diff --git a/src/dotnet/Gatekeeper/Services/AzureContentSafetyService.cs b/src/dotnet/Gatekeeper/Services/AzureContentSafetyService.cs
index eced31d7d8..b6a02d7c56 100644
--- a/src/dotnet/Gatekeeper/Services/AzureContentSafetyService.cs
+++ b/src/dotnet/Gatekeeper/Services/AzureContentSafetyService.cs
@@ -1,6 +1,5 @@
 using Azure;
 using Azure.AI.ContentSafety;
-using FoundationaLLM.Common.Authentication;
 using FoundationaLLM.Common.Constants;
 using FoundationaLLM.Common.Interfaces;
 using FoundationaLLM.Gatekeeper.Core.Interfaces;
@@ -11,7 +10,6 @@
 using System.Text;
 using System.Text.Json;
 
-
 namespace FoundationaLLM.Gatekeeper.Core.Services
 {
     /// <summary>
@@ -21,7 +19,6 @@ public class AzureContentSafetyService : IContentSafetyService
     {
         private readonly ICallContext _callContext;
         private readonly IHttpClientFactoryService _httpClientFactoryService;
-        private readonly ContentSafetyClient _client;
         private readonly AzureContentSafetySettings _settings;
         private readonly ILogger _logger;
 
@@ -43,52 +40,60 @@ public AzureContentSafetyService(
             _httpClientFactoryService = httpClientFactoryService;
             _settings = options.Value;
             _logger = logger;
-
-            //TODO: Use IHttpClientFactoryService
-            //_client = new ContentSafetyClient(new Uri(_settings.APIUrl), DefaultAuthentication.AzureCredential);
         }
 
         /// <inheritdoc/>
         public async Task<AnalyzeTextFilterResult> AnalyzeText(string content)
         {
-            var request = new AnalyzeTextOptions(content);
+            var client = await _httpClientFactoryService.CreateClient(HttpClientNames.AzureContentSafety, _callContext.CurrentUserIdentity);
 
-            Response<AnalyzeTextResult> response;
+            Response<AnalyzeTextResult>? results = null;
             try
             {
-                response = await _client.AnalyzeTextAsync(request);
+                var response = await client.PostAsync("/contentsafety/text:analyze?api-version=2023-10-01",
+                new StringContent(JsonSerializer.Serialize(new AnalyzeTextOptions(content)),
+                Encoding.UTF8, "application/json"));
+
+                if (response.IsSuccessStatusCode)
+                {
+                    var responseContent = await response.Content.ReadAsStringAsync();
+                    results = JsonSerializer.Deserialize<Response<AnalyzeTextResult>>(responseContent);
+                }
             }
             catch (RequestFailedException ex)
             {
                 _logger.LogError(ex, $"Analyze prompt text failed with status code: {ex.Status}, error code: {ex.ErrorCode}, message: {ex.Message}");
-                return new AnalyzeTextFilterResult { Safe = false, Reason = "The content safety service was unable to validate the prompt text due to an internal error." };
+                results = null;
             }
 
+            if (results == null)
+                return new AnalyzeTextFilterResult { Safe = false, Reason = "The content safety service was unable to validate the prompt text due to an internal error." };
+
             var safe = true;
             var reason = "The prompt text did not pass the content safety filter. Reason:";
 
-            var hateSeverity = response.Value.CategoriesAnalysis.FirstOrDefault(a => a.Category == TextCategory.Hate)?.Severity ?? 0;
+            var hateSeverity = results.Value.CategoriesAnalysis.FirstOrDefault(a => a.Category == TextCategory.Hate)?.Severity ?? 0;
             if (hateSeverity > _settings.HateSeverity)
             {
                 reason += $" hate";
                 safe = false;
             }
 
-            var violenceSeverity = response.Value.CategoriesAnalysis.FirstOrDefault(a => a.Category == TextCategory.Violence)?.Severity ?? 0;
+            var violenceSeverity = results.Value.CategoriesAnalysis.FirstOrDefault(a => a.Category == TextCategory.Violence)?.Severity ?? 0;
             if (violenceSeverity > _settings.ViolenceSeverity)
             {
                 reason += $" violence";
                 safe = false;
             }
 
-            var selfHarmSeverity = response.Value.CategoriesAnalysis.FirstOrDefault(a => a.Category == TextCategory.SelfHarm)?.Severity ?? 0;
+            var selfHarmSeverity = results.Value.CategoriesAnalysis.FirstOrDefault(a => a.Category == TextCategory.SelfHarm)?.Severity ?? 0;
             if (selfHarmSeverity > _settings.SelfHarmSeverity)
             {
                 reason += $" self-harm";
                 safe = false;
             }
 
-            var sexualSeverity = response.Value.CategoriesAnalysis.FirstOrDefault(a => a.Category == TextCategory.Sexual)?.Severity ?? 0;
+            var sexualSeverity = results.Value.CategoriesAnalysis.FirstOrDefault(a => a.Category == TextCategory.Sexual)?.Severity ?? 0;
             if (sexualSeverity > _settings.SexualSeverity)
             {
                 reason += $" sexual";
@@ -121,6 +126,7 @@ public async Task<AnalyzeTextFilterResult> AnalyzeText(string content)
                     return "The prompt text did not pass the safety filter. Reason: Prompt injection or jailbreak detected.";
                 }
             }
+
             return null;
         }
     }