Three mechanisms that should not be conflated
A rules-based screener returns records matching explicit filters. A language-model assistant can help translate a research request, summarize documents or call supported data tools. A predictive model estimates a defined target from inputs and a fitted procedure. A product may combine these mechanisms, but the label AI does not prove that all three are present.
Nor does a traditional screener necessarily use equal factor weights: many simply apply logical conditions. A language-generated explanation is not evidence of model calibration or profitable historical performance.
Inspect the data path
Identify the data provider, available stock universe, reporting period, quote timestamp and result limit. Financial statements, market quotes and analyst projections have different clocks. A refreshed page does not make every field real time.
Missing values should remain identifiable. Check currency, share adjustments and units before comparing companies. A limited result list is not a complete market ranking; a missing company may be outside the available universe or result limit.
What Pineify’s linked free screener actually does
The linked free Pineify screener filters by sector, market capitalization, share price, volume, exchange and country. It displays up to 20 matches with available quotes, selected trailing-twelve-month metrics and one dated annual analyst-estimate record. It does not implement a dividend-safety score, growth-prediction model, natural-language query box or automatic verification of the research criteria below. Missing data is unknown, not a zero or a failed investment test.
The screen requests company matches, then adds available quote, TTM metric and annual estimate data. Failed or unavailable enrichment can leave gaps. The presence of these data fields does not establish automatic sector normalization, analyst-revision history, an ensemble model or an 8–10 safety classification.
What evidence a predictive score would require
Ask what the score predicts: next-month return, benchmark-relative return, direction, or another outcome. Require the universe, horizon, training period, model version, available-at-the-time inputs and treatment of delisted companies. An arbitrary 1–10 scale is not a probability without a documented mapping and calibration.
Evaluate on chronologically separate data and compare a relevant baseline. Include turnover, costs, drawdown and uncertainty. Testing many variants before choosing one introduces selection bias. A paper about a model family cannot establish the performance of a different commercial implementation.
Use AI output as a verifiable research aid
A useful request asks for source links, reporting periods, formulas and an explicit list of unavailable inputs. Recalculate a sample of values against filings. Treat unsupported confidence or guaranteed-return language as a reason to investigate, not as validation.
This page does not claim a Pineify prediction accuracy, a quarterly retraining schedule, factor attribution or a live outperformance record. Those claims require evidence specific to the actual model and deployment.
Frequently asked questions
- Does an AI label mean the tool predicts prices?
- No. It may describe a language interface or data workflow. Check the actual inputs, outputs and documented model target.
- Does the linked free screener show an AI 1–10 score?
- No. Its current workflow provides structured filters and available financial data, not that predictive score.
- Can a research paper validate a commercial score?
- Only evidence about the actual implementation and evaluation can support its performance claims. Results from a different dataset or model do not transfer automatically.
Sources and further research
Educational research only. Screening results and illustrative calculations are not recommendations, forecasts or guarantees. Verify source documents, data periods and assumptions before making an investment decision.