7 Investors, 1 Stock: How My AI Committee Analyzed NSE: LATENTVIEW
Buffett and Graham said no. Lynch, Fisher, and Jhunjhunwala said buy. Taleb called it fragile. The weighted math came out at +0.15 — a split committee. The system bought anyway. Here’s every step.
In my last post I introduced the investor skill files — individual Claude Code skills that embody the complete investment frameworks of legendary investors like Warren Buffett, Ben Graham, and Rakesh Jhunjhunwala. Each skill file reads publicly available data, scores the business on the investor’s specific criteria, and returns a structured verdict: Bullish, Neutral, or Bearish, with a 0–100 confidence score.
This post is a live run. I assembled a seven-member Investment Committee, pointed it at NSE: LATENTVIEW — India’s first listed pure-play analytics firm — and let the system work. What follows is a complete account of every vote, the weighted math behind the consensus, the Portfolio Manager’s position decision, and the Risk Manager’s hard cap that almost overrode the buy signal.
No API keys. No subscriptions. Everything runs on Claude using free, publicly available data sources.
The Stock: NSE LATENTVIEW at a Glance
LatentView Analytics (NSE: LATENTVIEW) is India’s first listed pure-play analytics company — a data and analytics services firm that counts Fortune 500 technology, financial services, and consumer companies among its clients. At the time of this analysis (April 8, 2026), the stock was trading at ₹301, roughly 42% below its 52-week high of ₹518.
| Metric | Value | Context |
|---|---|---|
| Revenue (TTM) | ₹1,004 Cr | +27% YoY · 8 consecutive quarters of growth |
| EBITDA Margin | 22–23% | Stable across last 4 quarters |
| Net Profit (TTM) | ₹198 Cr | +20% YoY · but ₹81Cr from 'other income' |
| ROE (3yr avg) | 12.4% | Below Buffett's 15% bar |
| Debt / Equity | 0.02× | Effectively debt-free |
| Promoter Holding | 65.1% | 0% pledge — clean governance |
| Rev CAGR (3yr) | 28% | Compounding growth story |
| P/E (TTM) | 31–33× | EPS ~₹9.6 · strip other income → 45–50× |
The Committee Votes
Each investor skill was invoked separately on the same stock, on the same day, with access to the same public data. Seven different frameworks. Seven independent verdicts. No investor could see the others’ reasoning before casting their vote — the system is deliberately siloed at the analysis stage to prevent anchoring.
ROE of 12% & ₹81Cr 'other income' bloating earnings. No margin of safety at 31× P/E. Waits for a fat pitch.
P/B of ~4× fails net-net. 0% dividend payout. NCAV far below market cap. Would not touch at this price.
Likes the domain moat. Concerned about declining ROCE trend. Decision Point synergy unproven. Watching.
Revenue up 8 consecutive quarters. PEG ~1.3. Fortune 500 clients. Classic Lynch ten-bagger setup.
Passes quality screen. GenAI pipeline, contract renewals, CFO retention. Scarcity premium justified.
90%+ USD revenue concentration. Thin ADV (~9.8L). $3M Healtheon SAFE note = illiquid startup exposure. Fragile.
Down 40% YTD. Fin Services growing 20% QoQ. GenAI tailwinds. PL Capital Buy target ₹490. Classic setup.
The split is sharper than it looks on paper. The two bears — Graham and Taleb — are bearish for entirely different reasons. Graham is applying a 1930s-era balance sheet screen to a 2026 asset-light services business. Taleb is worried about fragility: the 90%+ USD revenue concentration, the illiquid startup investment, the thin daily trading volume. These are real concerns, but they’re orthogonal to each other and to the bull case.
“The interesting thing about a split committee isn’t the disagreement — it’s that each investor is right within their own framework. The system doesn’t resolve that. It quantifies it.”
The Consensus Formula
The Portfolio Manager doesn’t just count votes. It weights them by conviction. An investor with 90% confidence influences the outcome more than one expressing 55%. The formula maps each signal to a numeric value — Bullish = +1, Neutral = 0, Bearish = −1 — then computes a weighted average.
= 0 − 70 + 0 + 65 + 62 − 58 + 68 = +67
A weighted signal of +0.15 sits in the neutral/contested band (−0.2 to +0.2). The committee is not collectively bullish. It is not collectively bearish. It is split — three bulls vs two bears vs two neutrals, with moderate confidence across the board. What does the Portfolio Manager do with that?
The Portfolio Manager’s Decision
In most investment processes, a contested signal means paralysis. You wait for more data, you keep it on a watchlist, you revisit next quarter. The Portfolio Manager skill takes a more nuanced approach: a split signal doesn’t mean no position, it means a smaller one.
| Weighted Signal | Verdict | Max Allocation |
|---|---|---|
| > 0.7 | Strong Bullish | Up to 20% of portfolio |
| 0.5 – 0.7 | Moderate Bullish | Up to 15% |
| 0.2 – 0.5 | Lean Bullish | Up to 10% |
| −0.2 – +0.2 | ⚠️ Contested | Starter only — Risk capped |
| < −0.5 | Strong Bearish | SELL / Exit |
No portfolio drawdown active. No governance flags. No risk override triggered.
Cautious starter position initiated. Cash preserved for add-on if Q4 FY26 thesis confirms (May 4).
The logic here is important. The +0.15 signal sits just inside the contested band, but there are no hard exclusion flags — no negative FCF, no governance issues, no fraud or regulatory action. The bull case is coherent: a growing analytics firm in a structurally expanding market, down 42% from its peak, with three credible investors willing to buy. The bear case is also coherent, but it’s mostly a valuation and quality argument, not an existential one. So the system takes a starter position — small enough to be defensible, large enough to have skin in the game when the Q4 result lands.
The Risk Manager’s Veto
Even if the Portfolio Manager had come out with a strong bullish consensus, the Risk Manager would have intervened here. The 52-week range of ₹248–₹518 implies annualized volatility of roughly 70% — firmly in the “extreme” tier. The hard cap for extreme volatility stocks is 2% of portfolio. No exceptions, regardless of analyst conviction.
This is the critical distinction the Risk Manager makes: it controls size, not direction. The fact that the stock is highly volatile doesn’t make it a bad investment. It makes it a high-risk investment. Those are different things. The system accounts for that by reducing the position to where the maximum possible loss — if the stock goes to zero — is acceptable.
Stress Tests and the Final Order
Before any position is initiated, the Risk Manager runs four scenarios against the proposed 2% allocation on a ₹1 crore portfolio. All four pass.
| Scenario | Position Loss | % of Portfolio | Status |
|---|---|---|---|
| Market crash (−30%) | ₹60,000 | 0.6% | PASS |
| IT sector meltdown (−50%) | ₹1,00,000 | 1.0% | PASS |
| Rate spike +200bps | ₹55,000 | 0.55% | PASS |
| Stock goes to zero | ₹2,00,000 | 2.0% | PASS |
What to Watch: Catalysts vs Risks
The thesis isn’t a permanent buy-and-forget. It’s a conditional position built around specific monitors. The Q4 FY26 result on May 4 is the first major fork.
What This System Actually Does
The most useful thing about running a committee isn’t the final verdict — it’s what you learn from the disagreements. Ben Graham and Nassim Taleb are both bearish on LATENTVIEW, but they’re saying completely different things. Graham is applying a 1934 framework to a 2026 asset-light software-adjacent business. Taleb is flagging genuine structural fragilities in the revenue mix and balance sheet optionality. Both are “right” within their own models. Only you can decide which framework is most relevant to your own investing time horizon and risk tolerance.
The system doesn’t pretend to resolve that tension. It quantifies it. It takes the disagreement seriously, weights it by conviction, and translates it into a position size that reflects the uncertainty rather than ignoring it. That’s what a real investment process looks like — not the absence of doubt, but the appropriate calibration of it.
“The system doesn’t pretend to resolve tension between frameworks. It quantifies it — and translates uncertainty into position size rather than ignoring it.”
Run This on Any Stock. Free.
Everything in this analysis runs from these 9 Claude skill files. Drop them into ~/.claude/skills/ and invoke with /investor-name in Claude Code, or paste the file content into Custom Instructions on Claude Web.
Warren Buffett
Moat-based value investing with margin of safety. Evaluates durable competitive advantages, owner earnings DCF, and management quality.
Ben Graham
Deep value, margin of safety, and balance sheet strength. Net-net analysis, NCAV, Graham Number, and strict quantitative thresholds.
Charlie Munger
Multi-disciplinary inversion thinking for business quality. Latticework of mental models — avoid stupidity first, brilliance second.
Peter Lynch
Buy what you know — PEG ratio and ten-bagger GARP analysis. Individual investors have an edge in small and mid-cap stocks.
Phil Fisher
Deep qualitative research and management channel checks. 15-point quality screen — scarcity premium justified for exceptional businesses.
Nassim Taleb
Fragility, antifragility, and fat-tail risk. Avoid ruin. Size down anything that can blow up. Barbell strategy for asymmetric payoffs.
Rakesh Jhunjhunwala
India-focused macro + micro stock-picking. Structural India tailwinds, management integrity scoring, and multi-bagger potential.
Portfolio Manager
Synthesizes all investor votes into a confidence-weighted consensus signal. Produces position sizing tiers and an executable allocation decision.
Risk Manager
Hard position limits based on volatility tiers, liquidity risk, and fundamental flags. Final veto authority — controls size, not direction.