Following Microsoft through the full workflow

You sign in, land on the Universe page, and sort by archetype. MSFT shows up under Reinvestment Growth with a composite score of 80. It cleared every quality filter we have: 10-year revenue history, EPS growth above 3% CAGR, ROIC above WACC in 8 of the last 10 years, FCF positive in 9 of 10, manageable debt, stock-based compensation under the threshold.

What makes the score 80 and not 95: MSFT's revenue CAGR has been decelerating from the Azure-hyper-growth years. Strong in absolute terms, less so than 5 years ago.

You click over to Today's Buy Zone. MSFT sits at the top of the Deep section. Its current P/E is at the ~8th percentile of its own 5-year range — cheaper than it's been on 92 of the last 60 monthly snapshots.

That alone doesn't make it a buy. A P/E compression can mean the market is wrong (opportunity) or the market knows something (trap). The Deep grade is a flag to look harder, not a recommendation.

You click into the MSFT detail page. The score breakdown shows where the 80 comes from across 5 dimensions:

The 5yr P/E chart shows today's 24.5x at a clear extreme on the cheap side. Three shaded bands behind the line mark the 10th, 25th, and 40th percentile thresholds — today's dot is well inside the deepest band.

The valuation looks attractive on its own history. But why? Is the AI cycle over for them? Did earnings miss? Did Azure growth fall off a cliff? The screener can't answer those — that's when you reach for the analyst.

You click Deep Analysis. The screener shows a chronological list of every analysis ever run on MSFT — yours and other subscribers':

The latest is two months old — March 15. You open it for free. It outlines the bull case (Azure pricing power, Copilot adoption, GitHub margin expansion) and the bear case (capex acceleration on AI infra cutting near-term FCF). Verdict: Strong Entry with the caveat "FCF margin will compress short-term before AI ROI lands."

Every Pro subscriber can read every past analysis — no extra charge. The community spend on running these once benefits everyone.

MSFT reported quarterly earnings last week. The cached briefing is 2 months old and pre-earnings. You want today's take. You click Run new analysis.

The system can't quote you an exact price up front — every analysis is unique, and the final cost only resolves once Claude is done. It can confirm you have enough credit to cover the typical range, then start. Real-time financials get pulled, Claude searches the web for fresh news, and 5 sequential AI passes produce the 8-section briefing.

When it finishes, your balance shows the exact deduction: this run cost $1.40, charged after the fact. Failed runs don't cost anything.

You take this briefing, the cached ones, the score, the percentile, and your own judgment — and decide whether to act. The tool doesn't buy stocks for you. It tries to put the strongest evidence on one page so the decision is faster and less anchored on noise.