{"id":16160,"date":"2026-04-30T02:05:04","date_gmt":"2026-04-29T17:05:04","guid":{"rendered":"https:\/\/m365jp.net\/?p=16160"},"modified":"2026-04-30T02:08:17","modified_gmt":"2026-04-29T17:08:17","slug":"mc1265765-updated-auto-critique-and-model-council-features-in-researcher-frontier-program","status":"publish","type":"post","link":"https:\/\/m365jp.net\/index.php\/2026-04-30-mc1265765-updated-auto-critique-and-model-council-features-in-researcher-frontier-program","title":{"rendered":"MC1265765 | (Updated) Auto Critique and Model Council features in Researcher (Frontier program)"},"content":{"rendered":"<div class=\"postie-post\">\n<div>\n<hr>\n<table id=\"section\">\n<tbody>\n<tr>\n<th width=\"95%\">MC1265765 | (Updated) Auto Critique and Model Council features in Researcher (Frontier program)<\/th>\n<\/tr>\n<\/tbody>\n<\/table>\n<hr>\n<table id=\"data\">\n<tbody>\n<tr>\n<th>Classification<\/th>\n<td>stayInformed<\/td>\n<\/tr>\n<tr>\n<th>Last Updated<\/th>\n<td>04\/29\/2026 16:50:42<\/td>\n<\/tr>\n<tr>\n<th>Start Time<\/th>\n<td>03\/30\/2026 17:13:38<\/td>\n<\/tr>\n<tr>\n<th>End Time<\/th>\n<td>05\/15\/2026 07:00:00<\/td>\n<\/tr>\n<tr>\n<th>Message Content<\/th>\n<td>\n<p>Updated April 29, 2026: We have updated the content. Thank you for your patience.<\/p>\n<p><b>[Introduction]<\/b><\/p>\n<p>  <\/p>\n<p>We\u2019re introducing new capabilities in Researcher to help teams review AI\u2011generated research more confidently and reduce rework before using outputs in business decisions.<\/p>\n<p><b>Critique<\/b> is a built-in review step in Researcher that automatically checks and improves a draft before it\u2019s delivered. It uses a second reasoning pass to strengthen structure and completeness, prioritize reputable sources, and ensure key statements   are grounded in clear citations\u2014so the final report is more reliable and easier to trust.<\/p>\n<p><b>Council<\/b> is Researcher\u2019s multi-model comparison mode. It runs the same question through multiple deep\u2011reasoning research agents (for example, GPT and Claude) in parallel, preserves each model\u2019s full standalone report, and adds a lightweight synthesis   \u201ccover letter\u201d that highlights where the models agree, where they diverge, and what each uniquely contributes\u2014so you can compare perspectives and make decisions with higher confidence.<\/p>\n<p>These preview features are initially available only to organizations enrolled in the Microsoft 365 Copilot Frontier program, which provides early access to experimental Copilot capabilities.<\/p>\n<p>This message is associated with Roadmaps <a href=\"https:\/\/www.microsoft.com\/microsoft-365\/roadmap?filters=&amp;searchterms=553213#Roadmap\" target=\"_blank\">  553213<\/a> and <a href=\"https:\/\/www.microsoft.com\/microsoft-365\/roadmap?filters=&amp;searchterms=558538#Roadmap\" target=\"_blank\">  558538<\/a>.<\/p>\n<p><b>[When this will happen:]<\/b><\/p>\n<p>  <\/p>\n<p>Public preview: Rolling out to Frontier customers by <b>March 30, 2026.<\/b><\/p>\n<p><b>[How this affects your organization:]<\/b><\/p>\n<p>  <\/p>\n<p><b>Who is affected:<\/b><\/p>\n<ul>  <\/p>\n<li>Organizations enrolled in the <b>Microsoft 365 Copilot Frontier program using Researcher<\/b><\/li>\n<p>  <\/p>\n<li>Admins managing AI model access, including third-party models such as Anthropic Claude<\/li>\n<li>Users who generate long-form research reports using Researcher<\/li>\n<li>A Microsoft 365 Copilot (Premium) license is required to use this feature<\/li>\n<p>  <\/ul>\n<p>  <\/p>\n<p><b>What will happen:<\/b><\/p>\n<p>  <\/p>\n<ul>  <\/p>\n<li>Researcher drafts will run through Critique before delivery, improving structure, completeness, and citation quality, when \u201cCritique\u201d is selected as the option in the model picker.&nbsp;<\/li>\n<li>Users will be able to enable <b>Council<\/b>, which generates:\n<ul>\n<li>Two independent, full research reports from different reasoning models (for example, GPT and Claude).<\/li>\n<li>A concise synthesis \u201ccover letter\u201d summarizing areas of agreement, disagreement, and unique contributions.<\/li>\n<\/ul>\n<\/li>\n<li>Critique and Council availability is<b> tenant-admin controlled<\/b>, including the ability to allow or block third-party models such as Claude.<\/li>\n<li>There is <b>no change to existing user workflows<\/b> unless the feature is enabled.<\/li>\n<li>Existing security, compliance, and data handling policies continue to apply.<\/li>\n<\/ul>\n<p><b>[What you can do to prepare:]<\/b><\/p>\n<p>  <\/p>\n<ul>  <\/p>\n<li>Review your <b>AI model access settings<\/b> and confirm whether <b>Anthropic Claude<\/b> is enabled for your tenant.<\/li>\n<p>  <\/p>\n<li>Decide whether to allow Critique and Council based on your organization\u2019s governance and compliance requirements. There are no separate controls for these features; both are enabled when Anthropic and Claude report generation are turned on.<\/li>\n<p>  <\/p>\n<li>Inform helpdesk and support teams about the upcoming preview features.<\/li>\n<p>  <\/p>\n<li>Update internal guidance or training materials if Researcher is documented in your organization.<\/li>\n<\/ul>\n<p>No action is required if you do not plan to enable the preview features.<\/p>\n<p><b>Learn more:<\/b>&nbsp;<a href=\"https:\/\/go.microsoft.com\/fwlink\/?linkid=2356719\" target=\"_blank\">Use model choice in the Researcher agent | Microsoft Support<\/a><span><\/span><\/p>\n<p><b>[Compliance considerations:]<\/b><\/p>\n<p>No compliance considerations identified, review as appropriate for your organization.<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table><\/div>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>MC1265765 | (Updated) Auto Critique and Model Council features in Researcher (Frontier program) Classification [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-16160","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/m365jp.net\/index.php\/wp-json\/wp\/v2\/posts\/16160","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/m365jp.net\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/m365jp.net\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/m365jp.net\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/m365jp.net\/index.php\/wp-json\/wp\/v2\/comments?post=16160"}],"version-history":[{"count":0,"href":"https:\/\/m365jp.net\/index.php\/wp-json\/wp\/v2\/posts\/16160\/revisions"}],"wp:attachment":[{"href":"https:\/\/m365jp.net\/index.php\/wp-json\/wp\/v2\/media?parent=16160"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/m365jp.net\/index.php\/wp-json\/wp\/v2\/categories?post=16160"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/m365jp.net\/index.php\/wp-json\/wp\/v2\/tags?post=16160"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}