diff --git a/src/backends/neon/workloads/CMakeLists.txt b/src/backends/neon/workloads/CMakeLists.txt
index cbe1e3b..7db315f 100644
--- a/src/backends/neon/workloads/CMakeLists.txt
+++ b/src/backends/neon/workloads/CMakeLists.txt
@@ -22,8 +22,12 @@
     NeonConcatWorkload.hpp
     NeonConstantWorkload.cpp
     NeonConstantWorkload.hpp
+    NeonConvertBf16ToFp32Workload.cpp
+    NeonConvertBf16ToFp32Workload.hpp
     NeonConvertFp16ToFp32Workload.cpp
     NeonConvertFp16ToFp32Workload.hpp
+    NeonConvertFp32ToBf16Workload.cpp
+    NeonConvertFp32ToBf16Workload.hpp
     NeonConvertFp32ToFp16Workload.cpp
     NeonConvertFp32ToFp16Workload.hpp
     NeonConvolution2dWorkload.cpp
diff --git a/src/backends/neon/workloads/NeonConstantWorkload.cpp b/src/backends/neon/workloads/NeonConstantWorkload.cpp
index 08f0390..83a2692 100644
--- a/src/backends/neon/workloads/NeonConstantWorkload.cpp
+++ b/src/backends/neon/workloads/NeonConstantWorkload.cpp
@@ -6,6 +6,7 @@
 #include "NeonConstantWorkload.hpp"
 
 #include <arm_compute/core/Types.h>
+#include <BFloat16.hpp>
 #include <Half.hpp>
 #include <aclCommon/ArmComputeTensorUtils.hpp>
 #include <neon/NeonTensorHandle.hpp>
@@ -46,6 +47,11 @@
 
         switch (computeDataType)
         {
+            case arm_compute::DataType::BFLOAT16:
+            {
+                CopyArmComputeITensorData(data.m_LayerOutput->GetConstTensor<BFloat16>(), output);
+                break;
+            }
             case arm_compute::DataType::F16:
             {
                 CopyArmComputeITensorData(data.m_LayerOutput->GetConstTensor<Half>(), output);
diff --git a/src/backends/neon/workloads/NeonConvertBf16ToFp32Workload.cpp b/src/backends/neon/workloads/NeonConvertBf16ToFp32Workload.cpp
new file mode 100644
index 0000000..79d1f22
--- /dev/null
+++ b/src/backends/neon/workloads/NeonConvertBf16ToFp32Workload.cpp
@@ -0,0 +1,43 @@
+//
+// Copyright © 2020 Arm Ltd. All rights reserved.
+// SPDX-License-Identifier: MIT
+//
+
+#include "NeonConvertBf16ToFp32Workload.hpp"
+
+#include <armnnUtils/FloatingPointConverter.hpp>
+
+#include <BFloat16.hpp>
+
+#include <backendsCommon/WorkloadUtils.hpp>
+
+namespace armnn
+{
+
+NeonConvertBf16ToFp32Workload::NeonConvertBf16ToFp32Workload(const ConvertBf16ToFp32QueueDescriptor& descriptor,
+                                                             const WorkloadInfo& info)
+     : BFloat16ToFloat32Workload<ConvertBf16ToFp32QueueDescriptor>(descriptor, info)
+{
+    this->m_Data.ValidateInputsOutputs("NeonConvertBf16ToFp32Workload", 1, 1);
+    GatherTensorHandlePairs(descriptor, m_TensorHandlePairs);
+}
+
+void NeonConvertBf16ToFp32Workload::Execute() const
+{
+    ARMNN_SCOPED_PROFILING_EVENT_NEON("NeonConvertBf16ToFp32Workload_Execute");
+
+    auto convertFunc = [](uint8_t* dst, const uint8_t* src, size_t size)
+        {
+            auto input = reinterpret_cast<const BFloat16*>(src);
+            auto output = reinterpret_cast<float*>(dst);
+            size_t numElements = size/2; // 2 bytes per Bf16
+            armnnUtils::FloatingPointConverter::ConvertBFloat16ToFloat32(input, numElements, output);
+        };
+
+    for (const auto& pair : m_TensorHandlePairs)
+    {
+        CopyTensorContentsGeneric(pair.first, pair.second, convertFunc);
+    }
+}
+
+} //namespace armnn
diff --git a/src/backends/neon/workloads/NeonConvertBf16ToFp32Workload.hpp b/src/backends/neon/workloads/NeonConvertBf16ToFp32Workload.hpp
new file mode 100644
index 0000000..0969088
--- /dev/null
+++ b/src/backends/neon/workloads/NeonConvertBf16ToFp32Workload.hpp
@@ -0,0 +1,26 @@
+//
+// Copyright © 2020 Arm Ltd. All rights reserved.
+// SPDX-License-Identifier: MIT
+//
+
+#pragma once
+
+#include <backendsCommon/Workload.hpp>
+#include <backendsCommon/WorkloadData.hpp>
+#include <neon/workloads/NeonWorkloadUtils.hpp>
+
+namespace armnn
+{
+
+class NeonConvertBf16ToFp32Workload : public BFloat16ToFloat32Workload<ConvertBf16ToFp32QueueDescriptor>
+{
+public:
+    NeonConvertBf16ToFp32Workload(const ConvertBf16ToFp32QueueDescriptor& descriptor, const WorkloadInfo& info);
+    virtual void Execute() const override;
+
+private:
+    using TensorHandlePair = std::pair<const ITensorHandle*, ITensorHandle*>;
+    std::vector<TensorHandlePair> m_TensorHandlePairs;
+};
+
+} //namespace armnn
diff --git a/src/backends/neon/workloads/NeonConvertFp32ToBf16Workload.cpp b/src/backends/neon/workloads/NeonConvertFp32ToBf16Workload.cpp
new file mode 100644
index 0000000..e1aceec
--- /dev/null
+++ b/src/backends/neon/workloads/NeonConvertFp32ToBf16Workload.cpp
@@ -0,0 +1,44 @@
+//
+// Copyright © 2020 Arm Ltd. All rights reserved.
+// SPDX-License-Identifier: MIT
+//
+
+#include "NeonConvertFp32ToBf16Workload.hpp"
+
+#include <BFloat16.hpp>
+#include <Profiling.hpp>
+
+#include <armnnUtils/FloatingPointConverter.hpp>
+
+#include <backendsCommon/WorkloadUtils.hpp>
+
+namespace armnn
+{
+
+NeonConvertFp32ToBf16Workload::NeonConvertFp32ToBf16Workload(const ConvertFp32ToBf16QueueDescriptor& descriptor,
+                                                             const WorkloadInfo& info)
+    : Float32ToBFloat16Workload<ConvertFp32ToBf16QueueDescriptor>(descriptor, info)
+{
+    this->m_Data.ValidateInputsOutputs("NeonConvertFp32ToBf16Workload", 1, 1);
+    GatherTensorHandlePairs(descriptor, m_TensorHandlePairs);
+}
+
+void NeonConvertFp32ToBf16Workload::Execute() const
+{
+    ARMNN_SCOPED_PROFILING_EVENT_NEON("NeonConvertFp32ToBf16Workload_Execute");
+
+    auto convertFunc = [](uint8_t* dst, const uint8_t* src, size_t size)
+        {
+            auto input = reinterpret_cast<const float*>(src);
+            auto output = reinterpret_cast<BFloat16*>(dst);
+            size_t numElements = size/2; // 2 bytes per bf16
+            armnnUtils::FloatingPointConverter::ConvertFloat32ToBFloat16(input, numElements, output);
+        };
+
+    for (const auto& pair : m_TensorHandlePairs)
+    {
+        CopyTensorContentsGeneric(pair.first, pair.second, convertFunc);
+    }
+}
+
+} //namespace armnn
diff --git a/src/backends/neon/workloads/NeonConvertFp32ToBf16Workload.hpp b/src/backends/neon/workloads/NeonConvertFp32ToBf16Workload.hpp
new file mode 100644
index 0000000..bc96c16
--- /dev/null
+++ b/src/backends/neon/workloads/NeonConvertFp32ToBf16Workload.hpp
@@ -0,0 +1,26 @@
+//
+// Copyright © 2020 Arm Ltd. All rights reserved.
+// SPDX-License-Identifier: MIT
+//
+
+#pragma once
+
+#include <backendsCommon/Workload.hpp>
+#include <backendsCommon/WorkloadData.hpp>
+#include <neon/workloads/NeonWorkloadUtils.hpp>
+
+namespace armnn
+{
+
+class NeonConvertFp32ToBf16Workload : public Float32ToBFloat16Workload<ConvertFp32ToBf16QueueDescriptor>
+{
+public:
+    NeonConvertFp32ToBf16Workload(const ConvertFp32ToBf16QueueDescriptor& descriptor, const WorkloadInfo& info);
+    virtual void Execute() const override;
+
+private:
+    using TensorHandlePair = std::pair<const ITensorHandle*, ITensorHandle*>;
+    std::vector<TensorHandlePair> m_TensorHandlePairs;
+};
+
+} //namespace armnn
diff --git a/src/backends/neon/workloads/NeonWorkloads.hpp b/src/backends/neon/workloads/NeonWorkloads.hpp
index 2b7eabe..f255547 100644
--- a/src/backends/neon/workloads/NeonWorkloads.hpp
+++ b/src/backends/neon/workloads/NeonWorkloads.hpp
@@ -13,7 +13,9 @@
 #include "NeonBatchToSpaceNdWorkload.hpp"
 #include "NeonComparisonWorkload.hpp"
 #include "NeonConstantWorkload.hpp"
+#include "NeonConvertBf16ToFp32Workload.hpp"
 #include "NeonConvertFp16ToFp32Workload.hpp"
+#include "NeonConvertFp32ToBf16Workload.hpp"
 #include "NeonConvertFp32ToFp16Workload.hpp"
 #include "NeonConvolution2dWorkload.hpp"
 #include "NeonDepthToSpaceWorkload.hpp"