This article was written by Miles Grauberd – Investment Analyst at I Know First.

Where This Started

Trading started as a side interest, with a small account and an interest in charts. Early years involved a lot of guessing, following forums, and decisions based on hype rather than a defined process. Over time, one thing became clear: the traders who last are the ones who build a system and follow it, not the ones reacting to whatever is trending that day.

That led to treating trading as a discipline instead of a hobby. Managing real money changes the question from “what’s a hot stock” to “what gives an edge, and how is that edge confirmed before money moves.”

That question led to I Know First.

What I Know First Does

I Know First is an algorithmic forecasting service built on an AI model that analyzes historical prices across a large universe of tickers, including world indexes, sector ETFs, S&P500 names, Russell 1000 names, semiconductors, currencies, commodities, and more. Each day it generates two numbers per ticker:

• Signal: a score that can be positive or negative, representing the predicted direction and strength of a move
• Predictability: a score between 0 and 1, representing how reliable the model has historically been on that specific ticker

Both numbers are generated across six timeframes: 3 day, 7 day, 14 day, 1 month, 3 month, and 1 year. A stock can show a weak 3 day signal alongside a strong 1 year signal, which points to a longer term move that hasn’t shown up in the price yet.

The service arrives as a daily email with a report attached, covering dozens of tickers organized into ranked sections. Using it started with reading that report daily and tracking which signals lined up with actual price moves.

Turning Raw Data Into a System

A daily report with many rows is hard to use on its own without a system for tracking it over time.

An automated archive was the first step. A script saves every daily report into a Google Drive folder automatically, building a historical record instead of letting each report sit in an inbox.

Next was the dashboard. Working with Claude, that archive was turned into an Excel workbook with charts tracking signal and predictability over time for any ticker, a leaderboard ranking names by signal per timeframe, and a screen built to flag tickers where the signal was rising before the stock had moved. Months of daily reports became a searchable trend history.

A rising signal paired with rising predictability over several weeks is a useful pattern to track. It suggests the model is picking up something building rather than reacting to a single day’s noise.

A few specific patterns have stood out since the dashboard made trend tracking possible.

The first is a rising signal against a flat or slightly falling stock price. When the signal climbs over a stretch of weeks while the share price stays roughly flat or drifts down, that gap between signal and price is worth acting on.

The second is a divergence pattern, where a stock’s signal is trending downwards for a period, then flips and starts climbing into positive territory. Adobe (ADBE) is showing this pattern currently. Adobe is down over 70% from its highs back in 2024. On June 14th, the 1 year signal was at 43. Less than a month later, the signal was at 126, even though the stock was down 1.35% over that span. This past week, Adobe was upgraded by HSBC from hold to buy, with a price target of $308, about 40% upside from current levels. The signal now sits at 197. Adobe is an ongoing setup and stays on the watchlist.

More patterns like these continue to show up as the dashboard runs longer, and this list will keep getting updated.

Confirming Signals with Technical Analysis

I Know First functions as the alert system, the first signal pointing to where to look.

When a ticker’s signal and predictability both start climbing on the dashboard, that ticker goes on a watchlist. From there, standard technical analysis takes over: support and resistance levels, volume trends, moving averages, and chart patterns on the daily and weekly timeframes. The question at that stage is whether the chart agrees with what the algorithm is showing.

When both line up, an AI signal pointing to strength and a chart confirming accumulation or a breakout setup, that combination is used as confirmation before acting. I Know First scans a large universe of tickers and surfaces the ones worth a closer look. Technical analysis then confirms or contradicts that read on the specific chart.

Real Examples

Every signal referenced below is the 1 year signal.

Robinhood (HOOD): Robinhood hit a high near $153 in October. Seven months later, the stock had fallen to a low of $63. On May 1st, the 1 year signal was at 112, low relative to recent months. Seven days later, on May 7th, the signal had climbed to 312 while the stock stayed flat. That gap between a rising signal and a flat chart was the entry cue. Two months later, Robinhood was up 50%.

Alphabet (GOOGL): Google had slipped roughly 20% off its highs earlier in the year. On March 29th, the stock bottomed and the 1 year signal read 73. Four days later, with the stock up only 7%, the signal had roughly tripled to over 200. Google rallied 38% over the following month.

CrowdStrike (CRWD): On March 29th, CrowdStrike’s 1 year signal was 115. Less than a month later, the signal had climbed to 497, about a 4x increase, while the stock had moved up only about 11%. Two and a half months later, CrowdStrike was at an all time high, up over 83% since the sharp upswing in the signal.

Two names currently on the watchlist: Costco (COST) and ServiceNow (NOW). Both are showing the same pattern as the examples above, a signal rising sharply while the stock price stays roughly flat, the setup that has preceded every move covered so far.

Costco had a signal of 34 on May 29th. A little over a month later, that signal is sitting at 427, a jump of more than 10x, while the stock price has stayed essentially flat over that same stretch. ServiceNow shows a similar gap. On June 11th, the signal was at 174. It has since climbed to well over 500, while the stock itself is up only about 5% in that time. In both cases, the signal has moved far ahead of the price, which is exactly the kind of setup worth watching closely on the dashboard right now.

The Takeaway

I Know First is the base of this process. It processes scale that would be difficult to replicate manually, scanning thousands of tickers across six timeframes daily and surfacing where to look before a move shows up in the price. Without that scan, finding the right stocks at the right time would come down to guesswork. The edge comes from what gets built around that base. Claude turned a daily email report into an automated archive and an interactive dashboard, which made it possible to track signal and predictability over time instead of looking at one day in isolation, and to spot the patterns covered above. Technical analysis is the final check, confirming the entry once I Know First has already done the work of finding the opportunity.

That is the framework: I Know First as the foundation, tools like Claude to organize and track the data at scale, and a final chart check to find the right entry.