This exploratory action research examines the integration of communicative pronunciation activities into audio-lingual practice for Uzbek secondary school learners of English. The study was carried out with 28 students at the A2–B1 CEFR level over a six-week intervention period. Lessons combined traditional audio-lingual drills (minimal pairs, substitution, transformation) with short communicative activities such as role-plays, information-gap tasks, and storytelling designed to recycle target sounds in meaningful contexts. Data collection included pre- and post-tests of pronunciation accuracy, classroom recordings of communicative tasks, and learner self-reports on confidence and motivation. Results demonstrated significant improvement in the production of problematic English sounds (/θ, ð/, /w–v/, and tense–lax vowels /iː–ɪ/), as well as increased intelligibility during spontaneous speech. Students also reported higher motivation and enjoyment compared to drill-only lessons. Findings suggest that an integrated model combining audio-lingual structure with communicative practice can address the persistent pronunciation challenges of Uzbek EFL learners.