Background/Objectives: Anthracyclines carry considerable cardiotoxicity risk, and atrial fibrillation (AF) during cancer treatment adversely affects morbidity and mortality. We compared the incidence of clinically documented AF and explored its early clinical correlates between anthracycline-containing and non-anthracycline adjuvant regimens in triple-negative breast cancer (TNBC). Methods: This single-centre retrospective cohort included consecutive women with stage II TNBC treated with breast-conserving surgery and adjuvant doxorubicin–cyclophosphamide followed by docetaxel (AC-T) or docetaxel–cyclophosphamide (TC) between 2018 and 2023. ECG and echocardiography were performed at baseline, month one, and month six. Clinically documented AF was defined as AF documented on a scheduled or symptom-triggered 12-lead ECG or as AF lasting ≥30 s on symptom-triggered 72 h Holter monitoring. Patients with structural or coronary heart disease, complete bundle branch block, or non-sinus rhythm were excluded. Results: Among 140 women (mean age, 64 ± 12 years; 64 AC-T), 17 (12.1%) developed clinically documented AF within six months. AF was more frequent with AC-T (18.8% vs. 6.6%; absolute risk difference, 12.2% [95% CI, 1.1–23.3%]; risk ratio, 2.85 [1.07–7.58]; log-rank p = 0.030), and the association persisted after inverse probability of treatment weighting (OR, 3.28 [1.05–10.24]). At one month, PR prolongation > 20 ms was associated with incident AF (OR, 5.16 [1.79–14.89]), whereas conventional echocardiographic changes did not differ between regimens. Conclusions: The AC-T regimen was associated with more clinically documented incident AF than TC. Early PR prolongation, occurring without detectable changes in conventional echocardiographic parameters, may be an informative early electrical signal of AF risk, although prospective validation is required.