RPO for India's SMBs: Deep-Dive Research Note

Who does it: In an SMB with 20–200 employees, hiring is done by the founder, a co-founder, or a single HR person who also manages payroll, compliance, and office snacks. In companies below 50 people, nobody's full-time job is recruitment.

What they use and where it breaks:

• Naukri.com and Indeed India — companies post jobs, get 50–500 resumes per posting, and spend 3–6 hours manually filtering. Most resumes are irrelevant (the job description was written poorly, or the portal's matching is loose). The filter happens in the founder's inbox at 11pm.
• WhatsApp groups — hiring managers post "we need a React developer, 3 years exp, Mumbai" into 5–10 groups. Responses flood in unverified. No tracking. NoATS. Candidates drop out because nobody follows up. The manager forgets who they short-listed.
• Small staffing agencies and job brokers — these are neighbourhood operations, not brands. A broker in Pune charges ₹5,000–12,000 per hire for blue-collar or entry-level white-collar roles. For mid-level (2–5 years, ₹6–15 LPA), the broker markup goes to 8–10% of annual CTC. SMBs use brokers because they don't have time to source; they pay for speed and pre-filtering.
• LinkedIn — founders and senior staff cold-message candidates directly. Time-intensive. Unstructured. Works only for tech and senior roles.
• Excel and Google Sheets — candidate tracking is a spreadsheet with columns: Name, Role, Stage (Applied / Shortlisted / Interviewed / Offer / Joined / Rejected), Salary, Source. No automated follow-up. No reminder. Candidates go cold.

Where time and money leak:

• Founder time is the biggest leak nobody prices. A founder spending 10 hours/week on hiring at ₹5,000/hour opportunity cost = ₹50,000/week on hiring alone.
• Vacancy cost: a sales role vacant for 45 days at ₹15 lakh CTC means roughly ₹1.15 lakh in delayed revenue, unmeasured.
• Bad hire cost: in SMBs, a wrong hire at ₹6 LPA costs 1.5–3x annual salary in lost productivity and re-hiring. Unmeasured because the company doesn't track it.
• Recruiter markup leakage: SMBs pay staffing firms 8–10% for roles they could source themselves if they had a tool that screened first.

Who profits from it staying manual:

• Staffing brokers — their entire business model depends on SMBs being too busy to source and screen themselves. A tool that does the first 80% of a broker's work at 20% of the cost threatens this directly.
• Job boards (Naukri, Indeed) — their revenue model is sell postings and resume database access. If a bot handles the screening and candidate management, the portal still gets the posting fee but the SMB's dependency on the resume database diminishes.
• Recruitment agencies targeting mid-to-senior level — their margin depends on scarcity and relationship. A transparent AI sourcing layer erodes the scarcity story.

Who is hurt by it staying manual:

• Every SMB founder in India. The average SMB takes 45–60 days to fill an open position (industry estimates from staffing firm surveys, not cited here; treat as directional).
• Candidates in Tier 2 and Tier 3 cities who never see SMB openings because those companies never post anywhere they can be found.
• The Indian economy — if SMBs can't hire efficiently, they can't scale, and SMBs employ 110 million Indians.

Who would pay to change it:

• SMB founders and owners — directly, if the tool saves them more than it costs and doesn't require them to change behaviour significantly.
• Small HR teams (2–5 person HR department in a 100–500 person company) — as an internal productivity tool, they would buy it if it integrates with what they already use (WhatsApp, Google Sheets, Naukri).
• Not: company HR heads who have existing vendor relationships and budget approval cycles — the enterprise sale is too slow for a small team.

The trust problem: SMBs have been burned by staffing brokers who promise candidates and don't deliver. Any new vendor has to prove it works before it asks for money. The payment trigger in this market is evidence, not pitch decks.

The single narrow thing: A WhatsApp-native AI recruiting agent that handles end-to-end hiring for SMBs for one open role at a time.

What it does on Day One:

Receives a job description via WhatsApp voice note or text from the SMB owner.

Rewrites the JD into a clear, searchable format and posts it to 3–5 relevant channels (Naukri, WhatsApp groups, intern-specific boards).

Receives candidate responses and runs a screening conversation over WhatsApp: experience, notice period, salary expectation, location, key skills — a 7-question conversation.

Shortlists top 5 candidates with a one-paragraph summary of each, delivered to the owner on WhatsApp.

Sends automated status updates to all candidates who were screened out (something most SMBs never do — which is why candidates badmouth them on LinkedIn).

The owner makes the final call and schedules the interview themselves.

Who pays: The SMB owner (founder, HR person, director). Not the candidate. Not the employer in a corporate enterprise sense.

Pricing SHAPE: Per hire. Not per seat, not per month, not a subscription that accrues value whether or not a hire is made. The SMB pays when a candidate joins.

• Entry tier: ₹3,000–5,000 per hire for roles under 2 years experience (₹3–8 LPA). Covers non-tech, BPO, sales, operations, admin.
• Mid tier: ₹8,000–15,000 per hire for 2–5 year roles (₹6–15 LPA). Covers mid-level tech, marketing, finance, operations managers.
• No refund if candidate quits in 30 days (standard broker practice) — this is intentional. It signals confidence and matches SMB expectations from existing broker behaviour.

Why this SHAPE: Per hire aligns incentives. The SMB pays nothing if nobody joins. The tool earns nothing if it doesn't produce. This is the same risk-reward shape as the broker it replaces, which means the SMB's evaluation frame is already calibrated to it.

What the AI actually does vs. what it doesn't:

• Does: source, screen, summarise, update candidates.
• Doesn't (Day One): negotiate salary, conduct interviews, handle joining formalities, verify documents.
• The agent is not trying to replace the founder's judgment. It's trying to remove the 6 hours of screening and status-update drudgery per open role.

Job boards (listing + passive search):

• Naukri.com — dominant in India for mid-level white-collar hiring. Charges ₹1,000–₹25,000 per posting depending on visibility. SMBs use it but complain about resume quality and recruiter spam.
• Indeed India — smaller footprint than Naukri in India, primarily used by larger companies and multinationals.
• Foundit (formerly Monster India) — declining relevance, once a Naukri competitor.
• LinkedIn — used for senior and tech roles. Cost-effective if you have time; ineffective if you don't.

ATS and recruitment software (enterprise-tier):

• Greenhouse, Lever, Ashby — designed for 50+ person recruiting teams. Overkill and overpriced for a 30-person SMB.
• Zoho Recruit — exists but is primarily used as part of the Zoho ecosystem.
• Hire by greytHR — an ATS within the greytHR payroll/HRMS suite, popular among small Indian companies using greytHR for compliance.

AI recruiting tools (India-focused, verifiable):

• Cutshort — a professional network for tech roles in India. Uses AI for matching. Active as of 2024. Targets product and engineering roles. Not SMB-first; more for startups and mid-size tech companies.
• Belong — acquired by InfoEdge (Naukri's parent company) in 2019. Their enterprise AI recruiting platform is integrated into Naukri's ecosystem. Not SMB-accessible.
• Springworks (Trivial legal mention: Springworks TalentBird — not verified as still operating under that product name as of 2025. Springworks as a company exists and offers other HR tech products).
• Skillate — an AI resume screening platform used by large Indian enterprises (Infosys, TCS, HDFC). Not SMB-priced.

Staffing firms:

• TeamLease, Quess Corp, Innovsource — all listed or large enough to be verifiable. These handle volume hiring (BPO, retail, factory) for large clients. Not SMB-focused.
• Unverified: any startup claiming to be "AI-first recruitment for SMBs" in India without a named investor or published customer count — too many to list and most have pivoted or shut down.

The gap that exists: Nobody has built a WhatsApp-native, per-hire AI recruiting agent specifically for Indian SMBs (10–200 people, ₹3–20 LPA roles, Hindi-English bilingual capability, broker-price-point). The existing tools either require too much setup, too much training, too much money, or are desktop-only.

Kill Fact 1: Indian SMB candidates will not engage with a WhatsApp AI — they will only talk to a human.

How to check: Spend ₹0 and one weekend. Create a WhatsApp Business account. Send a 7-question screening quiz to 50 candidates who have applied to real SMB jobs on Naukri (cold outreach via Naukri's apply-email pattern). Track reply rate and completion rate. If fewer than 30% start the quiz and fewer than 50% of those complete it, this fact is true and the idea is severely compromised. This is the single most important cheap experiment.

Kill Fact 2: SMB founders will not pay per hire for a recruiting tool — they expect job postings to be free and brokers to be the fallback.

How to check: Spend ₹0. Interview 10 SMB founders in your target city (10–200 people, any industry, any open role). Ask them: "If a tool found and screened candidates for your open [role] and you paid ₹5,000 only when someone joined — would you use it? Why or why not?" Listen for whether they say they'd use it (stated) vs. whether they actually tried and paid for something similar before (behavioural). A majority saying they'd pay is weaker than a majority having already tried and abandoned something similar (meaning they have the habit, just not the tool).

Kill Fact 3: Naukri or Indeed adds WhatsApp-native screening or candidate management in the next 18 months, making a standalone tool redundant.

How to check: Spend ₹0. Set a Google Alert for "Naukri AI screening" and "Indeed India AI recruiting" and monitor for 30 days. Also check whether Naukri has published any product roadmap or announced any AI feature in the last 12 months (search: site:naukri.com AI, LinkedIn posts from Naukri product team). If they have announced WhatsApp integration or AI screening as a feature in their product, the window closes. If they haven't, the next check is whether they have the engineering velocity to build it quickly — Naukri's product cadence suggests slow iteration, so an 18-month window is plausible even if they announce it today.

Budget: ₹75,000

Month 1 — Build and Internal Test (₹30,000)

• ₹0 on software (use WhatsApp Business API via a wrapper like Twilio or a platform like Gupshup; Flask app on a ₹200/month VPS).
• ₹5,000 on a virtual phone number with WhatsApp Business capability.
• ₹15,000 on a contract developer for the screening logic and candidate summary generator (target: 40 hours, remote, Upwork or referral).
• ₹10,000 on legal/compliance review of automated messaging under TRAI and India's IT rules (particularly consent requirements for bulk WhatsApp messages).

Month 2 — Pilot with 5 SMBs (₹25,000)

• Recruit 5 SMBs personally — founder network, warm introductions. Criteria: 20–100 employees, at least one open role (any function, any level), founder willing to use WhatsApp for the process. Do not charge them for the pilot. Ask only for feedback and a testimonial if it works.
• Run the full process manually assisted by the AI for each role: post, screen, summarise, update candidates.
• Track: time spent by founder vs. time spent by tool (measure the AI's contribution vs. human assistance).
• Track: number of candidates screened per role, number shortlisted, number who attended interview, number who received offer, number who joined.

Month 3 — Convert or Kill (₹20,000)

• Offer pilot customers a paid conversion: ₹3,000–5,000 per hire (entry tier) for their next 3 hires.
• Minimum success bar: 3 of 5 pilot customers pay for at least 1 hire within 90 days of pilot start. This is the pass mark.
• If 3+ pay: you have proof of willingness-to-pay from real users. Raise a small round or go full-time.
• If fewer than 3 pay: run the falsification experiments above before deciding to kill, pivot, or wait.

What NOT to spend money on in 90 days:

• Branding, website, content marketing, SEO.
• Any enterprise sales effort.
• Hiring a sales person.
• Building an app (iOS/Android). WhatsApp is the distribution and the interface.

AGENCIFY first, then AI-FY, not the other way around.

The reason: the first deliverable is not software — it is a proof that an SMB will pay for a recruiting outcome delivered partly through WhatsApp. Running the service manually (with a human executing the screening and summary steps, guided by a script) in Month 2 answers the product question faster and cheaper than building the AI first. If the service works with humans, the AI layer replaces the human cost and scales margin. If the service doesn't work with humans, the AI won't fix it. Build the machine to do the work; don't build the work around the machine. The per-hire pricing SHAPE is correct because it matches what SMBs already pay brokers and aligns incentives from day one — so the first test is whether the broker-replacement story lands in a single conversation with a founder who has burned before.

Availability confirmed against the .in registry (RDAP) on 2026-09-22. Prices and ownership read from our own intelligence tables. Nothing here is estimated.

Single-word, available now

• proces.in — available
• proces.co.in — available
• automates.co.in — available

Already ours

• automates.in · developed (entropy 6.21)

Also available (compound)

• myautomates.in
• automateshub.in
• automatesmart.in
• automateskart.in
• automatesmandi.in
• automatesbazaar.in
• automatesdirect.in
• automatessupply.in
• automatesconnect.in

Listed for sale

• recruitments.com · price not listed on verifyhn · seller holds 72 domains

Taken and developed — do not chase

• recruitment.in · entropy 6.41
• myautomate.in · entropy 4.84
• goautomate.in · entropy 5.44
• myprocess.in · entropy 6.30

---

Generated 2026-09-22 08:40 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.