跳转到主要内容

监测：GenAI Lens

了解你的品牌、产品或竞争对手在主要大型语言模型 (LLM) 中的讨论情况。

Cheyenne V. avatar
作者：Cheyenne V.
本周更新

Meltwater的GenAI Lens让你可以了解自己的品牌在全球领先的生成式人工智能平台上的展示情况。该解决方案会收集大型语言模型 (LLM) 生成的响应，不仅显示品牌相关内容，还会显示这些信息的来源。这种程度的透明度有助于随时了解情况、发现不准确之处，并更好地理解影响人工智能生成内容的信息源。

GenAI Lens目前支持超过90%以上的最常用LLM，包括ChatGPT、Gemini、Perplexity、Claude、Grok和Deepseek。Meltwater正积极扩大覆盖范围，以纳入陆续进入市场的新产品。可以通过监测这些人工智能生成的响应，及早发现潜在的声誉风险，跟踪品牌可见度，并制定更多数据驱动的传播策略。

本文将介绍：

使用GenAI Lens

若要使用GenAI Lens，请按照以下步骤操作：

  1. 点击左侧导航栏中的“Monitor”（监测）

  2. 选择“GenAI Lens”

    Note: GenAI Lens is a shared environment across a Meltwater company, like Explore or Engage. This means that users accessing your Meltwater company will have access to both GenAI Lens and all saved prompts. That said, if your Meltwater company does not have any saved prompts in GenAI Lens, you will be greeted with a screen that invites you to create the first prompt. However, if someone else in your company has created and saved a prompt to GenAI Lens, it will open to that prompt.

  3. To get started, click Create a prompt in the top right-hand corner. This will open the prompt template library.

  4. We have several prompt templates to choose from, organized by type. These include 14 prompts across 6 tabs:

    1. All - all prompt templates

    2. Market

      1. Market trends

      2. Industry outlook

      3. Market movers

    3. 竞争对手

      1. Competitor reviews

      2. Competitor strategy

      3. Competitor buzz

    4. 品牌

      1. Brand strengths and weaknesses

      2. What people are saying

      3. Brand snapshot

    5. Audience and content

      1. Content intent

      2. Audience questions

    6. Products or services

      1. Product feedback

      2. Product vs competitors

      3. Product overview

  5. Select a template, then customize the prompt to match your brand, product, or competitive focus.

    Note: Templates are designed as starting points. Adjust the language and details to create a truly custom prompt tailored to your brand, products, and interests.

  6. Give your prompt a name

  7. 点击“Create”（创建）

  8. Results will appear in GenAI Lens in both a grid and word cloud view.

    Note: It may take up to 20 minutes for GenAI Lens to process your prompt across all AI models and return results the first time. Additionally, after initial results are returned, new results for each prompt are run once per day (every 24 hours).

  9. The results grid is customizable, and you can manage, pin, hide, and reorder columns. You can also sort results by the date column.

    1. Select any kebab menu to

      1. Pin to the left

      2. Pin to the right

      3. Hide Column

      4. Manage Columns

Understanding Your Results

Results Grid View

  1. The results grid view features eight columns:

    1. Model

    2. 提示

    3. 日期

    4. Emotion

    5. Key phrases

    6. 品牌

    7. Product/People

    8. 链接

  2. Each row features results for a prompt from a specific AI Assistant/LLM model broken out by column. Click View more details via the Prompt analysis slide-out panel.

  3. The prompt analysis slide-out panel expands the details of each result in the grid.

    1. Details include:

      1. AI Model and time of last data retrieval

      2. Prompt name

      3. Prompt text

      4. Verbatim LLM response (with copy option)

      5. Top emotion

      6. Top key phrases

      7. Top brands

      8. Top products and people

      9. Top links (with link outs to each source)

  4. The prompt analysis slide-out panel also includes an Export button, which includes all of the above data in a CSV format.

Results Word Cloud

The results word cloud includes:

  • 关键词

  • Organization

  • Product

  • 个人

  • 地点

Note: Word size in the cloud is dependent on the frequency of results. The more occurrences of the word, the larger it appears in the cloud.

  1. Hover over a keyword to view the number of instances it appears in your results.

  2. Hover over the entity type in the legend to isolate those entity types. Click to exclude them from the cloud.

  3. Click to exclude them from the cloud.

过滤

The results grid and word cloud are both dynamic and reflect the filters applied at the top of your GenAI Lens results.

筛选条件包括：

  1. 日期区间

    1. Responses from all models are collected daily. When you set up prompts, you can later filter results by a specific date or date range to view only the responses from that period. Each response includes its collection date, which also appears in a dedicated Day column in the results table.

  2. Saved prompt

    1. For all users

    2. Select any combination of prompts, and delete prompts from the dropdown

  3. LLM/AI Assistant Model

    1. Current models include:

      1. Anthropic Claude Opus 4.1

      2. Anthropic Claude Opus 4

      3. ChatGPT 5

      4. DeepSeek R1

      5. Google Gemini 2.5 Flash

      6. Google Gemini 2.5 Pro

      7. Meta AI Llama 4 Maverick

      8. Perplexity Sonar Pro

      9. xAI Grok 3

      10. xAI Grok 4

用例

个人

GenAI能为他们提供哪些协助

对关键绩效指标 (KPI) 产生积极影响

公关／传播主管

品牌在各种GenAI工具中的可见度 – 了解其品牌在ChatGPT和Gemini等主要语言模型中的呈现方式，为品牌提供目前所缺乏的全新可见度维度。

早期风险检测 – 通过丰富的情感分析、情绪分析及来源数据尽早识别并采取行动应对声誉风险或不准确信息，以便在危机升级前有效加以缓解。

数据驱动的传播策略 – 随着时间的推移跟踪情感分析、叙事趋势和竞争定位，为基于人工智能领域智能的更精准公关策略、媒体邮件联系及内容策略提供决策依据。

品牌情感分析指数

在各平台上保持媒体与消息的一致性

危机响应时间

市场行销

品牌在各种GenAI工具中的可见度 – 了解其品牌在ChatGPT和Gemini等主要语言模型中的呈现方式，为品牌提供目前所缺乏的全新可见度维度。

优化内容和推广活动 – 使用可自定义的提示来跟踪推广活动、产品或品牌信息，并识别在生成式搜索环境中影响绩效的内容缺口。

竞争情报 – 监测竞争对手的定位及其叙述内容如何随时间变化，根据人工智能领域真实的曝光数据制定更明智的内容和定位策略。

人工智能生成内容中的媒体占有率

在各种大型语言模型 (LLM) 中保持推广活动消息的一致性

生成式环境中的品牌可见度指数

💡 提示

需要更多帮助？请随时通过在线聊天与我们联系，或查看我们的“客户社区”。

寻找答案并向Meltwater支持和社区专家处获取帮助。

这是否解答了您的问题？