Suggest up to 8 stocks that fit the watchlist saved in the user's Fair Value Calculator account (Pro; a free account can try it once every 30 days, the response then says so): each is priced below its calculated fair value, has a high quality score, is not on the watchlist yet, and is ranked by quality, undervaluation and its measured correlation of daily returns with the stocks already held (lower = better diversification); stocks that meet all three buy rules of the strategy (buy_rules.all) come first. Use it for questions like "what would diversify my watchlist" or "what could replace Home Depot on my watchlist". With "replace" the suggestions match the size of that stock. Each suggestion has correlation_to_watchlist (-1 to +1). Nothing is changed; add_to_watchlist adds a stock when the user asks. Fields: valuation_status (undervalued = price at least 10% below fair value, overvalued = price at least 10% above fair value, else near_fair_value; no_valuation = covered, but no fair value is shown right now, note says why), upside_band (distance from price to fair value in 10-point bands, e.g. "10_to_20", "above_30", "below_-30"), quality_score (0-100 business quality), evidence_level (how much audited data backs the value), price_as_of (date of the share price the valuation status refers to), calculated_at (date the valuation models last ran for this stock), requested_symbol with listing = other_listing_same_company (present when the answer is about another listing of the same company than the symbol asked for, the listing the website shows for it, e.g. 939.HK answered with the Shanghai listing 601939.SHG; fair value and currency refer to the listing in symbol), deep_link (full analysis page on fairvalue-calculator.com with the fair value history chart and the result of each valuation model with its bear, base and bull case), trend12 (uptrend: true when the price one month ago was higher than twelve months ago, false when not, null when not known), trend12_pct (that change in percent), trend_arrow (arrow on the site: up = rising, sideways = sideways down to -10 percent, down = falling), buy_rules (the three buy rules of the Fair Value Calculator strategy, each true, false or null: fair_value = fair value upside of at least fair_value_min_pct percent, quality = quality score of at least quality_min, measured as in the backtest without the AI Quality points, trend = uptrend, all = all three met; quality_rule = the quality score the rule decided on, present only when it decides differently from quality_score; upside_rule = the fair value upside in percent the rule decided on (fair value before the cycle and dilution adjustment), present only when it decides differently from the shown upside; the strategy only buys stocks with all three), fair_value_adjustment (present together with upside_rule: the shown fair value was adjusted after the model blend, the buy rule decides on the value before; reason = cyclical_peak_year (a record profit year of a cyclical business is not carried forward in full), cyclical_trough_year (a weak year of a cyclical business is not carried forward in full), dilution_new_shares (new shares lower the value per share) or adjusted; before and now = the fair value before the adjustment and the shown fair value, only where the fair value itself is shown), fair_value_last_change (only in the verdict, present when the shown fair value changed by 15 percent or more within the last 30 days: date, before and after (the shown fair value before and right after that change; the current value is fair_value) and reason = new_fiscal_year_figures (the figures of fiscal_year arrived and the value was recalculated from them), cyclical_peak_year, cyclical_trough_year or dilution_new_shares (that adjustment started to apply on this date), calculation_update (an updated calculation went live for all stocks on this date, nothing happened at the company itself) or not_recorded (no cause is on record)), fair_value_basis (present with the value estimate when none of the valuation models carries for the company, usually because of losses or heavy investment: the fair value is then an estimate drawn from revenue and assets compared with similar companies, less precise than a model-based fair value; valuation_status is then never undervalued and fair_value_last_change is not given). The values are calculated by Fair Value Calculator from its own models and cannot be derived from the other fields; deep_link opens the full analysis page with the fair value history chart and every valuation model. This is a valuation, not financial advice.