Lite starts at ₹12,000–₹20,000 per month.

Staffing agencies — the incumbent:

• Adecco India, ManpowerGroup India, Randstad India: all serve large enterprises primarily. Their SMB business is either non-existent or handled through very small branch offices in each city. Not interested in ₹5,000–12,000 per hire engagements.
• Local staffing brokers and placement consultants in every Indian city: unorganised sector. Charge 8–15% of annual salary. Work by personal network. No technology layer.

Technology-enabled recruiting platforms:

• Cutshort: professional network for tech roles in India. Uses AI matching. Targets product and engineering roles at startups and mid-size tech companies. Not SMB-general.
• Betterplace: blue collar workforce management platform. Provides payroll, compliance, and worker management for large enterprises and gig platforms. Not an SMB placement service.
• Belong: AI-driven recruitment platform for mid-senior level tech and product roles. Targets larger companies (500+ employees). Not SMB-accessible.
• greytHR: HRMS platform with a hiring module, popular among small Indian companies already using greytHR for payroll and compliance. Not a standalone hiring solution.

SMB-focused job portals:

• Jobdoga (unverified): claimed to focus on SMB hiring but operational details limited.
• Internfire (unverified): targeted at entry-level and internship hiring.

What is genuinely missing: A managed shortlisting and screening service for SMBs hiring non-tech roles at ₹3–15 LPA. The gap is not in listings — Naukri handles that. The gap is in the 10 hours of human work per hire that Naukri does not touch.

---

Kill condition 1: SMBs will not pay for this under any framing. How to check: post 30 hiring requests on LinkedIn and in 3 SMB owner WhatsApp groups (or cold-call 30 SMB founders in one city — Chennai, Pune, or Ahmedabad) asking if they would pay ₹5,000 on hire for a shortlisting service. Count responses. If fewer than 3 say yes, the market does not exist. Budget: ₹2,000 for LinkedIn outreach credits and phone call costs. Pass mark: 3+ willing to pay advance or escrow.

Kill condition 2: SMBs have a candidate shortage, not a screening problem. How to check: ask the same 30 SMBs what happens when they post a job. If most say "we barely get any applications," then better screening solves no real problem. The bottleneck is candidate supply, not candidate evaluation. In that case, the product must include candidate sourcing, not just screening. Budget: same outreach above. Pass mark: fewer than 5 of 30 saying "we get too many, can't filter them" kills this falsifier.

Kill condition 3: An AI agent cannot build enough trust to conduct first-round screening conversations with Indian SMB candidates. How to check: run a pilot where an AI agent (using a voice agent via Twilio or similar, or a structured WhatsApp conversation) screens 20 candidates for 5 real roles. Measure what percentage of shortlists the employer accepts without major revisions. If acceptance rate is below 50%, the agent is not reliable enough. If it is above 70%, the agent can operate at meaningful quality. Budget: ₹5,000–15,000 for API costs and setup. Pass mark: 3+ hires placed, with employer satisfaction that the shortlist was better than what they could have produced themselves.

---

Budget: ₹25,000–40,000

Week 1–2: Outreach and demand validation

• Cold outreach to 100 SMB founders in one city (Pune or Ahmedabad recommended — less saturated than Bangalore or Mumbai) via LinkedIn, phone calls, and WhatsApp. Target: manufacturing SMEs, B2B service companies, distribution businesses — companies with 30–150 employees hiring 2+ non-tech roles per year.
• Ask directly: "When you have an open sales rep or operations role, where do you post, who screens, and what would you pay for a guaranteed shortlist of 3 candidates in 5 days?"
• Cost: ₹3,000 (phone + data)

Week 2–4: Run 5 real placements as a pure service

• Take 5 real hiring mandates from the companies that express interest.
• Use a human-assisted process: post on Naukri (candidate's existing account), do phone screening yourself, present shortlists.
• Charge ₹5,000–8,000 per hire on placement.
• Target: 2+ paid placements to validate willingness to pay.
• Cost: ₹5,000 (Naukri posting credits) + ₹5,000 (phone/time)

Week 5–8: Introduce AI screening for inbound applications

• For 3 of the 5 mandates, use an AI agent (structured WhatsApp bot or voice agent) to conduct first-round screening of inbound candidates.
• Compare shortlist quality against your own manual shortlist.
• Measure: did the AI shortlist produce the same final hires? Did the employer notice a quality difference?
• Cost: ₹8,000 (agent API costs + setup)

Week 9–12: Evaluate and decide

• Count: placements completed, revenue earned, employer satisfaction, AI shortlist accuracy.
• Pass mark: at least 3 paid placements, at least 2 where the AI shortlist was used as the final shortlist, positive feedback from at least 2 employers on shortlist quality.
• Cost: ₹5,000 (ongoing outreach + minor fixes)

If pass marks are met: Scale by taking more mandates and building a WhatsApp-first AI agent to handle volume. Move toward the agencify model first, then identify what software pieces become sticky enough to productize.

If pass marks are not met: Kill the idea or pivot to a different wedge — possibly candidate sourcing for SMBs in Tier 2 cities (supply-side bet) rather than screening-side.

---

AGENCIFY first, then PRODUCTIZE.

A small team cannot productize a hiring tool for SMBs without first proving that SMBs will pay for the outcome the tool is supposed to produce — and the fastest proof is running the service manually, using AI tools as an internal efficiency layer, not a customer-facing product. The SMB buyer in India does not buy software from an unknown vendor; they buy results from someone they can hold accountable, which is exactly what an agency provides. The AI agent is not the product — it is the cost-reduction tool inside the agency that allows a small team to deliver broker-quality shortlists at half the broker price. Once 10–15 SMBs have paid for placements and the unit economics are validated, the playbooks and tools that make the agency efficient become the product.

Availability confirmed against the .in registry (RDAP) on 2026-09-24. Prices and ownership read from our own intelligence tables. Nothing here is estimated.

Single-word, available now

• flexible.co.in — available
• flexibles.co.in — available

Also available (compound)

• myflexible.in
• flexiblehub.in
• flexiblekart.in
• flexiblemandi.in
• flexiblebazaar.in
• flexibledirect.in
• flexiblesupply.in
• flexibleconnect.in

Taken and developed — do not chase

• flexible.in · entropy 4.67
• smes.in · entropy 4.67
• smeconnect.in · entropy 6.68

---

Generated 2026-09-24 08:41 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.