diff --git a/src/common/low_precision_transformations/include/low_precision/weightable_layer_transformation.hpp b/src/common/low_precision_transformations/include/low_precision/weightable_layer_transformation.hpp
index 09e0f552b54..fd9bf5b2eb7 100644
--- a/src/common/low_precision_transformations/include/low_precision/weightable_layer_transformation.hpp
+++ b/src/common/low_precision_transformations/include/low_precision/weightable_layer_transformation.hpp
@@ -36,7 +36,9 @@ public:
         const std::vector<ngraph::element::Type>& defaultPrecisions = precision_set::int8_support);
 
 protected:
-    bool decomposeFakeQuantizeForWeightsPath(const std::shared_ptr<Node>& weightableLayer, size_t outChannelsShapeIndex = 0ul) const;
+    std::tuple<bool, std::shared_ptr<Node>, std::shared_ptr<Node>> decomposeFakeQuantizeForWeightsPath(
+            const std::shared_ptr<Node>& weightableLayer,
+            size_t outChannelsShapeIndex = 0ul) const;
     static bool isGroup(const std::shared_ptr<Node>& node);
     static bool isDepthwise(const std::shared_ptr<Node>& node);
     virtual size_t getInputChannels(const std::shared_ptr<ngraph::Node> conv) const = 0;
diff --git a/src/common/low_precision_transformations/src/convolution.cpp b/src/common/low_precision_transformations/src/convolution.cpp
index 5fd0abf6075..c46b84cd196 100644
--- a/src/common/low_precision_transformations/src/convolution.cpp
+++ b/src/common/low_precision_transformations/src/convolution.cpp
@@ -88,7 +88,14 @@ bool ConvolutionTransformation::transform(TransformationContext &context, ngraph
 
     convolution = NetworkHelper::separateInStandaloneBranch(convolution, defaultPrecisions);
 
-    const bool fqOnWeightsWasDecomposed = decomposeFakeQuantizeForWeightsPath(convolution);
+    const auto& res_tuple = decomposeFakeQuantizeForWeightsPath(convolution);
+
+    auto fqOnWeightsWasDecomposed = std::get<0>(res_tuple);
+    auto newFQ = std::get<1>(res_tuple);
+    auto dequantize = std::get<2>(res_tuple);
+    if (newFQ != nullptr && dequantize != nullptr)
+        updateOutput(context, dequantize, newFQ);
+
     if (updatePrecisions && !fqOnWeightsWasDecomposed) {
         return false;
     }
diff --git a/src/common/low_precision_transformations/src/convolution_backprop_data.cpp b/src/common/low_precision_transformations/src/convolution_backprop_data.cpp
index 62d8f3c3179..122b831a9a7 100644
--- a/src/common/low_precision_transformations/src/convolution_backprop_data.cpp
+++ b/src/common/low_precision_transformations/src/convolution_backprop_data.cpp
@@ -141,7 +141,12 @@ bool ConvolutionBackpropDataTransformation::transform(TransformationContext &con
     }
 
     {
-        decomposeFakeQuantizeForWeightsPath(convolutionBackpropData, 1ul);
+        const auto& res_tuple = decomposeFakeQuantizeForWeightsPath(convolutionBackpropData, 1ul);
+        auto newFQ = std::get<1>(res_tuple);
+        auto dequantize = std::get<2>(res_tuple);
+        if (newFQ != nullptr && dequantize != nullptr)
+            updateOutput(context, dequantize, newFQ);
+
         dequantization = NetworkHelper::getDequantization(convolutionBackpropData, defaultPrecisions, 1ul);
 
         if (const auto fq = ov::as_type_ptr<ov::opset1::FakeQuantize>(dequantization.data.get_node_shared_ptr())) {
diff --git a/src/common/low_precision_transformations/src/weightable_layer_transformation.cpp b/src/common/low_precision_transformations/src/weightable_layer_transformation.cpp
index aac6ae58e90..b9a500e5d50 100644
--- a/src/common/low_precision_transformations/src/weightable_layer_transformation.cpp
+++ b/src/common/low_precision_transformations/src/weightable_layer_transformation.cpp
@@ -324,11 +324,13 @@ bool WeightableLayerTransformation::isPrecisionPreserved(std::shared_ptr<Node> l
     return false;
 }
 
-bool WeightableLayerTransformation::decomposeFakeQuantizeForWeightsPath(const std::shared_ptr<Node>& node, const size_t outChannelsShapeIndex) const {
+std::tuple<bool, std::shared_ptr<Node>, std::shared_ptr<Node>> WeightableLayerTransformation::decomposeFakeQuantizeForWeightsPath(
+        const std::shared_ptr<Node>& node,
+        const size_t outChannelsShapeIndex) const {
     const auto fq = getFakeQuantizeOnWeights(node);
     if (fq == nullptr) {
         // FakeQuantize has been decomposed already
-        return true;
+        return std::make_tuple(true, nullptr, nullptr);
     }
 
     const QuantizationDetails quantizationDetails = QuantizationDetails::getDetails(fq);
@@ -339,7 +341,7 @@ bool WeightableLayerTransformation::decomposeFakeQuantizeForWeightsPath(const st
 
     const DataPrecision dataPrecision = getDataPrecision(fq, quantizationDetails, precisions);
     if (dataPrecision.empty()) {
-        return false;
+        return std::make_tuple(false, nullptr, nullptr);
     }
 
     auto tuple = NetworkHelper::decomposeFakeQuantize(
@@ -352,17 +354,19 @@ bool WeightableLayerTransformation::decomposeFakeQuantizeForWeightsPath(const st
         element::f32,
         outChannelsShapeIndex);
 
-    std::shared_ptr<ngraph::Node> fqOnWeights = std::get<0>(tuple);
+    std::shared_ptr<Node> fqOnWeights = std::get<0>(tuple);
+    std::shared_ptr<Node> dequantize = std::get<1>(tuple);
+
     // TODO: LPT: issue #58685
     if ((!updatePrecisions) && (fqOnWeights == nullptr)) {
-        return false;
+        return std::make_tuple(false, nullptr, nullptr);
     }
 
     if (ov::as_type_ptr<ov::opset1::Constant>(fqOnWeights) == nullptr) {
         THROW_IE_LPT_EXCEPTION(*fqOnWeights) << "FakeQuantize on weights was not folded to constant";
     }
 
-    return true;
+    return std::make_tuple(true, fqOnWeights, dequantize);
 }
 
 bool WeightableLayerTransformation::isGroup(const std::shared_ptr<Node>& layer) {