Recently, we introduced an efficient high-throughput protocol for backbone assignment of small folded proteins based on two-dimensional (2D) projections of HN(C)N suite of experiments and its automation [Borkar et al., J. Biomol. NMR 2011, 50(3), 285-297]. This strategy provides complete sequence-specific assignment of backbone ((1)H, (15)N, (13)C(α), and (13)C') resonances in less than a day; thus, it has great implications for high-throughput structural proteomics. However, in cases when such small folded protein exhibits substantial amide (1)H shift degeneracy (typically seen in alpha-helical proteins), the strategy may fail or lead to ambiguities. Another limitation is with respect to the identification of checkpoints from the variants of 2D-hncNH spectrum. For example, a protein with many GG, GA, AA, SS, TS, TT, and TS types of dipeptide stretches along its sequence, thus the identification of NH cross-peak corresponding to second G, A, S, or T becomes difficult. In this backdrop, we present here two improvements to enhance the utility of the proposed high-throughput AUTOmatic Backbone Assignment protocol: (i) use of 2D-hNnH spectrum and its variants that display additional (1)H-(15)N correlations and thus help to resolve ambiguities arising because of amide (1)H shift degeneracy and (ii) optimization of the τ(CN) delay in the 2D-hncNH experiment that, when properly adjusted, is observed to help remove ambiguities in the identification of the checkpoints. These improvements have also been incorporated in the automation program AUTOmatic Backbone Assignment. Finally, the performance of the strategy and the automation has been demonstrated using the chicken SH3 domain protein.