Latest in AI
MorningAI Website Agent: How the New Site and Social Workflows Work
MorningAI's October update connects website edits, campaign scheduling and channel analytics, with approval and integration limits worth checking first.
MorningAI announced a Website Agent on October 12 that can edit an existing site through a GitHub connection and show a preview before publication. The same update adds social campaign planning from chat and brings email and web analytics into its Channels view.
For a small marketing team, the interesting part is the combination. A launch rarely needs only a caption or only a landing page. Someone has to connect the page, the distribution plan and the measurement afterwards. Moving those steps closer together could reduce handoffs, provided the team can still inspect what is about to change.
How the Website Agent works
The release notes describe a GitHub connection followed by a generated plan, website changes and a preview.
It can change copy, images and layout, create pages, and edit buttons or links. MorningAI says users can preview desktop, tablet and mobile versions before choosing Publish. A review link can be shared with an engineer first.
That preview is an important part of the workflow. A page can look correct while a form submits to the wrong place or a mobile menu stops working. Before approving an agent-generated page, I would check the actual action the visitor is supposed to complete, alongside the visual result.
What changes for social campaigns
The update also lets users plan, edit and schedule social posts from a conversation, with previews and approval. Related posts can be grouped into campaigns on the calendar. Channels now combines social metrics with email data from Klaviyo and web data from Google Analytics and Search Console.
The practical benefit depends on how the team uses that context. Last week's highest-engagement post might be a useful writing reference, but it is not automatically the best template for this week's launch. A campaign aimed at existing customers can succeed with fewer impressions and more qualified visits.
A useful brief would therefore include the audience, the intended action and what would count as success. Otherwise, bringing analytics into the same chat may simply make it easier to repeat whatever received the most visible engagement.
Check the supported channels and permissions
MorningAI's resource center identifies Facebook, Instagram and LinkedIn as publishing connections. It specifies requirements such as an Instagram business account and a LinkedIn organization page. Its X guide says native X publishing is not supported, so X search in the new release should not be confused with posting to X.
There is also a documentation gap to notice. The older permissions reference says agents do not publish social content from chat, while the newer release describes chat-based scheduling and approval. The newer announcement describes the changed workflow, but teams should verify how the actual approval controls behave in their account.
That older reference also documents per-tool permission settings, with connector writes requiring approval by default. Connecting a tool, generating a draft and publishing a change are different stages. It is worth checking each one when introducing an agent into a live brand workflow.
Pricing and practical limits
The release announcement invites existing users to try the updates, but does not specify a separate Website Agent price, eligible plan list or detailed framework compatibility matrix. MorningAI's billing documentation directs users to Plan & Billing for their account's seats, brands, credits and integrations.
That leaves some straightforward questions for a trial: can it work with your repository, does its preview match production, and what usage consumes credits? The documentation alone does not establish how reliably it can change an unfamiliar codebase, or whether a particular customer setup is supported.
I think the useful shift is that a marketing request can increasingly span several tools without becoming several separate tickets. The next test is whether the review process remains clear enough that the people responsible for the site and the campaign can confidently approve the result.
I'm Paras. I work across technical writing, founder content and distribution.
If you're building an AI product and want help making its workflows and launches understandable to the people who might use it, I'd be happy to talk.