Abstract Aims Cardiac involvement, a major determinant of prognosis in AL (light-chain immunoglobulin) amyloidosis, is characterized by an impairment of longitudinal strain (LS%). We sought to evaluate the utility of LS% in a prospectively observed series of patients. Methods and results A total of 915 serial newly diagnosed AL patients with comprehensive baseline assessments, inclusive of echocardiography, were included. A total of 628/915 (68.6%) patients had cardiac involvement. The LS% worsened with advancing cardiac stage with mean −21.1%, −17.1%, −12.9%, and −12.1% for stages I, II, IIIa, and IIIb, respectively (P < 0.0001). There was a highly significant worsening of overall survival (OS) with worsening LS% quartile: LS% ≤−16.2%: 80 months, −16.1% to −12.2%: 36 95% confidence interval (CI) 20.9–51.1 months, −12.1% to −9.1%: 22 (95% CI 9.1–34.9) months, and ≥−9.0%: 5 (95% CI 3.2–6.8) months (P < 0.0001). Improvement in LS% was seen at 12 months in patients achieving a haematological complete response (CR) (median improvement from −13.8% to −14.9% in those with CR and difference between involved and uninvolved light chain <10 mg/L). Strain improvement was associated with improved OS (median not reached at 53 months vs. 72 months in patients without strain improvement, P = 0.007). Patients achieving an LS% improvement and a standard N-terminal pro-brain natriuretic peptide-based cardiac response survived longer than those achieving a biomarker-based cardiac response alone (P < 0.0001). Conclusion Baseline LS% is a functional marker that correlates with worsening cardiac involvement and is predictive of survival. Baseline LS% and an absolute improvement in LS% are useful additional measures of prognosis and response to therapy in cardiac AL amyloidosis, respectively. Graphical Abstract Graphical Abstract