1\begin{tabular}{llrrrrr}2\toprule3Contrast & Period & Estimate (pp) & $t$ & 95\% CI & MDE (80\%) & $n$ \\4\midrule5Any AI project vs none, cohort-level (wins.) & full & 1.49 & 0.33 & [-9.22, 12.21] & 12.69 & 8 cohorts \\6Any AI project vs none, firm-level (wins., cohort FE, size, momentum) & full & +2.79 & 0.77 & [-4.3, +9.9] & 10.1 & 529 treated / 26{,}933 \\7 & pre & -0.93 & -0.19 & [-14.37, 12.5] & 13.56 & 5 cohorts \\8 & pre & -1.22 & -0.33 & [-8.6, +6.1] & 10.5 & 241 treated / 20{,}080 \\9 & post & 5.54 & 0.55 & [-37.59, 48.66] & 28.08 & 3 cohorts \\10 & post & +6.64 & 1.19 & [-4.3, +17.6] & 15.7 & 288 treated / 6{,}853 \\11\addlinespace12Any technology project vs none, cohort-level (wins.) & full & -0.49 & -0.40 & [-3.13, 2.15] & 3.45 & 15 cohorts \\13Any technology project vs none, firm-level (wins., cohort FE, size, momentum) & full & +0.09 & 0.12 & [-1.5, +1.6] & 2.2 & 6{,}379 treated / 26{,}933 \\14 & pre & 0.04 & 0.04 & [-2.35, 2.44] & 3.05 & 12 cohorts \\15 & pre & +0.81 & 1.11 & [-0.6, +2.2] & 2.0 & 4{,}518 treated / 20{,}080 \\16 & post & -2.64 & -0.54 & [-23.78, 18.5] & 13.76 & 3 cohorts \\17 & post & -1.92 & -0.93 & [-6.0, +2.1] & 5.8 & 1{,}861 treated / 6{,}853 \\18\addlinespace19T3 $-$ T1 among firms with $>$ 0 technology projects (wins.) & full & +1.48 & 0.70 & [-3.03, 5.99] & & 15 cohorts \\20\bottomrule21\end{tabular}22\par\vspace{2pt}{\footnotesize\emph{Notes:} Any-AI prevalence 2.0\% of firm-years (4.2\% post); cohorts enter the cohort-level contrast only with $\geq$ 15 treated firms. The legacy quintile sort on a variable with median and P75 equal to zero (MDE 10.71 pp) is retained only in Table~\ref{tab:sorts} for comparison. Extractor: gpt-4o-mini (FIRC-0001 pipeline snapshot), run 2026-10-27/28; human precision frame of 200 mentions delivered (results/r3k\_v2/validation), labels pending.}