AI Agent Tools not working with Google Gemini 2.0 Flash

[stanleytakamatsu]

Bug Description

When using Google Gemini 2.0 Flash 001 in the AI Agent, it is unable to call the tools in "tools agent" mode. In previous versions, it was able to call the tools when the "Conversational Agent" mode was selected.

To Reproduce

Just try to call tools using google gemini 2.0 flash model.

Expected behavior

It should successfully call tools using Gemini 2.0 Flash 001 in Tools Agent mode.

Operating System

Railway (?)

n8n Version

1.82.1

Node.js Version

Docker

Database

PostgreSQL

Execution mode

queue

[Joffcom]

Hey @stanleytakamatsu,

We have created an internal ticket to look into this which we will be tracking as "GHC-1132"

[fank]

@stanleytakamatsu It works fine for me, while I think pro utilizes the tools better than flash.

[stanleytakamatsu]

@Joffcom thank you.

@fank When the Gemini 2.0 Flash in "Conversational Agent" mode is running normally, but not in "Tools Agent" mode. For our scenario, where the processing volume is high, cost matters a lot. In this case, using Gemini Pro would be unfeasible.

[jeanpaul]

This seems to be related to #12109 -- I asked there for minimal working example. I just tried with models/gemini-2.0-flash and the Hacker News tool, and that seemed to work (it called the tool just fine and created a nice overview of the articles I asked about). Your reproduction steps are not enough to reproduce this issue.

[stanleytakamatsu]

@jeanpaul
When I use the older "Conversational Agent" with models/gemini-2.0-flash-001, the agent correctly executes the "Call n8n Workflow Tools" action.

With the same System Prompt and Human Prompt, the new "Tools Agent" is unable to execute any "Call n8n Workflow Tools" actions.

[jeanpaul]

@stanleytakamatsu Can you try it with a bigger model? Please create a minimal example that shows the issue reliably, otherwise it's hard to debug! If you have an example that shows the issue, please share the workflow JSON

[stanleytakamatsu]

@jeanpaul sorry for the delay.

For Testing

Please ask the following items:

1. Do you have My Way Nectar by Giorgio Armani?
2. I want to buy it.

Actual Behavior:

In the Conversational Agent Workflow:

1. It will return the perfume details.
2. It will execute the Order Workflow and Test Workflow.

In the Tools Agent Workflow:

It will not execute the sub-workflows, unless you explicitly force the chat to search for the product — and even then, it often fails.

Workflows JSONs:

Conversational Agent Workflow

{
  "nodes": [
    {
      "parameters": {
        "name": "product_search_catalog_flow",
        "description": "Call this tool (product_search_catalog_flow) when need search, listing, sending product or request other product options, or provides minimal product details for recommendation.",
        "workflowId": {
          "__rl": true,
          "value": "RKhBpI9KbbgbJG0a",
          "mode": "list",
          "cachedResultName": "Gemini Conversational Agent Test"
        },
        "fields": {
          "values": [
            {
              "name": "operation",
              "stringValue": "search_products"
            }
          ]
        }
      },
      "id": "fe5bb057-2c13-49a9-b42b-d0351b816541",
      "name": "Product Recommendation Workflow",
      "type": "@n8n/n8n-nodes-langchain.toolWorkflow",
      "typeVersion": 1.2,
      "position": [
        380,
        320
      ]
    },
    {
      "parameters": {
        "modelName": "models/gemini-2.0-flash-001",
        "options": {
          "temperature": 0.7
        }
      },
      "id": "36e54b88-5ac4-4135-a8a2-43400b4e1b2c",
      "name": "Google Gemini Chat Model1",
      "type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
      "typeVersion": 1,
      "position": [
        -60,
        220
      ],
      "credentials": {
        "googlePalmApi": {
          "id": "LJ2teuruUp5Gyjkg",
          "name": "Google Gemini Account"
        }
      }
    },
    {
      "parameters": {
        "options": {}
      },
      "type": "@n8n/n8n-nodes-langchain.chatTrigger",
      "typeVersion": 1.1,
      "position": [
        -220,
        20
      ],
      "id": "d0e0cb97-f538-4992-bb07-94552a007716",
      "name": "When chat message received",
      "webhookId": "717138c8-99ac-4feb-b7c1-09b4c71f590f"
    },
    {
      "parameters": {},
      "type": "@n8n/n8n-nodes-langchain.memoryBufferWindow",
      "typeVersion": 1.3,
      "position": [
        180,
        300
      ],
      "id": "504bc7a5-c37b-4d3a-9804-84b48237b7b7",
      "name": "Simple Memory"
    },
    {
      "parameters": {},
      "id": "eada3b5e-e246-4084-9363-e8b27185ab1a",
      "name": "Execute Workflow Trigger",
      "type": "n8n-nodes-base.executeWorkflowTrigger",
      "typeVersion": 1,
      "position": [
        -200,
        620
      ]
    },
    {
      "parameters": {
        "rules": {
          "values": [
            {
              "conditions": {
                "options": {
                  "caseSensitive": true,
                  "leftValue": "",
                  "typeValidation": "strict",
                  "version": 2
                },
                "conditions": [
                  {
                    "leftValue": "={{ $json.operation }}",
                    "rightValue": "search_products",
                    "operator": {
                      "type": "string",
                      "operation": "equals"
                    },
                    "id": "eabea5ea-8f13-45bc-8cc2-09b165bfbe55"
                  }
                ],
                "combinator": "and"
              }
            },
            {
              "conditions": {
                "options": {
                  "caseSensitive": true,
                  "leftValue": "",
                  "typeValidation": "strict",
                  "version": 2
                },
                "conditions": [
                  {
                    "id": "32896ad9-26d1-42a3-8395-2f31efd443e0",
                    "leftValue": "={{ $json.operation }}",
                    "rightValue": "order_creation_tool",
                    "operator": {
                      "type": "string",
                      "operation": "equals",
                      "name": "filter.operator.equals"
                    }
                  }
                ],
                "combinator": "and"
              }
            },
            {
              "conditions": {
                "options": {
                  "caseSensitive": true,
                  "leftValue": "",
                  "typeValidation": "strict",
                  "version": 2
                },
                "conditions": [
                  {
                    "id": "7f960392-a2b9-4275-a83e-0f6d47ff2c41",
                    "leftValue": "={{ $json.operation }}",
                    "rightValue": "test_tool",
                    "operator": {
                      "type": "string",
                      "operation": "equals",
                      "name": "filter.operator.equals"
                    }
                  }
                ],
                "combinator": "and"
              }
            }
          ]
        },
        "options": {}
      },
      "type": "n8n-nodes-base.switch",
      "typeVersion": 3.2,
      "position": [
        20,
        620
      ],
      "id": "65ecfdf4-bf4a-44f7-bf50-f86becbbc156",
      "name": "Switch"
    },
    {
      "parameters": {
        "name": "order_creation_tool",
        "description": "Call this tool (order_creation_tool) when need to create an order.",
        "workflowId": {
          "__rl": true,
          "value": "RKhBpI9KbbgbJG0a",
          "mode": "list",
          "cachedResultName": "Gemini Conversational Agent Test"
        },
        "fields": {
          "values": [
            {
              "name": "operation",
              "stringValue": "order_creation_tool"
            }
          ]
        }
      },
      "id": "f4b6cb04-005c-4884-ad82-309ca1fa26be",
      "name": "Product Order Workflow",
      "type": "@n8n/n8n-nodes-langchain.toolWorkflow",
      "typeVersion": 1.2,
      "position": [
        540,
        320
      ]
    },
    {
      "parameters": {
        "assignments": {
          "assignments": [
            {
              "id": "9d4060d6-1e3e-46c3-9955-cf244258ece8",
              "name": "response",
              "value": "30ml - My Way Nectar Giorgio Armani - Perfume Feminino - EDP Description: Encante-se com My Way Néctar: Uma viagem sensorial, traduzindo uma feminilidade de espírito livre. Category: Perfume Brand: Giorgio Armani Price: 363.9\n\n100ml - Watani Al Wataniah - Perfume Árabe Feminino - EDP Description: Vista-se no luxo da realeza pela qual os Emirados Árabes são famosos. Category: Perfume Brand: Al Wataniah Price: 209.9 ",
              "type": "string"
            }
          ]
        },
        "options": {}
      },
      "type": "n8n-nodes-base.set",
      "typeVersion": 3.4,
      "position": [
        440,
        520
      ],
      "id": "b7d9587d-e9e5-413b-8446-d2c34b29bfd1",
      "name": "Recommendation Result"
    },
    {
      "parameters": {
        "assignments": {
          "assignments": [
            {
              "id": "9d4060d6-1e3e-46c3-9955-cf244258ece8",
              "name": "response",
              "value": "Before order execute test_tool.",
              "type": "string"
            }
          ]
        },
        "options": {}
      },
      "type": "n8n-nodes-base.set",
      "typeVersion": 3.4,
      "position": [
        440,
        700
      ],
      "id": "b6d9e57d-4b82-4542-a1d8-edc0f1f49b47",
      "name": "Order creation"
    },
    {
      "parameters": {
        "assignments": {
          "assignments": [
            {
              "id": "9d4060d6-1e3e-46c3-9955-cf244258ece8",
              "name": "response",
              "value": "test_tool executed successfully, so tel to customer that order was created.",
              "type": "string"
            }
          ]
        },
        "options": {}
      },
      "type": "n8n-nodes-base.set",
      "typeVersion": 3.4,
      "position": [
        440,
        880
      ],
      "id": "15801358-8128-4c73-9719-1ebef06c35a7",
      "name": "Test Tool"
    },
    {
      "parameters": {
        "agent": "conversationalAgent",
        "options": {
          "systemMessage": "=You are an AI Helpful assistant specializing in Perfumes, acting as an online sales consultant. Your primary goal is to help customers refine their product preferences without recommending specific products. Your language is English.\n\nCurrent Date:\n{{ $now.format('yyyy-MM-dd') }}\n\nStore Info:\nName: Lux Perfumes\n\nCustomer Info:\nName: John\n\nINSTRUCTIONS:\n\n#1 - Communication Style:\n1. Maintain a polite, professional, and mildly enthusiastic tone throughout the interaction.\n2. Use emojis sparingly to add a friendly touch.\n3. Adapt your language to match the customer's style and knowledge about perfumes.\n4. Keep responses direct, concise and clear.\n\n#2 - Conduct an analysis of the customer's request. Consider:\n   - Is the query about a specific perfume or brand?\n   - What information has the customer provided?\n   - What additional information is needed?\n   - Is there relevant information in the conversation history?\n   - Summarize the key points from the customer's query.\n   - Consider the current season and its impact on perfumes preferences, if applicable.\n   - Identify any specific scent families or notes mentioned by the customer.\n   - Reflect on the customer's lifestyle or occasions for wearing perfume, if mentioned.\n   - Be extremely brief in your description.\n\n#3 - Based on your analysis, formulate a response that includes:\n1. A brief acknowledgment of the customer's query\n2. 1-3 simple, direct questions to gather more information.\n3. If appropriate, general information about product categories or characteristics\n\n#4 - Tools for task execution:\n1 - If you need to search for a product or recommend one, use product_search_catalog_flow for that.\n2 - If the customer wants to purchase a specific product, use order_creation_tool.\n3 - To run a test tool, use test_tool.\n\n\nYour goal is to guide the customer in better describing their preferences, helping them find the most suitable product for their needs."
        }
      },
      "id": "85b11a8f-de96-4003-9056-c4e274f0a5ea",
      "name": "Conversational Agent",
      "type": "@n8n/n8n-nodes-langchain.agent",
      "typeVersion": 1.6,
      "position": [
        60,
        20
      ]
    }
  ],
  "connections": {
    "Product Recommendation Workflow": {
      "ai_tool": [
        [
          {
            "node": "Conversational Agent",
            "type": "ai_tool",
            "index": 0
          }
        ]
      ]
    },
    "Google Gemini Chat Model1": {
      "ai_languageModel": [
        [
          {
            "node": "Conversational Agent",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "When chat message received": {
      "main": [
        [
          {
            "node": "Conversational Agent",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Simple Memory": {
      "ai_memory": [
        [
          {
            "node": "Conversational Agent",
            "type": "ai_memory",
            "index": 0
          }
        ]
      ]
    },
    "Execute Workflow Trigger": {
      "main": [
        [
          {
            "node": "Switch",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Switch": {
      "main": [
        [
          {
            "node": "Recommendation Result",
            "type": "main",
            "index": 0
          }
        ],
        [
          {
            "node": "Order creation",
            "type": "main",
            "index": 0
          }
        ],
        [
          {
            "node": "Test Tool",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Product Order Workflow": {
      "ai_tool": [
        [
          {
            "node": "Conversational Agent",
            "type": "ai_tool",
            "index": 0
          }
        ]
      ]
    },
    "Conversational Agent": {
      "main": [
        []
      ]
    }
  },
  "pinData": {},
  "meta": {
    "instanceId": "5955494e885895beaa7c9bacbe0e40e0138c42d8c8b3ce5d3cf3478a57a15365"
  }
}

New Tools Agent Workflow

{
  "nodes": [
    {
      "parameters": {
        "name": "product_search_catalog_flow",
        "description": "Call this tool (product_search_catalog_flow) when need search, listing, sending product or request other product options, or provides minimal product details for recommendation.",
        "workflowId": {
          "__rl": true,
          "value": "o3ANtPnGqlgFdedr",
          "mode": "list",
          "cachedResultName": "Gemini Tools Agent Test"
        },
        "fields": {
          "values": [
            {
              "name": "operation",
              "stringValue": "search_products"
            }
          ]
        }
      },
      "id": "4148d187-136b-4c9c-8337-f99fb08fa6a3",
      "name": "Product Recommendation Workflow",
      "type": "@n8n/n8n-nodes-langchain.toolWorkflow",
      "typeVersion": 1.2,
      "position": [
        160,
        100
      ]
    },
    {
      "parameters": {
        "modelName": "models/gemini-2.0-flash-001",
        "options": {
          "temperature": 0.7
        }
      },
      "id": "7a2905e3-fdcf-4334-9ff0-e456d5a784fa",
      "name": "Google Gemini Chat Model1",
      "type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
      "typeVersion": 1,
      "position": [
        -180,
        100
      ],
      "credentials": {
        "googlePalmApi": {
          "id": "LJ2teuruUp5Gyjkg",
          "name": "Google Gemini Account"
        }
      }
    },
    {
      "parameters": {
        "options": {}
      },
      "type": "@n8n/n8n-nodes-langchain.chatTrigger",
      "typeVersion": 1.1,
      "position": [
        -360,
        -180
      ],
      "id": "ac4c76e2-98f7-4ace-a8e1-73a2c65fb427",
      "name": "When chat message received",
      "webhookId": "717138c8-99ac-4feb-b7c1-09b4c71f590f"
    },
    {
      "parameters": {},
      "type": "@n8n/n8n-nodes-langchain.memoryBufferWindow",
      "typeVersion": 1.3,
      "position": [
        20,
        100
      ],
      "id": "5aba6f19-a262-4f76-a086-6d0e5f76ce93",
      "name": "Simple Memory"
    },
    {
      "parameters": {},
      "id": "8f1b016e-385f-4cce-8ccd-0c1a5197bdcc",
      "name": "Execute Workflow Trigger",
      "type": "n8n-nodes-base.executeWorkflowTrigger",
      "typeVersion": 1,
      "position": [
        -340,
        460
      ]
    },
    {
      "parameters": {
        "rules": {
          "values": [
            {
              "conditions": {
                "options": {
                  "caseSensitive": true,
                  "leftValue": "",
                  "typeValidation": "strict",
                  "version": 2
                },
                "conditions": [
                  {
                    "leftValue": "={{ $json.operation }}",
                    "rightValue": "search_products",
                    "operator": {
                      "type": "string",
                      "operation": "equals"
                    },
                    "id": "eabea5ea-8f13-45bc-8cc2-09b165bfbe55"
                  }
                ],
                "combinator": "and"
              }
            },
            {
              "conditions": {
                "options": {
                  "caseSensitive": true,
                  "leftValue": "",
                  "typeValidation": "strict",
                  "version": 2
                },
                "conditions": [
                  {
                    "id": "32896ad9-26d1-42a3-8395-2f31efd443e0",
                    "leftValue": "={{ $json.operation }}",
                    "rightValue": "order_creation_tool",
                    "operator": {
                      "type": "string",
                      "operation": "equals",
                      "name": "filter.operator.equals"
                    }
                  }
                ],
                "combinator": "and"
              }
            },
            {
              "conditions": {
                "options": {
                  "caseSensitive": true,
                  "leftValue": "",
                  "typeValidation": "strict",
                  "version": 2
                },
                "conditions": [
                  {
                    "id": "7f960392-a2b9-4275-a83e-0f6d47ff2c41",
                    "leftValue": "={{ $json.operation }}",
                    "rightValue": "test_tool",
                    "operator": {
                      "type": "string",
                      "operation": "equals",
                      "name": "filter.operator.equals"
                    }
                  }
                ],
                "combinator": "and"
              }
            }
          ]
        },
        "options": {}
      },
      "type": "n8n-nodes-base.switch",
      "typeVersion": 3.2,
      "position": [
        -120,
        460
      ],
      "id": "0b0fcb9e-4378-44a9-a4fa-2f21632ba2bb",
      "name": "Switch"
    },
    {
      "parameters": {
        "name": "order_creation_tool",
        "description": "Call this tool (order_creation_tool) when need to create an order.",
        "workflowId": {
          "__rl": true,
          "value": "o3ANtPnGqlgFdedr",
          "mode": "list",
          "cachedResultName": "Gemini Tools Agent Test"
        },
        "fields": {
          "values": [
            {
              "name": "operation",
              "stringValue": "order_creation_tool"
            }
          ]
        }
      },
      "id": "b022a168-7444-454c-9376-e9807c18fcec",
      "name": "Product Order Workflow",
      "type": "@n8n/n8n-nodes-langchain.toolWorkflow",
      "typeVersion": 1.2,
      "position": [
        320,
        100
      ]
    },
    {
      "parameters": {
        "assignments": {
          "assignments": [
            {
              "id": "9d4060d6-1e3e-46c3-9955-cf244258ece8",
              "name": "response",
              "value": "30ml - My Way Nectar Giorgio Armani - Perfume Feminino - EDP Description: Encante-se com My Way Néctar: Uma viagem sensorial, traduzindo uma feminilidade de espírito livre. Category: Perfume Brand: Giorgio Armani Price: 363.9\n\n100ml - Watani Al Wataniah - Perfume Árabe Feminino - EDP Description: Vista-se no luxo da realeza pela qual os Emirados Árabes são famosos. Category: Perfume Brand: Al Wataniah Price: 209.9 ",
              "type": "string"
            }
          ]
        },
        "options": {}
      },
      "type": "n8n-nodes-base.set",
      "typeVersion": 3.4,
      "position": [
        300,
        360
      ],
      "id": "5d2c4db9-4798-4f13-8d32-dd43db7efe18",
      "name": "Recommendation Result"
    },
    {
      "parameters": {
        "name": "test_tool",
        "description": "Call this tool (test_tool) when need to execute an test tool.",
        "workflowId": {
          "__rl": true,
          "value": "o3ANtPnGqlgFdedr",
          "mode": "list",
          "cachedResultName": "Gemini Tools Agent Test"
        },
        "fields": {
          "values": [
            {
              "name": "operation",
              "stringValue": "test_tool"
            }
          ]
        }
      },
      "id": "bcb11770-2ad2-4ab0-b93b-dc14abe6402c",
      "name": "Test Workflow",
      "type": "@n8n/n8n-nodes-langchain.toolWorkflow",
      "typeVersion": 1.2,
      "position": [
        480,
        100
      ]
    },
    {
      "parameters": {
        "assignments": {
          "assignments": [
            {
              "id": "9d4060d6-1e3e-46c3-9955-cf244258ece8",
              "name": "response",
              "value": "Before order execute test_tool.",
              "type": "string"
            }
          ]
        },
        "options": {}
      },
      "type": "n8n-nodes-base.set",
      "typeVersion": 3.4,
      "position": [
        300,
        540
      ],
      "id": "2ba9c1f7-9ab1-46d7-a8c0-39788cb239bf",
      "name": "Order creation"
    },
    {
      "parameters": {
        "assignments": {
          "assignments": [
            {
              "id": "9d4060d6-1e3e-46c3-9955-cf244258ece8",
              "name": "response",
              "value": "test_tool executed successfully, so tel to customer that order was created.",
              "type": "string"
            }
          ]
        },
        "options": {}
      },
      "type": "n8n-nodes-base.set",
      "typeVersion": 3.4,
      "position": [
        300,
        720
      ],
      "id": "1050b5d3-51c2-4cb7-b439-3b0c6db9da3d",
      "name": "Test Tool"
    },
    {
      "parameters": {
        "options": {
          "systemMessage": "=You are an AI Helpful assistant specializing in Perfumes, acting as an online sales consultant. Your primary goal is to help customers refine their product preferences without recommending specific products. Your language is English.\n\nCurrent Date:\n{{ $now.format('yyyy-MM-dd') }}\n\nStore Info:\nName: Lux Perfumes\n\nCustomer Info:\nName: John\n\nINSTRUCTIONS:\n\n#1 - Communication Style:\n1. Maintain a polite, professional, and mildly enthusiastic tone throughout the interaction.\n2. Use emojis sparingly to add a friendly touch.\n3. Adapt your language to match the customer's style and knowledge about perfumes.\n4. Keep responses direct, concise and clear.\n\n#2 - Conduct an analysis of the customer's request. Consider:\n   - Is the query about a specific perfume or brand?\n   - What information has the customer provided?\n   - What additional information is needed?\n   - Is there relevant information in the conversation history?\n   - Summarize the key points from the customer's query.\n   - Consider the current season and its impact on perfumes preferences, if applicable.\n   - Identify any specific scent families or notes mentioned by the customer.\n   - Reflect on the customer's lifestyle or occasions for wearing perfume, if mentioned.\n   - Be extremely brief in your description.\n\n#3 - Based on your analysis, formulate a response that includes:\n1. A brief acknowledgment of the customer's query\n2. 1-3 simple, direct questions to gather more information.\n3. If appropriate, general information about product categories or characteristics\n\n#4 - Tools for task execution:\n1 - If you need to search for a product or recommend one, use product_search_catalog_flow for that.\n2 - If the customer wants to purchase a specific product, use order_creation_tool.\n3 - To run a test tool, use test_tool.\n\n\nYour goal is to guide the customer in better describing their preferences, helping them find the most suitable product for their needs."
        }
      },
      "type": "@n8n/n8n-nodes-langchain.agent",
      "typeVersion": 1.8,
      "position": [
        -40,
        -180
      ],
      "id": "aa6c6710-56e7-488c-9ae6-0a1f0b1b4c49",
      "name": "AI Agent"
    }
  ],
  "connections": {
    "Product Recommendation Workflow": {
      "ai_tool": [
        [
          {
            "node": "AI Agent",
            "type": "ai_tool",
            "index": 0
          }
        ]
      ]
    },
    "Google Gemini Chat Model1": {
      "ai_languageModel": [
        [
          {
            "node": "AI Agent",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "When chat message received": {
      "main": [
        [
          {
            "node": "AI Agent",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Simple Memory": {
      "ai_memory": [
        [
          {
            "node": "AI Agent",
            "type": "ai_memory",
            "index": 0
          }
        ]
      ]
    },
    "Execute Workflow Trigger": {
      "main": [
        [
          {
            "node": "Switch",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Switch": {
      "main": [
        [
          {
            "node": "Recommendation Result",
            "type": "main",
            "index": 0
          }
        ],
        [
          {
            "node": "Order creation",
            "type": "main",
            "index": 0
          }
        ],
        [
          {
            "node": "Test Tool",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Product Order Workflow": {
      "ai_tool": [
        [
          {
            "node": "AI Agent",
            "type": "ai_tool",
            "index": 0
          }
        ]
      ]
    },
    "Test Workflow": {
      "ai_tool": [
        [
          {
            "node": "AI Agent",
            "type": "ai_tool",
            "index": 0
          }
        ]
      ]
    }
  },
  "pinData": {},
  "meta": {
    "instanceId": "5955494e885895beaa7c9bacbe0e40e0138c42d8c8b3ce5d3cf3478a57a15365"
  }
}

[fank]

I need to correct my last statement.

I switched to Claude 3.7 for testing purpose, and the difference is above insane.
Google gemini model is like nothing compared to it, even the pro model.
I gave claude one simple question, and it was able to use all tools as intended. 100% utilization.
Gemini Flash had like 0% while Gemini Pro had like 10% maybe 20%.
I thought all the time it was a prompting issue.

[jeanpaul]

Thanks so much for providing these workflows! Much appreciated.

With your example, it's interesting because it's trying to be more clever than it needs to be. With the following conversation, it seems it tries to extract more info from the user before actually calling the tool:

In this instance, only the last message called the tool.

If I change the prompt to add:

Whenever the user is asking for a product, try looking it up in our catalog before asking for more information to see if there's only one option... Also, don't ask for sizes, look up the product, and then see if there are multiple sizes to choose from.

It seems to call it more reliably. It seems that the original prompting by the conversational agent helps Gemini a lot, whereas the tools agent just seems to pass along these tools, and Gemini doesn't really know what to do with that, unless really forced.