{"id":237302,"date":"2026-01-14T07:30:00","date_gmt":"2026-01-14T07:30:00","guid":{"rendered":"https:\/\/osmosys.co\/ca\/?p=237302"},"modified":"2026-01-13T11:48:00","modified_gmt":"2026-01-13T11:48:00","slug":"ai-checklist-dynamics-365-power-platform","status":"publish","type":"post","link":"https:\/\/osmosys.co\/ca\/ai-checklist-dynamics-365-power-platform\/","title":{"rendered":"AI Governance Checklist for Dynamics 365 + Power Platform"},"content":{"rendered":"<div id=\"bsf_rt_marker\"><\/div>\n<p><em><a href=\"https:\/\/osmosys.co\/blog\/digital-transformation-marketing-2026\/\">From copilots to agents: guardrails you can ship without freezing delivery<\/a><\/em><\/p>\n\n\n\n<p>Picture this.<\/p>\n\n\n\n<p>Your sales lead asks for <strong>an AI agent that follows up on every inbound lead<\/strong>.<br>Your service head wants <strong>Copilot to summarize cases and suggest next actions<\/strong>.<br>Your ops team wants <strong>flows that auto-create tasks and route approvals<\/strong>.<\/p>\n\n\n\n<p>All reasonable asks.<\/p>\n\n\n\n<p>Then the uncomfortable question shows up:<\/p>\n\n\n\n<p><strong>If the AI makes a bad call, who owns it?<\/strong><\/p>\n\n\n\n<p>That\u2019s the shift from <strong>copilots (assist)<\/strong> to <strong>agents (act)<\/strong>. When AI moves from suggestions to execution, governance stops being a policy document and becomes an operating habit.<\/p>\n\n\n\n<p>Microsoft\u2019s Responsible AI framing is a useful anchor\u2014<strong>fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability<\/strong>.<\/p>\n\n\n\n<p>This post turns that into a practical, buildable checklist for <strong>Dynamics 365 + Power Platform + Copilot Studio<\/strong> teams: guardrails you can implement, review, and improve without killing momentum.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">What changes when AI becomes \u201cagentic\u201d<\/h2>\n\n\n\n<p>Traditional automation is mostly deterministic: <strong>\u201cif X, then Y.\u201d<\/strong><\/p>\n\n\n\n<p>Agentic AI is different. It can <strong>plan steps<\/strong>, <strong>choose tools<\/strong>, and <strong>chain actions<\/strong> toward a goal. That\u2019s why AI agents are powerful\u2014and why they need controls that are easier to audit than \u201ctrust the model.\u201d<\/p>\n\n\n\n<p>A solid baseline should answer:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>What data can the AI see?<\/strong><\/li>\n\n\n\n<li><strong>What actions can the AI take?<\/strong><\/li>\n\n\n\n<li><strong>What requires approval?<\/strong><\/li>\n\n\n\n<li><strong>How do we review outcomes and catch drift?<\/strong><\/li>\n<\/ul>\n\n\n\n<p>If you can\u2019t answer these quickly, you don\u2019t have AI governance\u2014you have hope.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">AI governance goals (keep them simple)<\/h2>\n\n\n\n<p>If your goals are vague, your controls will be vague. Use three that map to real buyer concerns:<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">1) Safety: reduce harmful actions<\/h3>\n\n\n\n<p>Not only security harm. <strong>Operational harm<\/strong> counts too: wrong emails, wrong customer updates, wrong entitlements, wrong routing.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">2) Privacy: prevent accidental data paths<\/h3>\n\n\n\n<p>Especially through connectors, environment sprawl, and cross-boundary sharing.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">3) Auditability: make outcomes explainable<\/h3>\n\n\n\n<p>You don\u2019t need perfect explainability. You need enough to answer:<br><strong>\u201cWhy did the AI do this?\u201d<\/strong> without detective work.<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><img decoding=\"async\" data-src=\"https:\/\/osmosys.co\/wp-content\/uploads\/2026\/01\/2.png\" alt=\"Need to put some guardrail to your ai.\" class=\"wp-image-239650 lazyload\" src=\"data:image\/svg+xml;base64,PHN2ZyB3aWR0aD0iMSIgaGVpZ2h0PSIxIiB4bWxucz0iaHR0cDovL3d3dy53My5vcmcvMjAwMC9zdmciPjwvc3ZnPg==\" \/><\/figure>\n\n\n\n<h2 class=\"wp-block-heading\">The AI guardrails that matter first in the Microsoft stack<\/h2>\n\n\n\n<h3 class=\"wp-block-heading\">1) Identity and permissions: start with least privilege<\/h3>\n\n\n\n<p>AI features and agents still act through <strong>identities, roles, and connector permissions<\/strong>. The fastest path to trouble is over-permissioning \u201cso it works.\u201d<\/p>\n\n\n\n<p><strong>Practical steps<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Separate who can <strong>build<\/strong>, <strong>test<\/strong>, and <strong>publish<\/strong> AI-enabled assets.<\/li>\n\n\n\n<li>Keep AI agents in environments where access is already controlled.<\/li>\n\n\n\n<li>Treat privileged connectors (finance, customer comms, ERP) like production-grade capabilities\u2014because they are.<\/li>\n<\/ul>\n\n\n\n<p><strong>Quick check<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>If an AI agent can update customer records, ask: <strong>which role grants it that right, and who approved that role?<\/strong><\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">2) Environments: your control plane<\/h3>\n\n\n\n<p>Environments aren\u2019t just containers. They\u2019re a boundary for data, apps, flows, and agents. Microsoft describes a Power Platform environment as a space to store\/manage\/share business data, apps, chatbots, and flows\u2014often to separate assets by roles and security needs.<\/p>\n\n\n\n<p><strong>Practical steps<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Don\u2019t pilot AI agents in the same place where production customer data lives unless guardrails are already in place.<\/li>\n\n\n\n<li>Create a clear promotion path: <strong>dev \u2192 test \u2192 prod<\/strong> (for agents and AI-enabled flows).<\/li>\n\n\n\n<li>Use environment strategy intentionally so \u201cquick pilots\u201d don\u2019t become permanent production.<\/li>\n<\/ul>\n\n\n\n<p><strong>Common failure mode<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u201cWe built it in Default because it was fast.\u201d<br>That\u2019s how AI gets adopted before governance exists.<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">3) DLP: prevent accidental data sharing<\/h3>\n\n\n\n<p>Data loss prevention (DLP) policies in Power Platform act as guardrails to reduce the risk of unintentionally exposing organizational data, including through connector use.<\/p>\n\n\n\n<p><strong>Practical steps<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Start with a baseline: allow required business connectors; restrict consumer connectors by default.<\/li>\n\n\n\n<li>Review DLP like you review firewall rules: regularly, with named ownership.<\/li>\n\n\n\n<li>Document exceptions. If an exception can\u2019t be explained simply, it probably shouldn\u2019t exist.<\/li>\n<\/ul>\n\n\n\n<p><strong>DLP in one sentence<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>DLP is how you stop \u201chelpful AI\u201d from creating unplanned data exits through connectors.<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">4) Copilot Studio governance: treat agents like products<\/h3>\n\n\n\n<p>Copilot Studio documents security and governance controls such as <strong>data residency, DLP, environment routing, and regional customization<\/strong>.<\/p>\n\n\n\n<p><strong>Practical steps<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Define who can publish agents to users (and who can\u2019t).<\/li>\n\n\n\n<li>Require review for any agent that connects to sensitive systems.<\/li>\n\n\n\n<li>Maintain a simple <strong>agent register<\/strong>:\n<ul class=\"wp-block-list\">\n<li>owner<\/li>\n\n\n\n<li>purpose<\/li>\n\n\n\n<li>data sources<\/li>\n\n\n\n<li>permissions<\/li>\n\n\n\n<li>connectors<\/li>\n\n\n\n<li>escalation\/support contact<\/li>\n\n\n\n<li>environments used<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n\n\n\n<p>If AI is in production, it needs a production owner.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Approval design: what can run unattended vs what must be approved<\/h2>\n\n\n\n<p>This is where AI governance becomes real.<\/p>\n\n\n\n<p>Build a simple action-tiering model that your business and security teams can both understand.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Tier 0: Read-only assistance (usually safe to run)<\/h3>\n\n\n\n<p><strong>Examples<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Summarize a case<\/li>\n\n\n\n<li>Draft an email<\/li>\n\n\n\n<li>Suggest knowledge articles<\/li>\n<\/ul>\n\n\n\n<p><strong>Controls<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>user can edit before action<\/li>\n\n\n\n<li>references included (what record(s) the AI used)<\/li>\n\n\n\n<li>clear boundary text (\u201csuggestion, not action\u201d)<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">Tier 1: Low-risk actions (auto-run with guardrails)<\/h3>\n\n\n\n<p><strong>Examples<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Create a follow-up task<\/li>\n\n\n\n<li>Route a case to a queue<\/li>\n\n\n\n<li>Tag a record for review<\/li>\n<\/ul>\n\n\n\n<p><strong>Controls<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>bounded scope (which entities\/tables can be touched)<\/li>\n\n\n\n<li>clear logging<\/li>\n\n\n\n<li>easy rollback (defined undo steps)<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">Tier 2: High-impact actions (require approval)<\/h3>\n\n\n\n<p><strong>Examples<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Send external communication<\/li>\n\n\n\n<li>Change entitlement\/SLA fields<\/li>\n\n\n\n<li>Close cases, cancel orders, update contract terms<\/li>\n\n\n\n<li>Trigger refunds or credit approvals (even indirectly)<\/li>\n<\/ul>\n\n\n\n<p><strong>Controls<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>human approval workflow<\/li>\n\n\n\n<li>audit evidence attached (what the AI used + what it proposed)<\/li>\n\n\n\n<li>separation of duties where appropriate (maker \u2260 approver)<\/li>\n<\/ul>\n\n\n\n<p><strong>Rule of thumb<\/strong><br>If an AI action changes customer experience or financial outcomes, design approval into the workflow.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Logging and review cadence: the part most teams skip<\/h2>\n\n\n\n<p>AI governance fails when nobody reviews outcomes.<\/p>\n\n\n\n<p>Set a lightweight \u201cAI ops\u201d rhythm:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Weekly:<\/strong> agent outcomes review<br>(wrong routes, false positives, user feedback, \u201cwhat surprised us\u201d)<\/li>\n\n\n\n<li><strong>Monthly:<\/strong> permissions + DLP + environment review<br>(new connectors, role creep, access drift)<\/li>\n\n\n\n<li><strong>Quarterly:<\/strong> scenario expansion review<br>(\u201cwhat\u2019s safe to automate next, and what\u2019s still Tier 2?\u201d)<\/li>\n<\/ul>\n\n\n\n<p>This is accountability in practice\u2014supervision as an operating habit, not a slide deck.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Region lens: ANZ, US, Canada (high-level, not legal advice)<\/h2>\n\n\n\n<p>If you support ANZ, US, and Canada, AI governance gets easier when you standardize controls and document data flows.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">ANZ: document and control data flows<\/h3>\n\n\n\n<p>Australia\u2019s privacy framework is anchored in the <strong>Australian Privacy Principles (APPs)<\/strong> under the Privacy Act.<br>New Zealand\u2019s Privacy Act sets privacy principles for how agencies collect, store, use, and share personal information.<\/p>\n\n\n\n<p><strong>Implementation habit<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Maintain a simple \u201cAI data flow sheet\u201d per agent:\n<ul class=\"wp-block-list\">\n<li>sources \u2192 processing \u2192 outputs<\/li>\n\n\n\n<li>environments involved<\/li>\n\n\n\n<li>connectors used<\/li>\n\n\n\n<li>who can access results<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">Canada: be explicit about purpose and access<\/h3>\n\n\n\n<p>PIPEDA applies to private-sector organizations in Canada that collect, use, or disclose personal information in the course of commercial activity.<\/p>\n\n\n\n<p><strong>Implementation habit<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>For each AI scenario, document:\n<ul class=\"wp-block-list\">\n<li>purpose<\/li>\n\n\n\n<li>data used<\/li>\n\n\n\n<li>who can access outputs<\/li>\n\n\n\n<li>retention expectations<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">US: build a baseline that can handle variability<\/h3>\n\n\n\n<p>US privacy requirements vary by state and sector; NCSL tracks ongoing consumer privacy legislation activity.<\/p>\n\n\n\n<p><strong>Implementation habit<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Use a consistent baseline across regions:\n<ul class=\"wp-block-list\">\n<li>least privilege<\/li>\n\n\n\n<li>environment separation<\/li>\n\n\n\n<li>DLP<\/li>\n\n\n\n<li>logging<\/li>\n\n\n\n<li>approval tiering<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n\n\n\n<p>That way you don\u2019t rebuild controls every time the map changes.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">A lightweight rollout that earns trust (without slowing delivery)<\/h2>\n\n\n\n<h3 class=\"wp-block-heading\">Step 1: Pick one scenario with clear boundaries<\/h3>\n\n\n\n<p>Start with Tier 0 (assist) before Tier 2 (act).<\/p>\n\n\n\n<p><strong>Example progression<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Draft follow-up emails \u2192<\/li>\n\n\n\n<li>Suggest follow-up tasks \u2192<\/li>\n\n\n\n<li>Create follow-up tasks \u2192<\/li>\n\n\n\n<li>Send follow-up emails (only with approvals)<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">Step 2: Lock down who can build and publish<\/h3>\n\n\n\n<p>Keep maker roles limited. Make publishing a review gate.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Step 3: Implement DLP and environment separation early<\/h3>\n\n\n\n<p>If you add DLP after sprawl, you create outages and politics. Put boundaries in early while the surface area is small.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Step 4: Add approvals before scaling actions<\/h3>\n\n\n\n<p>Let AI propose, then approve. Expand autonomy only when outcomes are consistently safe.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Step 5: Establish a review cadence + feedback loop<\/h3>\n\n\n\n<p>If users can\u2019t flag bad outputs easily, you won\u2019t improve safely.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">AI governance checklist for Dynamics 365 + Power Platform (copy\/paste)<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Define AI agent scenarios and classify actions (Tier 0 \/ 1 \/ 2)<\/li>\n\n\n\n<li>Assign owners for each AI agent and each environment<\/li>\n\n\n\n<li>Separate build\/test\/prod environments for AI work<\/li>\n\n\n\n<li>Apply baseline DLP policy and document exceptions<\/li>\n\n\n\n<li>Restrict publish permissions; create a review gate<\/li>\n\n\n\n<li>Define what requires human approval (Tier 2 actions)<\/li>\n\n\n\n<li>Enable logging and keep a simple AI agent register<\/li>\n\n\n\n<li>Set weekly outcome review and monthly access\/DLP review<\/li>\n\n\n\n<li>Document AI data flows and purposes (especially multi-region)<\/li>\n\n\n\n<li>Create a rollback plan (what happens when outputs are wrong)<\/li>\n\n\n\n<li>Train users on \u201cAI assists unless approved\u201d<\/li>\n\n\n\n<li>Decide how you\u2019ll measure success (see below)<\/li>\n<\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">What \u201csuccess\u201d looks like (qualitative metrics leaders care about)<\/h2>\n\n\n\n<p>You don\u2019t need vanity metrics. You need early signals that AI is helping without adding risk:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>fewer manual handoffs in service <strong>without<\/strong> increased escalations<\/li>\n\n\n\n<li>higher consistency in case notes and summaries<\/li>\n\n\n\n<li>fewer \u201cwho did this?\u201d incidents due to logging and ownership<\/li>\n\n\n\n<li>improved trust: users adopt AI suggestions because boundaries are clear<\/li>\n<\/ul>\n\n\n\n<p>When AI is governed well, adoption grows because people feel safe using it.<\/p>\n\n\n<div id=\"rank-math-faq\" class=\"rank-math-block\">\n<div class=\"rank-math-list \">\n<div id=\"faq-question-1768290885361\" class=\"rank-math-list-item\">\n<h3 class=\"rank-math-question \">What is AI governance in Dynamics 365 and Power Platform?<\/h3>\n<div class=\"rank-math-answer \">\n\n<p>AI governance is the set of controls that define what AI can access, what AI can do, what requires approval, and how outcomes are logged and reviewed.<\/p>\n\n<\/div>\n<\/div>\n<div id=\"faq-question-1768290894782\" class=\"rank-math-list-item\">\n<h3 class=\"rank-math-question \">Do AI agents need different controls than copilots?<\/h3>\n<div class=\"rank-math-answer \">\n\n<p>Yes. Copilots assist users. AI agents can take actions. The moment AI can change data or trigger workflows, you need approval tiering, least privilege, and audit-ready logging.<\/p>\n\n<\/div>\n<\/div>\n<div id=\"faq-question-1768290903217\" class=\"rank-math-list-item\">\n<h3 class=\"rank-math-question \">What should be governed first: prompts, data, or actions?<\/h3>\n<div class=\"rank-math-answer \">\n\n<p>Start with actions and access. Control identity\/permissions, environment boundaries, and DLP. Then refine prompts and experience patterns.<\/p>\n\n<\/div>\n<\/div>\n<div id=\"faq-question-1768290912879\" class=\"rank-math-list-item\">\n<h3 class=\"rank-math-question \">How do DLP policies help with AI?<\/h3>\n<div class=\"rank-math-answer \">\n\n<p>DLP reduces accidental data exposure by restricting which connectors can be used (and together) in each environment, acting as guardrails for AI-enabled apps and flows.<\/p>\n\n<\/div>\n<\/div>\n<div id=\"faq-question-1768290925602\" class=\"rank-math-list-item\">\n<h3 class=\"rank-math-question \">How do we keep AI governance lightweight?<\/h3>\n<div class=\"rank-math-answer \">\n\n<p>Use a tier model, a small agent register, and a regular review cadence. Governance becomes lighter when it\u2019s built into operating rhythm instead of one-time approvals.<\/p>\n\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n\n\n<h2 class=\"wp-block-heading\">Closing: AI governance is what makes AI usable at scale<\/h2>\n\n\n\n<p>AI agents are not just another feature. They\u2019re a new operational surface area inside CRM and automation. If governance is an afterthought, adoption becomes fragile\u2014people stop trusting the system the first time it makes a costly mistake.<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><img decoding=\"async\" data-src=\"https:\/\/osmosys.co\/wp-content\/uploads\/2026\/01\/3.png\" alt=\"Osmosys can help for AI governed processes\" class=\"wp-image-239651 lazyload\" src=\"data:image\/svg+xml;base64,PHN2ZyB3aWR0aD0iMSIgaGVpZ2h0PSIxIiB4bWxucz0iaHR0cDovL3d3dy53My5vcmcvMjAwMC9zdmciPjwvc3ZnPg==\" \/><\/figure>\n\n\n\n<p>If you\u2019re planning AI agents in Dynamics 365 or Copilot Studio, <a href=\"https:\/\/osmosys.co\/book-a-demo-2\/\"><strong>Osmosys<\/strong> can run a short <strong>AI governance readiness workshop<\/strong><\/a>: scenario selection, tiering, DLP and environment strategy, approval design, and a rollout plan your security and business teams can both live with.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>From copilots to agents: guardrails you can ship without freezing delivery Picture this. Your sales lead asks for an AI agent that follows up on every inbound lead.Your service head wants Copilot to summarize cases and suggest next actions.Your ops team wants flows that auto-create tasks and route approvals. All reasonable asks. Then the uncomfortable [&hellip;]<\/p>\n","protected":false},"author":44,"featured_media":237303,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_et_pb_use_builder":"off","_et_pb_old_content":"","_et_gb_content_width":"","_lmt_disableupdate":"","_lmt_disable":"","jetpack_post_was_ever_published":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[74],"tags":[58,50,66,55],"class_list":["post-237302","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai","tag-ai-in-crm","tag-copilot-ai","tag-dynamics-365","tag-microsoft-dynamics-365"],"modified_by":"mounika","jetpack_featured_media_url":"https:\/\/osmosys.co\/ca\/wp-content\/uploads\/sites\/5\/2026\/01\/1.png","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/osmosys.co\/ca\/wp-json\/wp\/v2\/posts\/237302","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/osmosys.co\/ca\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/osmosys.co\/ca\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/osmosys.co\/ca\/wp-json\/wp\/v2\/users\/44"}],"replies":[{"embeddable":true,"href":"https:\/\/osmosys.co\/ca\/wp-json\/wp\/v2\/comments?post=237302"}],"version-history":[{"count":1,"href":"https:\/\/osmosys.co\/ca\/wp-json\/wp\/v2\/posts\/237302\/revisions"}],"predecessor-version":[{"id":237304,"href":"https:\/\/osmosys.co\/ca\/wp-json\/wp\/v2\/posts\/237302\/revisions\/237304"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/osmosys.co\/ca\/wp-json\/wp\/v2\/media\/237303"}],"wp:attachment":[{"href":"https:\/\/osmosys.co\/ca\/wp-json\/wp\/v2\/media?parent=237302"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/osmosys.co\/ca\/wp-json\/wp\/v2\/categories?post=237302"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/osmosys.co\/ca\/wp-json\/wp\/v2\/tags?post=237302"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}