{"id":4097,"date":"2026-04-08T15:22:30","date_gmt":"2026-04-08T15:22:30","guid":{"rendered":"https:\/\/developers-test.one2five.digital\/?post_type=docs&p=4097"},"modified":"2026-04-09T10:28:42","modified_gmt":"2026-04-09T10:28:42","slug":"ai-summary","status":"publish","type":"docs","link":"https:\/\/developers-test.one2five.digital\/?docs=ai-summary","title":{"rendered":"AI Summary"},"content":{"rendered":"\n

Overview<\/mark><\/h4>\n\n\n\n

This feature creates AI-powered summaries of product reviews<\/strong> for specific markets, optimizing API usage and token consumption by only processing incremental review changes<\/strong> instead of re-evaluating all reviews repeatedly.<\/p>\n\n\n\n

Data Hierarchy<\/mark><\/h4>\n\n\n\n

Market
\u2514\u2500\u2500 Product
\u2514\u2500\u2500 Reviews<\/p>\n\n\n\n

\n
\n
Property name<\/td>Description<\/td>Default Value<\/td><\/tr>
AI Summaries Generation<\/td>Toggle to enable or disable summary generation for the market<\/td>false<\/td><\/tr>
AI Summaries Distribution<\/td>Toggle to enable or disable the inclusion of summaries in the standard client feed XML file for the specified market<\/td>false<\/td><\/tr>
active.ai.model<\/td>Active ai model in use<\/td>GPT (can be set to ‘COHERE’, ‘MISTRAL’, ‘DEEPSEEK’)<\/td><\/tr>
product.review.size<\/td>The number of least reviews required for the product to generate a summary for<\/td>8<\/td><\/tr>
ai.summary.cron<\/td>cron expression<\/td>0 0 13 *\/3 * ?<\/td><\/tr><\/tbody><\/table><\/figure>\n<\/div>\n<\/div>\n\n\n\n

Workflow<\/mark><\/h4>\n\n\n\n

High level workflow

The system generates AI-based summaries of product reviews at the market level, where each market contains multiple products and each product has multiple reviews. A scheduled cron job runs every three days and processes only those markets where the AI summary feature is enabled. For each product, the system fetches all approved reviews and generates a summary while maintaining a state of which reviews have already been processed. In subsequent runs, instead of reprocessing all reviews, the system identifies only new or previously unprocessed reviews and generates incremental summaries for them. These new summaries are then merged with the existing summary, ensuring that only delta changes are handled. This approach significantly reduces API calls, token usage, and overall billing cost while keeping the summaries up to date efficiently.<\/p>\n\n\n\n

Cron Trigger\n   \u2193\nFetch Enabled Markets\n   \u2193\nFor Each Market\n   \u2193\nFetch Products\n   \u2193\nFor Each Product\n   \u2193\nFetch Incremental Reviews\n   \u2193\nGenerate AI Summary (Partial)\n   \u2193\nMerge with Existing Summary\n   \u2193\nUpdate State of Review<\/code><\/pre>\n\n\n\n
Sample Summary<\/mark><\/h4>\n\n\n\n
{\n  \"overall_sentiment\": \"neutral\",\n  \"summary\": \"The washing machine has received mixed feedback from customers. Many appreciate its quiet operation, ease of use, effective wash cycles, and range of features. However, concerns about durability, reliability, lengthy cycle times, and minor design issues have also been noted.\",\n  \"top_positives\": [\n    \"Quiet operation\",\n    \"Easy to use with a variety of programs\",\n    \"Effective wash cycles\",\n    \"Good build quality with useful features\",\n    \"Eco-friendly settings\"\n  ],\n  \"top_negatives\": [\n    \"Durability issues\",\n    \"Mechanical failures reported\",\n    \"Long cycle times\",\n    \"Problems with washing wool garments\",\n    \"Installation issues leading to water leakage\"\n  ]\n}<\/code><\/pre>\n","protected":false},"excerpt":{"rendered":"
Overview This feature creates AI-powered summaries of product reviews for specific markets, optimizing API usage and token consumption by only processing incremental review changes instead of re-evaluating all reviews repeatedly. Data Hierarchy Market\u2514\u2500\u2500 Product \u2514\u2500\u2500 Reviews Property name Description Default Value AI Summaries Generation Toggle to enable or disable summary generation for the market false … Read more<\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"closed","template":"","meta":{"_eb_attr":"","inline_featured_image":false,"footnotes":""},"doc_category":[11],"doc_tag":[],"class_list":["post-4097","docs","type-docs","status-publish","hentry","doc_category-knowledge-hub"],"jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/developers-test.one2five.digital\/index.php?rest_route=\/wp\/v2\/docs\/4097","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/developers-test.one2five.digital\/index.php?rest_route=\/wp\/v2\/docs"}],"about":[{"href":"https:\/\/developers-test.one2five.digital\/index.php?rest_route=\/wp\/v2\/types\/docs"}],"author":[{"embeddable":true,"href":"https:\/\/developers-test.one2five.digital\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/developers-test.one2five.digital\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=4097"}],"version-history":[{"count":13,"href":"https:\/\/developers-test.one2five.digital\/index.php?rest_route=\/wp\/v2\/docs\/4097\/revisions"}],"predecessor-version":[{"id":4120,"href":"https:\/\/developers-test.one2five.digital\/index.php?rest_route=\/wp\/v2\/docs\/4097\/revisions\/4120"}],"wp:attachment":[{"href":"https:\/\/developers-test.one2five.digital\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=4097"}],"wp:term":[{"taxonomy":"doc_category","embeddable":true,"href":"https:\/\/developers-test.one2five.digital\/index.php?rest_route=%2Fwp%2Fv2%2Fdoc_category&post=4097"},{"taxonomy":"doc_tag","embeddable":true,"href":"https:\/\/developers-test.one2five.digital\/index.php?rest_route=%2Fwp%2Fv2%2Fdoc_tag&post=4097"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}