Oozh.AI
Pitch · October 2026
The Oozh pitch
Oozh uses AI to explore the web of cause and effect from real-world events and map their impact onto the Indian stock market. We built the first version in three months, part-time, and are now looking for the support to launch it and land our first institutional pilot.
What we do
Oozh uses AI to explore the web of cause and effect from real-world events and map their impact onto the Indian stock market.
For investors · B2C
See which Indian stocks a news event could move, which way, and why.
For institutions · B2B & B2G
Build and explore the cause-and-effect web, then steer it on your own terms.
How it works · for investors
Read the story. Swipe left twice.
Real screens from the Oozh app, on a CNBC story from August 2026. Illustrative only, not investment advice.
How it works · for institutions
Pick an event. Build the web. Steer it.
Selected · Iranian oil output and exports weaken · 64%
Go deeper
Branch
Add an effect
Change an assumption
Concept design on real engine output. The team-added effect is an example.
Market
Two kinds of buyers, one engine.
InvestorsPeople who buy Indian stocks · B2C₹299 a month
TAM13.5 croreregistered investors on NSENSE, Aug 2026 · 135 million
SAM10 lakh+already pay for market researchMoneycontrol Pro alone, FY25
SOM10,000subscribers, our 3-year goalTijori, a similar research product, has 15,000 paying
InstitutionsResearch, risk, strategy and policy teams · B2B & B2G₹4 lakh a year
TAM1.08 lakhorganisations with 100+ staff, plus research firmsCompanyData, RevenueBase, 2026
SAM2,350already pay for one Indian research platformTracxn alone, June 2026
SOM100organisations, our 3-year goalPrewave, which maps world events to business impact, serves 200+ enterprises
Every figure is sourced in appendix A1, with the detail in A2 and A3.
Progress
MVP built in 3 months, part-time.
- BuildApp and AI engine
Working end to end on real news.
- ValidateBrokerage feedback
Investors want near-term moves, not long reports.
- IterateIMM, our way in
Immediate Market Movement, a short-horizon read of market mood.
In progressSEBI-mandated Research Analyst certification and registration.
The brokerage feedback came from one informal conversation. It is not a partnership, an endorsement or a study.
Insight
Hidden impacts for investors. A steerable web for institutions.
For investors · B2C
Oozh builds a deep, multi-level web from each event and finds its immediate and hidden impacts on Indian stocks, at any level.
For institutions · B2B & B2G
Teams build and steer their own cause-and-effect chain, as deep as they need, and project it onto markets, supply chains or regulation.
Immediate-impact analysis is still in development.
Business model
We sell subscriptions.
₹299a month
Monthly or annual plan.
Proposed, to tune after launchTijori charges ₹330 to ₹500 a month.₹4 lakha year
Annual licence, starting with a pilot.
Proposed, to validate in pilotsTracxn earns about ₹3.6 lakh a year per account.₹3.6 crore10,000 subscribers × ₹299 a month
+ ₹4 crore100 organisations × ₹4 lakh a year
≈ ₹7.6 crore a yearYear-3 goal at full run rate
A target, not a forecast. Assumes every customer pays the stated price. Benchmarks in appendix A2 and A3.
Roadmap
Retail in December. First pilot in February.
- Nov 2026Get ready
Research Analyst certification, registration and incubation.
- Dec 2026Retail launch
The investor app goes live once registration allows. IMM built alongside.
- Jan 2027B2B discovery
Talk to research, risk and strategy teams. Recruit pilot partners.
- Feb 2027First B2B pilot
A real partner builds and steers its own cause-and-effect web.
- Apr 2027B2B launch
Institutional subscriptions go on sale.
In progressAccuracy validation. The AI models behind Oozh already know how past events played out, so back-tests would flatter us. We’re designing a forward test with market experts and academics, and will publish the method and the scores, possibly after the retail launch.
Target dates. December depends on certification and registration. B2B discovery runs in parallel. Accuracy plan in appendix A6.
Team
Khavin Shankar
Solo technical founder
- 5 years building software in regulated health-tech.
- Built the Oozh MVP, app and AI engine, in 3 months.
- Pursuing SEBI-mandated Research Analyst certification.
We built Oozh because we wanted it as investors ourselves, my father and I. India is the market we know best, so we validate here first.
The ask
We’re seeking mentorship, AI credits, industry connections, and early-stage support to launch Oozh and secure our first B2B pilot.
- admin@oozh.ai
- Web
- oozh.ai
Sources and detail for every claim follow in the appendix.
Take the deck with you
All eleven slides and the appendix, ready to forward or present offline.
Appendix
Sources and detail
Everything behind the claims in the deck: where each number comes from, how the markets were sized, one real analysis from start to finish, how the engine reasons, and how we will measure accuracy.
Appendix · A1
Where every number comes from
| Figure | What it is | Source | As of |
|---|---|---|---|
| 13.5 crore | Investors registered on NSE (unique PANs) | NSE Market Pulse, Sep 2026 | Aug 2026 |
| 10 lakh+ | Moneycontrol Pro paid subscribers | Network18 annual report | FY25 |
| 15,000 · ₹330 to ₹500 | Tijori paid subscribers, quoted by its founder, and its monthly price by plan | Entrackr · Zerodha | 2025 |
| 1.08 lakh | 105,314 Indian companies with 100+ staff, plus 2,403 research and advisory firms | CompanyData · RevenueBase | Oct 2026 |
| 2,350 · ₹3.6 lakh | Tracxn paying accounts, up 16% in a year, and revenue per account a year (₹21.1 crore a quarter) | Tracxn Q1 FY27 investor deck, NSE | Jun 2026 |
| 200+ | Prewave enterprise customers | prewave.ai | Oct 2026 |
| 6,500+ · ~$200M | AlphaSense customers and Dataminr’s annual recurring revenue, both cited in A3 | AlphaSense · Dataminr | 2025 |
| 42 · 82 · 27 | Effects traced, companies within reach and stock calls in one real analysis | Oozh database, final backup | Aug 2026 |
Appendix · A2 · Retail market
The numbers behind 10,000 subscribers
TAM13.5 crore
Everyone registered to invest on NSE in August 2026. This is the audience, not the number willing to pay. It grew by 61.6 lakh between April and August 2026 alone.
SAM10 lakh+
Moneycontrol Pro crossed 1 million paid subscribers in FY25, which shows Indian investors pay for market news and research. It is a broader product than Oozh, so we do not count its users as ours.
SOM10,000
Our 3-year goal, 1% of that paying base. The closest research product, Tijori, has 15,000 paid subscribers at ₹330 to ₹500 a month. Oozh is priced at ₹299.
Distribution matters.
Tijori’s founder expects its Zerodha partnership to grow its paid base. Industry connections are in our ask for the same reason.
Other paid research products.
Tickertape, Screener, Trendlyne, Value Research and MarketsMojo all sell paid plans. We could not verify their paid counts, so none are used here.
Appendix · A3 · Institutional market
The numbers behind 100 organisations
TAM1.08 lakh
105,314 Indian companies with 100 or more staff (53,222 with 100 to 199, 33,239 with 200 to 499, 18,853 with 500+), plus 2,403 research and advisory firms with 1 to 50 staff. Third-party databases; government bodies are not counted.
SAM2,350
Tracxn, a Bengaluru research platform, has 2,350 paying accounts, up 16% in a year. Half are investment firms and 45% are corporates, including strategy teams and consulting firms. India is its largest market by accounts.
SOM100
Our 3-year goal, about 4% of Tracxn’s paying base. The closest product idea, Prewave, maps world events to their impact on a company’s suppliers and serves 200+ enterprises.
Intelligence software sells.
AlphaSense has 6,500+ customers and over $500M in annual recurring revenue. Dataminr is approaching $200M, serving 100+ U.S. agencies and 20+ governments. Tracxn earns about ₹3.6 lakh a year per account.
How we will validate.
List 100 named Indian organisations in research, risk and strategy. Hold 20 to 30 discovery conversations, test paid-pilot terms, and replace these estimates with real conversion.
Appendix · A4 · Proof of work
One real story, start to finish
“U.S. Iran pressure campaign raises energy-market risk” · CNBC · analysed 15 August 2026
42
effects, three levels deep
5,147
listed Indian companies checked
82
within reach of the effects
27
with a call
One CNBC story about U.S. pressure on Iran set off 42 effects, three levels deep. Oozh checked which of 5,147 listed Indian companies they could reach and found 82. It weighed 250 links between effects and companies, kept the 119 that held up, and made calls on 27 stocks: 3 bullish, 7 bearish, 16 to watch and 1 neutral. Its closest historical parallel was the 1980s Tanker War, a 65% match.
Oil India ▲ Bullish+2 to +6% · 5 to 45 days · conviction 55
It sells crude, so a higher price lifts what it earns per barrel.
Indian Oil ▼ Bearish−9 to −4% · 1 to 60 days · conviction 65
It buys crude, so the same rise squeezes its refining margins.
Figures as recorded in the analysis run. Illustrative, not a forecast or a recommendation.
Appendix · A5 · Methodology
How Oozh reasons
1Watch
Reads trending financial news every day, merges duplicate coverage into one event, and scores each event for market relevance. Low-value stories are dropped.
2Understand
Builds context for each event that survives: a summary, what, why, when, where and who, and its sources.
3Trace
AI proposes effects level by level, each with a reason and a confidence score. Every effect is challenged before it’s kept.
4Recall
Compares the event with past events to find historical parallels, and says where each parallel stops holding.
5Match
Maps the surviving effects onto 5,147 listed Indian companies. Every company link is challenged, and weak ones are dropped.
6Call
For each company that remains: a stance, a conviction score out of 100, a projected price band and a time horizon.
7Reality check
Each call will be compared with what actually happened. We’ll publish the hits, the misses and the shortcomings, and use them to keep improving.
AI does the reading. Analysts do the steering.
For investors, in-house SEBI-registered Research Analysts will navigate and steer Oozh’s analysis. Registration is in progress, and the retail launch waits for it.
Appendix · A6 · Accuracy validation
Back-tests would flatter us. So we’re building a better test.
Why not back-tests
It’s like grading a student on a test they’ve already seen.
Oozh runs on frontier models such as GPT and Claude, trained on recent data. They have already read how past events ended, so a back-test would score their memory, not their reasoning.
What we’re working on
1Designing a better method
Working out how calls get recorded and scored, and against which simple baselines, with no hindsight.
2Next, a second opinion
Market experts and academics will review and verify the method before any test runs.
3Then, test and publish
We’ll run the accuracy tests and publish the results, hits and misses alike, verified through SEBI’s PaRRVA framework.
Until then, we make no accuracy claims.