KRAS and EGFR are known essential mediators of pancreatic cancer development. In addition, KRAS and EGFR have both been shown to interact with and perturb the function of Argonaute 2 (AGO2), a key regulator of RNA-mediated gene silencing. Here, we employed a genetically engineered mouse model of pancreatic cancer to define the effects of conditional loss of AGO2 in KRAS G12D -driven pancreatic cancer. Genetic ablation of AGO2 does not interfere with development of the normal pancreas or KRAS G12D -driven early precursor pancreatic intraepithelial neoplasia (PanIN) lesions. Remarkably, however, AGO2 is required for progression from early to late PanIN lesions, development of pancreatic ductal adenocarcinoma (PDAC), and metastasis. AGO2 ablation permits PanIN initiation driven by the EGFR-RAS axis, but rather than progressing to PDAC, these lesions undergo profound oncogene-induced senescence (OIS). Loss of Trp53 (p53) in this model obviates the requirement of AGO2 for PDAC development. In mouse and human pancreatic tissues, increased expression of AGO2 and elevated co-localization with RAS at the plasma membrane is associated with PDAC progression. Furthermore, phosphorylation of AGO2 Y393 by EGFR disrupts the interaction of wild-type RAS with AGO2 at the membrane, but does not affect the interaction of mutant KRAS with AGO2. ARS-1620, a G12C-specific inhibitor, disrupts the KRAS G12C -AGO2 interaction specifically in pancreatic cancer cells harboring this mutant, demonstrating that the oncogenic KRAS-AGO2 interaction can be pharmacologically targeted. Taken together, our study supports a biphasic model of pancreatic cancer development: an AGO2 -independent early phase of PanIN formation reliant on EGFR-RAS signaling, and an AGO2 -dependent phase wherein the mutant KRAS-AGO2 interaction is critical to prevent OIS in PanINs and allow progression to PDAC.