Which part of your hiring
is actually broken?
The most common mistake: implementing AI everywhere at once. The teams that get results start with one bottleneck. Click yours and jump straight to the right workflow.
✍️
Writing JDs takes too long and they still sound the same
You rewrite the same requirements slightly differently for every role. It takes 2 hours per JD and still misses good candidates.
Go to Workflow 01 →
📋
200 CVs came in. You can’t review them all properly.
High-volume roles mean you’re skimming CVs and likely missing strong candidates who don’t have perfect-format resumes.
Go to Workflow 02 →
🔍
Everyone you’re finding is already talking to 5 other companies
Active candidates on Naukri and LinkedIn have multiple offers. You need to reach people who aren’t looking yet.
Go to Workflow 03 →
💬
Good candidates are going cold because follow-up takes too long
You have 40 candidates at different stages. Manual follow-up falls behind. Candidates wait 2 weeks and accept another offer.
Go to Workflow 04 →
Implement one this week.
See results immediately.
Not a theoretical overview. Actual steps with specific tools. Start with the bottleneck you identified above.
Write Better Job Descriptions in 3 Minutes
Most JDs are a copy of last year’s JD with slightly different words. AI can help you write skills-first, bias-reduced JDs that attract better candidates — starting today, for free.
⌄
1
Brief Claude with what you actually need
Don’t ask Claude to “write a JD for a Marketing Manager.” That gives you a generic template. Instead: “Write a JD for a Marketing Manager at [COMPANY TYPE], [SIZE], based in [CITY]. Must-have skills: [3-4 skills]. Day-to-day: [what they’ll actually do]. Report to: [title]. The role is [IN-OFFICE/HYBRID/REMOTE]. We’re a [CULTURE DESCRIPTION] team.” Specificity is everything.
2
Use the JD review prompt to catch bias
Paste the draft JD into Claude and ask: “Review this JD for language that might discourage qualified candidates from applying — including gendered language, unnecessarily exclusive requirements, jargon that signals insider culture, or qualification inflation. List specific phrases to change and suggest alternatives.” Research shows inclusive JDs receive 42% more applications.
3
Create role-specific screening questions
Once your JD is ready: “Based on this JD, write 5 pre-screening questions that distinguish genuinely qualified candidates from those who just read the JD carefully. Questions should be practical, not theoretical. Include one question that reveals how they think, not just what they know.”
Claude free tier: ₹0. This workflow costs nothing and can be implemented in the next 30 minutes.
Screen 100 CVs Overnight with a Custom Screening Bot
Build a pre-screening chatbot trained on your exact role requirements. Candidates interact with it after applying. You get a structured summary of each candidate’s fit before you review a single CV.
⌄
1
Build your screening bot with CustomGPT
CustomGPT lets you build a chatbot trained on specific documents — your JD, role requirements, culture notes. Upload your JD and the screening questions from Workflow 01. The bot asks each candidate the questions, captures answers, and gives you a structured summary. No API setup. No code. Takes one afternoon to configure.
2
Share the screening link with applicants
Add the CustomGPT chatbot link to your application confirmation email: “As a next step, please take 10 minutes to answer a few role-specific questions. This helps us review your application more thoroughly.” Most serious candidates complete it. Drop-off is a signal too — candidates who don’t engage with a 10-minute screening usually don’t engage in the job either.
3
Review summaries, not raw CVs
CustomGPT exports conversation summaries. You review 100 structured summaries — each with the candidate’s answers to your specific questions — instead of 100 CVs in 100 different formats. You make better decisions faster.
📋 Real-world proof
MIT Entrepreneurship used CustomGPT to manage high-volume candidate interactions at scale. Read the MIT case study →
CustomGPT: 14-day free trial. Then $49/month. For high-volume hiring: the time saved on one role alone typically covers the monthly cost.
Source Passive Candidates Who Aren’t Applying Anywhere
The best candidates for senior and specialist roles are already employed and not checking Naukri. You need to find them before your competitors do.
⌄
1
Build a target company and profile list
Use Claude to define your ideal candidate profile: “For a [ROLE] at a [COMPANY TYPE] in India, which types of companies would have someone with exactly this background? List 10-15 specific company names and the job titles that would map to this experience.” This gives you a targeted sourcing list instead of a broad search.
2
Scrape LinkedIn and job boards with Octoparse
Octoparse is a no-code web scraping tool. Point it at LinkedIn search results, Naukri profiles, or industry-specific directories matching your profile. It extracts name, title, company, LinkedIn URL and contact information into a structured spreadsheet. No technical setup — visual point-and-click interface. Build your sourcing list in hours, not days.
3
Personalise outreach with Claude
For each shortlisted candidate: “Write a 60-word LinkedIn InMail to [NAME], currently [TITLE] at [COMPANY]. We have a [ROLE] opening at [YOUR COMPANY]. Be specific about why their background is relevant. Don’t use the phrase ’exciting opportunity.’ Sound like a real person, not a recruiter template.” Personalised messages get 4-5x better response rates than templates.
⚠️ LinkedIn scraping — know the rules
LinkedIn’s terms of service restrict automated data collection. Use Octoparse within LinkedIn’s permitted usage — public data, manual exports, and within rate limits. Do not scrape private profile information. For compliant large-scale sourcing, LinkedIn Recruiter (paid) is the appropriate tool. Octoparse is best used for publicly available data on job boards, directories and company websites.
Octoparse: free for basic scraping (10,000 rows/month). Paid plans from $89/month. Claude: free. LinkedIn outreach: free with your existing account.
Automate the 65% of Candidate Communication That Costs You Nothing to Personalise
65% of candidates report inconsistent communication during hiring. 82% lose trust in employers because of it. Most of this is preventable with a one-afternoon automation setup.
⌄
1
Map your communication touchpoints
List every moment a candidate should hear from you: Application received, CV under review, Moving to screening, Interview scheduled, Post-interview update, Offer / Rejection. Most teams currently automate only the application confirmation. All of these can be automated.
2
Build the automation in Make.com
Connect your candidate tracker (Google Sheets or ATS) to Make.com. When a candidate’s status changes in the sheet, Make.com triggers the relevant email or WhatsApp message automatically. For WhatsApp, connect AiSensy as the delivery layer. For email, connect Gmail directly. The messages are personalised by pulling the candidate name and role from the sheet.
3
Write the messages once with Claude
Use Claude to write each message in your company voice — not corporate HR speak. Prompt: “Write a [POST-INTERVIEW / REJECTION / OFFER] message to a candidate who applied for [ROLE]. Our company tone is [DESCRIBE]. The message should be warm, specific, under 100 words, and not sound like it came from an HR template.” Review and save. These messages run automatically from then on.
Make.com: free tier (1,000 ops/month — covers most hiring teams). Gmail: free. WhatsApp via AiSensy: ₹999/month if needed. Start with email only at ₹0.
Generate Role-Specific Interview Question Banks and Scorecards
Most interview panels walk in with the same 5 generic questions for every role. AI builds a custom, structured question bank and debrief template for any position in minutes — making every panel interview more consistent and defensible.
⌄
1
Generate a structured question bank
Paste your JD into Claude and prompt: “Generate a structured interview question bank for this role. Include: 5 technical/functional questions that test core skills (not Googled answers), 3 situational questions using the STAR method, 2 culture and working-style questions specific to [COMPANY TYPE], and 2 questions that reveal how the candidate thinks under ambiguity. Include what a strong answer looks like for each.” The question bank is role-specific, not recycled.
2
Create an interview scorecard
After generating questions, ask Claude: “Create a 1-page interview scorecard for this role. Columns: candidate name, date, interviewer. Rows: one row per question cluster with a 1-5 rating scale and a notes column. Add a final overall recommendation field with four options: Strong Yes / Yes / No / Strong No. Format as a simple table I can paste into Google Sheets or Docs.” Share with all panel members before the interview. Debrief becomes a structured data exercise, not a memory contest.
3
Use NotebookLM for panel preparation
Upload the candidate’s CV, the JD, and any work samples to NotebookLM. Before the interview, ask: “What are the 3 most important things to probe in this interview given the candidate’s background vs what this role requires?” Gets every panel member focused on the right gaps — not just reviewing their own area.
Claude free tier: ₹0. NotebookLM: free. This workflow costs nothing and takes 20 minutes per role to implement.
Generate Offer Letters, 30-60-90 Plans and Onboarding Checklists Automatically
The hiring win means nothing if onboarding is chaotic. Most Indian companies still send the same generic joining kit to every new hire. AI builds a role-specific onboarding package in under 30 minutes — and Make.com delivers it automatically when an offer is accepted.
⌄
1
Generate a role-specific offer letter
Prompt Claude: “Write a professional offer letter for [ROLE] at [COMPANY NAME], joining on [DATE]. CTC: ₹[AMOUNT] per annum (fixed: ₹X, variable: ₹Y). Reporting to [MANAGER TITLE]. Probation period: 3 months. Location: [CITY / WFH]. Include standard Indian employment clauses: confidentiality, IP assignment, notice period. Tone: professional but warm. Avoid legalese. Under 400 words.” Have your legal team review the first template. After that, it’s plug-and-play.
2
Build a 30-60-90 day plan for the role
Paste the JD and ask Claude: “Write a 30-60-90 day onboarding plan for a new [ROLE] joining [COMPANY TYPE]. First 30 days: learning — what they need to understand about the company, team, tools and processes. Days 31-60: contributing — their first independent deliverable. Days 61-90: owning — a project or metric they are now accountable for. Be specific, not generic. Include 3 measurable success indicators per phase.” New hires who have a structured 90-day plan have 58% better 12-month retention (Glassdoor data).
3
Automate delivery with Make.com
When a candidate marks “Offer Accepted” in your Google Sheets candidate tracker, Make.com can automatically: send the offer letter PDF via Gmail, trigger a WhatsApp welcome message via AiSensy, create a Google Doc from the 30-60-90 template and share it with the new hire and manager, and add a calendar invite for day-one orientation. Set it up once — every subsequent hire triggers the full sequence automatically.
Make.com free tier: 1,000 ops/month — covers most hiring volumes. Claude: free. AiSensy: ₹999/month for WhatsApp. Gmail: free. Total: ₹0–999/month.
What AI can do.
What it absolutely cannot.
The best AI-assisted hiring teams are those who know exactly where the boundary is. AI is a force multiplier for human judgment — not a replacement for it.
✓ AI does this well
⚡ Write, review and optimise job descriptions in minutes
⚡ Screen high-volume applications against specific criteria
⚡ Draft and send personalised follow-up communications at scale
⚡ Schedule interviews by matching panel availability
⚡ Summarise CVs and screening responses into structured formats
⚡ Identify passive candidates from public data sources
✗ AI cannot do this
❌ Assess whether a candidate will fit your specific team culture
❌ Read hesitation in a candidate’s voice when discussing a past employer
❌ Negotiate an offer with a senior candidate who has competing interests
❌ Build the long-term relationship that brings a candidate back when they’re ready
❌ Judge whether ambition and humility are balanced in the right proportion
❌ Make the final call. That always stays with the hiring manager.
💡 The right frame
AI handles the mechanical 60-70% — sourcing, screening, scheduling, follow-up. This frees recruiters to spend more time on the judgment 30-40% — candidate relationships, hiring manager alignment, offer strategy, onboarding experience. The teams winning in 2026 are not the ones with the most AI. They’re the ones where AI handles volume and humans handle judgment.
Every tool. Honest verdict.
Confirmed free tiers.
| Tool | Best for in recruiting | Free tier | Paid from | Link |
|---|---|---|---|---|
| Claude | JD writing, outreach drafts, interview question generation, offer letter templates, rejection messages, 30-60-90 plans. Privacy-safe for candidate data — check Settings → Privacy first. Best for long-form drafting. | Yes — generous | ₹1,650/mo | Visit → |
| ChatGPT | JD ideation, screening question generation, outreach message drafting. Free GPT-4o access on the web. Turn off data training in Settings → Data Controls before pasting candidate information. Works in Hindi for Indian-language JDs. | Yes — GPT-4o | ₹1,650/mo | Visit → |
| Google Gemini | Best option for teams already in Google Workspace. Drafts JDs inside Google Docs, responds to Sheets data, integrates with Gmail for follow-up drafting. Free Gemini in Gmail is enough for most recruiting communication tasks. | Yes — Workspace | Free | Visit → |
| NotebookLM | Upload your company handbook, role briefings and past interview notes. Ask questions across all documents. Useful for onboarding new recruiters on role history and context. Upload CVs + JD to prep panel interviews. | Free | Free | Visit → |
| Tool | Best for in recruiting | Free tier | Paid from | Link |
|---|---|---|---|---|
| CustomGPT | Build a pre-screening chatbot trained on your JD and requirements. Candidates self-screen. You get structured summaries. MIT Entrepreneurship uses this for high-volume candidate interactions. | 14-day trial | $49/mo | Try Demo → |
| Manatal | AI-powered ATS built for Asian markets including India. Candidate scoring, LinkedIn enrichment, pipeline management and team collaboration. INR-friendly pricing. Strongest affordable ATS option for Indian companies hiring 5-100 people per year. | 14-day trial | $15/mo | Visit → |
| iSmartRecruit | India-based ATS with local support, AI candidate matching and WhatsApp integration. Built for the Indian hiring context. Good option for recruitment agencies and mid-size corporates wanting India-hosted data. | Demo available | Custom INR | Visit → |
| Emergent | Build your own internal candidate tracker, ATS-lite or role briefing tool. Chat with AI agents that design, code and deploy the tool. No developer needed. YC S24. | Yes | Paid plans | Visit → |
| Tool | Best for in recruiting | Free tier | Paid from | Link |
|---|---|---|---|---|
| Octoparse | No-code web scraping for passive sourcing. Extract candidate profiles from public LinkedIn search results, Naukri, industry directories. Visual interface — no code. Use within platform terms. | 10,000 rows/mo | $89/mo | Visit → |
| Apollo.io | B2B contact database with 275M+ profiles including Indian professionals. Find verified work emails and phone numbers for passive candidates. AI-powered search filters by title, company size, funding stage and location. Free tier: 5 email exports/month. | 5 exports/mo | $49/mo | Visit → |
| Hunter.io | Find professional email addresses from any company domain. Useful for reaching candidates who aren’t active on LinkedIn. 25 free searches/month. Works well for finding decision-maker contacts at target companies for lateral hiring. | 25 searches/mo | $34/mo | Visit → |
| LinkedIn Recruiter | The compliant paid option for large-scale passive sourcing. AI-suggested matches, InMail credits, full search filters. Required for enterprise-scale sourcing workflows. | No | Custom pricing | Visit → |
| Tool | Best for in recruiting | Free tier | Paid from | Link |
|---|---|---|---|---|
| Make.com | Automate candidate status communications, interview scheduling reminders, post-interview follow-ups and offer workflows. Connects Google Sheets, Gmail, WhatsApp, and your ATS. | 1,000 ops/mo | ₹750/mo | Visit → |
| AiSensy | WhatsApp Business API for candidate communication — application confirmations, interview reminders, offer notifications. Built in India with INR pricing and local support. 25,000+ Indian businesses use it. Most cost-effective WhatsApp automation for HR teams. | Free trial | ₹999/mo | Visit → |
| Calendly | Eliminate the email back-and-forth for interview scheduling. Share a Calendly link — candidates pick a slot that works for both sides. Connects to Google Calendar and Zoom. Free plan covers individual recruiters. Paid plan for team scheduling and round-robin panel scheduling. | Yes — 1 event type | $10/mo | Visit → |
| Google Sheets | Your candidate pipeline tracker and the trigger source for all Make.com automations. Simple, free, works with every tool on this list. No ATS needed to start — a well-structured Sheet handles 90% of what paid ATS tools do for small hiring volumes. | Free | Free | Visit → |
8 prompts HR teams use daily.
Copy. Paste. Customise.
Replace text in [brackets]. Works in Claude or ChatGPT (turn off training data first in ChatGPT Settings → Data Controls before pasting candidate information).
JD Writing
Skills-first job description
Use when: writing any new JD from scratch
Write a job description for a [ROLE TITLE] at a [COMPANY TYPE, SIZE] company based in [CITY], India. Must-have skills: [3-4 specific skills]. Day-to-day responsibilities: [3-5 actual tasks]. Reports to: [TITLE]. Team size: [N]. Must avoid: [any specific deal-breakers]. Company culture: [2-line description]. Format: role summary (50 words), responsibilities (6 bullets), requirements (5 bullets — separate must-have from nice-to-have), what we offer (3 bullets). Avoid gendered language and unnecessary qualification inflation.
JD Writing
Review existing JD for bias
Use when: reviewing a JD before posting — your own or a hiring manager's draft
Review this job description for language that may discourage qualified candidates from applying. Check for: (1) Gendered language or masculine-coded words, (2) Qualification inflation — requirements that aren't truly necessary, (3) Experience requirements that exclude non-traditional backgrounds, (4) Cultural fit language that signals exclusivity, (5) Jargon that only insiders would understand. For each issue found, suggest specific alternative phrasing. [PASTE JD]
Screening
Generate screening questions
Use when: setting up a pre-screening process for any role
Based on this job description for a [ROLE], create 5 pre-screening questions that: (1) Distinguish genuinely qualified candidates from those who just read the JD, (2) Are practical and scenario-based, not theoretical, (3) Include at least one question that reveals how the person thinks, not just what they know, (4) Can be answered in under 200 words each. Avoid questions with obvious "right" answers. [PASTE JD]
Screening
Interview question set
Use when: preparing for a first or second round interview
Create a structured interview question set for a [ROLE] role. Include: (2) Role-specific technical questions with what a good answer looks like, (2) Behavioural questions using the STAR format, (2) Questions that assess culture fit without being exclusionary, (1) A question that reveals ambition and self-awareness, (1) A question for the candidate to ask us. For each question, note what you're looking to assess. Keep the full interview under 60 minutes.
Outreach
LinkedIn InMail to passive candidate
Use when: reaching out to a candidate who isn't actively looking
Write a LinkedIn InMail to [NAME], currently [TITLE] at [COMPANY]. We have a [ROLE] opening at [YOUR COMPANY]. Be specific about why their background is a fit — mention their [SPECIFIC EXPERIENCE OR SKILL]. Keep it under 75 words. Don't use the phrase "exciting opportunity," "hope this finds you well," or "I came across your profile." Sound like a real person who has actually read their profile. End with a soft, low-commitment CTA.
Outreach
Follow-up after no response
Use when: a candidate hasn't replied to first outreach after 1 week
Write a follow-up LinkedIn message to [NAME] who didn't respond to my first message about a [ROLE] opportunity at [YOUR COMPANY]. The follow-up should: be genuinely brief (under 40 words), add one new piece of information (a specific reason this role is relevant to them or a recent development at our company), not be apologetic or pushy, and make it easy for them to decline if they're not interested. Don't reference the previous message directly.
Communication
Rejection message with dignity
Use when: communicating a rejection — applies to any stage
Write a rejection message for a candidate named [NAME] who applied for [ROLE] at our company. They reached [STAGE — e.g. first interview, final round]. Reason for rejection: [BRIEF HONEST REASON — e.g. stronger candidate for current needs / not the right fit for this specific role]. The message should: be warm and specific (not a template), thank them genuinely without being hollow, not leave the door closed permanently if they were strong, and be under 100 words. Our company tone: [DESCRIBE].
Communication
Offer letter narrative paragraph
Use when: writing the personal section of an offer letter — not the legal terms
Write the opening paragraph of an offer letter for [CANDIDATE NAME] who is joining as [ROLE TITLE]. The paragraph should: express genuine excitement about them specifically (mention one thing that impressed us — [WHAT IMPRESSED YOU]), set a warm tone for what to expect in the first week, and be written as if from the hiring manager personally, not HR. Under 80 words. Not corporate. Sounds human.
AI in hiring + DPDPA.
What Indian HR teams need to know.
The compliance landscape for AI in hiring is evolving fast. Here is what is relevant for Indian HR teams right now.
🔒 Compliance checklist — AI in Indian hiring
India’s DPDPA 2023 + global precedent set in 2026
📌
Candidate consent is required. Under India’s Digital Personal Data Protection Act 2023, you need explicit consent before collecting and processing candidate personal data. Your application form must state clearly what data you collect and why.
📌
Disclose AI use in screening. If you use AI to score or rank candidates, disclose this to applicants. A simple line in the application process: “We use AI tools to assist in initial application review.” Globally, this is becoming mandatory — Canada ruled OpenAI violated privacy law in May 2026, setting precedent.
📌
Human review before rejection. Do not auto-reject candidates based solely on AI screening scores without human review of at least edge cases. In January 2026, a class action lawsuit was filed against Eightfold AI in the US for AI-based hiring discrimination. Indian employment law will catch up.
📌
ChatGPT free tier is not safe for candidate data. If you paste candidate CVs or personal details into ChatGPT free or Plus, that data may be used to train OpenAI’s models. Turn off “Improve the model for everyone” in Settings → Data Controls before pasting any candidate information. Claude is privacy-safe by default.
📌
Delete data when it’s no longer needed. DPDPA requires data minimisation — store candidate data only as long as necessary for the stated purpose. Set a deletion schedule for rejected candidate data (typically 6-12 months).
What Indian HR teams
actually ask.
Which AI tool should an HR team of 2 people start with?
Start with Claude for JD writing and candidate communications. No setup, no integration, no cost. Open Claude, paste the JD writing prompt from above, and have a better JD in 3 minutes. Once that’s your habit, add Make.com for communication automation. Add CustomGPT for screening only when you have consistent volume (50+ applications per role). Sequence matters — don’t implement everything at once.
Can AI screen CVs better than humans?
At volume, yes — for specific measurable criteria. AI is faster, more consistent, and doesn’t suffer from resume fatigue after 50 CVs. But AI screening is only as good as the criteria you give it. Vague criteria produce poor screening. The other limit: AI cannot detect the things that make a CV interesting despite the format — unconventional backgrounds, unusual career paths, potential that doesn’t fit a template. Always have a human review the borderline cases that AI flags as near-misses.
How do I handle candidate privacy when using AI tools?
Two rules. First: never paste identifiable candidate data (name, phone, email, PAN, Aadhaar) into ChatGPT free tier without turning off training data in Settings → Data Controls first. Use Claude instead — it doesn’t train on conversations by default. Second: disclose to candidates that you use AI in your screening process. A one-sentence note in your application form covers this. Under DPDPA 2023, transparency about how personal data is processed is a legal requirement.
Will AI in hiring reduce diversity or increase bias?
It can — if implemented carelessly. AI trained on historical hiring data will replicate historical biases. The mitigation: use AI for criteria-based screening with explicitly defined criteria (not “culture fit”), review rejected candidates periodically for patterns, use the JD bias review prompt before posting roles, and ensure humans make all final decisions. Used carefully, AI can actually reduce bias by applying criteria consistently — something humans struggle with across 200 CVs.
What does a complete AI recruiting setup cost for an Indian SME?
To automate JD writing, candidate communication and basic screening: Claude (free) + Make.com free tier (₹0) + CustomGPT if you have volume ($49/month ≈ ₹4,100). Total: ₹0 to ₹4,100/month. At a conservative recruiter billing rate of ₹500/hour, recovering 8 hours per month from automation pays for everything. For passive sourcing add Octoparse at $89/month (≈ ₹7,500). AiSensy for WhatsApp: ₹999/month. The full stack is under ₹13,000/month — less than one day of agency fees.