The business case

What building it yourself costs, worked out with sources.

An Indian company's in-house people-analytics build against IntRpret, for the same 10 people asking the same questions: the team, role by role; the infrastructure, line by line; the AI, per question; and the total. Every figure has its source below.

The team

Who an in-house build needs, and what they cost

Annual pay (CTC) in India for each role, from public salary data. CTC already includes the employer's PF, gratuity and insurance.

  • Benchmark team2 peopleOne people-analytics practitioner per 2,500 employees — about two people for a 5,000-person company — with the data engineering borrowed from IT.
    • Head of people analyticsOwns the questions, the definitions and the roadmap₹25–50 lakh[1–2]
    • HR / people analystBuilds the analyses and answers the business₹6–8 lakh[3–4]
    Total CTC
    ₹31–58 lakh
    With employer costs
    ₹31–58 lakh[15]
    Per year
    $33k–61k
  • Full in-house build5 peopleEverything needed to build and run the analytical layer, the dashboards and the models without help.
    • Head of people analyticsOwns the questions, the definitions and the roadmap₹25–50 lakh[1–2]
    • HR / people analystBuilds the analyses and answers the business₹6–8 lakh[3–4]
    • Power BI developerBuilds and keeps the dashboards and reports₹7–11 lakh[5–6]
    • Data engineerConnects the HR systems and keeps the data layer₹11–18 lakh[7–8]
    • Machine learning engineerBuilds the forecasts and risk models₹13–25 lakh[9–10]
    Total CTC
    ₹62 lakh – ₹1.1 crore
    With employer costs
    ₹62 lakh – ₹1.1 crore[15]
    Per year
    $65k–118k
The infrastructure

What the analytical layer runs on

Each line is a published unit price times what an HR analytics build uses of it. The usage is our assumption, stated on the line.

On SnowflakeA warehouse that pauses when idle, and Power BI for the reports.
  • Snowflake compute$1k–12k / yr[11]$2–3 a credit (Standard–Enterprise, AWS Mumbai); X-Small = 1 credit an hour · assumed busy 2–8 hours a working day, 260 days a year
  • Snowflake storage$28–276 / yr[11]$23 a TB a month (AWS Mumbai) · assumed 0.1–1 TB of HR data
  • Power BI Pro$2k / yr[13]$14 a user a month · the same 10 people
Per year$3k–14k
On DatabricksA serverless SQL warehouse, and Power BI for the reports.
  • Databricks SQL Serverless$2k–6k / yr[12, 14]$0.77 a DBU, compute included (AWS Mumbai, Premium); 2X-Small = 4 DBU an hour · assumed 520–2,080 hours a year ($3.08 an hour)
  • Power BI Pro$2k / yr[13]$14 a user a month · the same 10 people
Per year$3k–8k
Like for like

The same 10 people, both ways

10 people asking 5–20 questions each working day — 12,500–50,000 a year — answered by AI on the same models. A year, before GST.

Per yearBuild it yourselfIntRpret
TeamHead of people analytics, HR / people analyst, Power BI developer, Data engineer, Machine learning engineer$65k–118k[1–10]₹62 lakh – ₹1.1 croreNo specialists to hire
Warehouse and BISnowflake or Databricks for the HR data, and Power BI for the 10 people$3k–14k[11–14]₹3–14 lakhIncluded — the CDM is the analytical layer
AI for the same questions12,500–50,000 questions at about 24¢ each: the models, the agent runtime and its memory$3k–12k[23–25]₹3–11 lakhIncluded
Total$71k–144k₹67 lakh – ₹1.4 crore₹15 lakh$16k, everything included

Dollars at ₹94.96 to the dollar[19]. The low ends go together — the cheaper team, stack and workload — and so do the high ends.

78–89%Saved every year with IntRpret — $55k–128k (₹52 lakh – ₹1.2 crore)For the same 10 people: an in-house build of $71k–144k a year against IntRpret's ₹15 lakh, and the in-house build takes 6–18 months[16–18] before its first answer. With the 2-person benchmark team instead, the build is $38k–87k a year and IntRpret still saves 59–82% — though that team leans on IT for the data engineering and the models.
  • The workload, on both sides: 10 people asking 5–20 questions each working day, 250 working days a year — 12,500–50,000 questions.
  • Salaries are annual CTC in India, from salary sites read on 19 October 2026; CTC already includes the employer's PF, gratuity and insurance [15].
  • Hiring the five through an agency adds a one-time ₹5.6–16.1 lakh, at 8.33–25% of first-year CTC by seniority [22].
  • AI is priced per question as IntRpret answers one: 23.6¢ — the models 8.3¢, the agent runtime 11.1¢, memory and the safety filter 4.2¢. Measured on 16 real IntRpret answers on 18–19 October 2026 (memory and the filter from October's bill), at AWS list prices [23–25]. A build answering the same questions on the same models pays at least that.
  • Infrastructure prices are published list prices; the usage on each line is our assumption for HR data alone. As a company-wide contract, a Snowflake warehouse is typically $20,000–100,000 a year and Databricks $31,630–333,333 [20, 21].
  • IntRpret is ₹15 lakh a year for the 10 people, everything included. All figures are before GST.
  • Not included on the in-house side: data-loading connectors, HRIS fees, hosting for the app the team builds, laptops and office space — each would add to it.

Sources

  1. [1]AmbitionBox — People Analytics Manager salary — average ₹24.5 lakh a year (68 salaries); used as the low end for the lead (15 Oct 2026)
  2. [2]PayScale India — Head of Analytics salary — median base ₹49.7 lakh a year; used as the high end for the lead (21 Mar 2025)
  3. [3]AmbitionBox — HR Analyst salary — ₹5.8–6.4 lakh a year at 3–6 years (about 4,000 salaries) (18 Oct 2026)
  4. [4]AmbitionBox — HR Data Analyst salary — ₹6.7–7.8 lakh a year at 3–6 years (15 Oct 2026)
  5. [5]AmbitionBox — Power BI Developer salary — ₹6.7–7.4 lakh a year at 3–6 years (31,100 salaries) (18 Oct 2026)
  6. [6]PayScale India — BI Developer with Power BI — average total pay ₹11.2 lakh a year at 5–9 years (8 Apr 2026)
  7. [7]AmbitionBox — Data Engineer salary — ₹11.3–12.5 lakh a year at 3–6 years (about 1.1 lakh salaries) (18 Oct 2026)
  8. [8]AmbitionBox — Senior Data Engineer salary — ₹16.4–18.2 lakh a year at 3–6 years (18 Oct 2026)
  9. [9]AmbitionBox — Machine Learning Engineer salary — ₹13.1–14.5 lakh a year at 3–6 years (9,800 salaries) (15 Oct 2026)
  10. [10]AmbitionBox — Senior Machine Learning Engineer salary — ₹22.6–25.0 lakh a year at 3–6 years (16 Oct 2026)
  11. [11]Snowflake — Service Consumption Table — credit prices by edition and region, storage per TB, credits per hour by warehouse size (effective 15 Oct 2026)
  12. [12]Databricks — pricing — SQL Serverless $0.77 a DBU in AWS Mumbai, Premium tier (19 Oct 2026)
  13. [13]Microsoft — Power BI pricing — Pro $14 and Premium Per User $24 a user a month, paid yearly (19 Oct 2026)
  14. [14]Microsoft Learn — Databricks SQL warehouse sizes — DBU an hour by warehouse size (2X-Small = 4); the Azure page, as the AWS one now redirects (11 Sep 2026)
  15. [15]Naukri — What is CTC — CTC includes employer PF, gratuity and insurance (29 Sep 2025)
  16. [16]Agile HR Analytics — pre-built vs custom HR data models — 6–18 months to production-ready analytics on a custom model (12 Mar 2026)
  17. [17]HRBench — People analytics — 12–18 months to actionable insights for a broad programme (26 Aug 2026)
  18. [18]Insight222 — People Analytics Trends 2025 — one people-analytics practitioner per 2,500 employees (372 companies) (2025)
  19. [19]European Central Bank — euro reference rates — ₹94.96 to the dollar (EUR/INR 106.864 ÷ EUR/USD 1.1253) (16 Oct 2026)
  20. [20]Vendr — Snowflake contracts — company-wide contracts: median $100,000 a year, low $20,000 (738 purchases) (Feb 2026)
  21. [21]Vendr — Databricks contracts — company-wide contracts: median $333,333 a year, low $31,630 (205 purchases) (Feb 2026)
  22. [22]WSNE Consulting — recruitment agency fees in India — 8.33–25% of first-year CTC, by seniority (an agency's own published rates) (20 Feb 2026)
  23. [23]AWS Price List — Amazon Bedrock models, Asia Pacific (Mumbai) — Claude Haiku 5.5 (global): $0.10 input, $0.50 output, $0.01 cache-read and $0.125–0.20 cache-write per million tokens (published 12 Oct 2026)
  24. [24]AWS Price List — Amazon Bedrock AgentCore, Asia Pacific (Mumbai) — Runtime $0.0895 a vCPU-hour and $0.00945 a GB-hour (published 16 Oct 2026)
  25. [25]AWS Price List — Amazon Bedrock, Asia Pacific (Mumbai) — DeepSeek V3.2 $0.74 input and $2.22 output, Kimi K2.5 $0.72 and $3.60, Nova Pro $0.94 and $3.76 per million tokens; Guardrails sensitive-information filter $0.10 per 1,000 text units (published 16 Oct 2026)

Your next people decision, with the why attached.

Request a demo. We'll email you a two-minute walkthrough today, then set up a time that suits you.

We'll use your email to set up your demo. Remove it any time. Privacy Policy