Aims. Active galactic nuclei (AGN) are typically identified through their distinctive X-ray or radio emissions, mid-infrared (MIR) colors, or emission lines. However, each method captures different subsets of AGN due to signal-to-noise (S/N) limitations, redshift coverage, and extinction effects, underscoring the necessity for a multiwavelength approach for comprehensive AGN samples. This study explores the effectiveness of spectral energy distribution (SED) fitting as a robust method for AGN identification. Methods. Using CIGALE optical-MIR SED fits on DESI Early Data Release galaxies, we compare SED-based AGN selection (AGNFRAC ≥ 0.1) with traditional methods including BPT diagrams, WISE colors, X-ray, and radio diagnostics. Resuts. The SED fitting identifies ∼70% of narrow- and broad-line AGN and 87% of WISE-selected AGN. Incorporating high S/N WISE photometry reduces star-forming galaxy contamination from 62% to 15%. Initially, ∼50% of SED-AGN candidates are undetected by standard methods, but additional diagnostics classify ∼85% of these sources, revealing low-ionization nuclear emission-line regions and retired galaxies potentially representing evolved systems with weak AGN activity. Further spectroscopic and multiwavelength analysis will be essential to determine the true AGN nature of these sources. Conclusions. SED fitting provides complementary AGN identification, unifying multiwavelength AGN selections. This approach enables more complete – albeit somewhat contaminated – AGN samples, which are essential for upcoming large-scale surveys where spectroscopic diagnostics may be limited.