Skip to main content
General AI models like ChatGPT, Claude, and Gemini are remarkably capable, and many investors use them to think through stocks. Driven starts from the same kind of model but is built specifically for investing. The difference comes down to four layers: the data the model can see, how analysis is structured, what it remembers, and what it can do on its own.

​
1. Data: training memory vs live professional data

A general model’s financial knowledge comes from its training data and, sometimes, web search. The results are uneven. Search can return mixed-quality sources; without it, the model is working from memory. Either way, timeliness and accuracy are not guaranteed. Ask for a gross margin and you might get last quarter’s figure, or one that was never quite right. Hallucinated financial numbers are especially dangerous because a number that is roughly plausible is harder to catch than one that is obviously wrong. Driven connects to 245 professional data endpoints covering real-time quotes, historical data, financial statements, filings, institutional holdings, insider activity, technical indicators, ETF data, and more. The figures come from databases, not from the model’s recall. When the Agent reports a margin, it pulled it; it did not remember it.

​
2. Analysis: general Q&A vs specialized Skill workflows

A general model answers an investing question by writing a fresh response each time. It has no preset framework; it starts from scratch, and the structure and quality depend heavily on how you phrase the question. Ask the same thing two different ways and you can get two different conclusions. Driven’s Skills are pre-built research workflows. Each one defines an analytical framework, data sources, execution steps, output format, and quality checks, and Driven selects the right one for your question automatically. Deep Research, Valuation Matrix, Stock Screener, Competitor Analysis, Smart Money tracking, and more. This makes the analysis consistent and auditable rather than dependent on prompt phrasing.

​
3. Memory: single conversation vs persistent strategy

A general model does not know you. Each conversation is a fresh start, so your style, preferences, holdings, and risk rules have to be restated every time. Even with a memory feature, it is a generic user profile, not a structured, strategy-level memory of how you invest. Driven’s Playbook is persistent investing strategy memory. It lives across sessions and records your investment universe, watchlists, risk preferences, position-sizing rules, screening criteria, and research style. Every analysis the Agent runs is shaped by the Playbook, and the Playbook gets richer as you keep working.

​
4. Capability: a chat window vs a full research workspace

A general model is, fundamentally, a chat window. You ask, it answers, and when you close the tab the work is done. There is no concept of a workbench that keeps running. Driven provides a complete research workflow:
  • Scheduled tasks run 24/7 automatically, pre-market briefings, monitoring, earnings summaries
  • Portfolio and paper trading track positions and let you test ideas with natural-language orders on real market data
  • Multiple Agents run different strategies independently, without interfering with each other
  • Messaging delivery brings monitoring results to Telegram and WeChat without opening the web app

​
The underlying idea: Agent = Model + Harness

A useful way to frame this comes from AI engineering: an Agent is a model plus a harness. The model is the reasoning core; the harness is everything around it that makes the reasoning useful, the data access, the workflows, the memory, the runtime. A general model gives you the model layer with no harness. You get a smart brain, but with no eyes (it can’t see live data), no hands (it can’t act), no memory (it doesn’t remember your strategy), and no clock (it can’t work on a schedule). Driven is a complete, investing-specific harness built around an equally capable model. The reasoning core is comparable; the difference is the harness. Driven’s value is not “our model is smarter.” It is “we built the best harness for investors.”

​
What general models are still good for

General models are excellent for learning concepts, thinking through ideas, and general reasoning. If you want to understand how a DCF works or talk through the logic of a thesis, they are a fine tool. Driven is for when you want that reasoning wired into live data, structured into repeatable workflows, anchored to your strategy, and able to work while you are away.