TL;DR
- Meta is rolling out an AI assistant called Creative Partner inside its Edits video-editing app, Social Media Today reported on September 30, 2026.
- The assistant answers a creator's typed questions by combining their own Instagram performance metrics, comments and platform trends.
- It draws on five specific metrics: follows, views, retention, likes and shares, all tied to the creator's own content.
- Creators reach it through a chat icon inside the Edits editor; Meta did not disclose how its recommendations are generated or how accurate they are.
Meta is rolling out an AI assistant called Creative Partner inside its Edits video-editing app, answering a creator's typed questions by combining their own Instagram performance data with their comments and with platform-wide trends. Social Media Today reported the rollout on September 30, 2026.
What the assistant actually does
The assistant sits inside the Edits editor and responds to a creator's conversational questions, meaning a creator can type a question in normal language rather than digging through a separate analytics tab. According to the report, it answers by combining three distinct inputs: the creator's own Instagram performance metrics, their own comments, and platform trends. The result, per the report, is that the assistant surfaces performance insights and content recommendations directly inside the editing workflow, rather than requiring the creator to leave Edits to check a dashboard elsewhere.
The three inputs behind each answer
The creator's own Instagram performance metrics
The report names five specific metrics the assistant draws on: follows, views, retention, likes and shares. These are the creator's own numbers, tied to their own content, not industry-wide benchmarks or competitor data.
Comments on the creator's content
Alongside the performance numbers, the assistant also factors in comments left on the creator's content. The report does not specify whether this means it reads comment text for sentiment, counts comment volume, or something else; it states only that comments are one of the inputs combined into an answer.
Platform trends
The third input is platform trends. The report does not define what counts as a trend, over what time window, or how broad the platform scope is in this context. It is listed as a third ingredient alongside the creator's own metrics and comments, without further detail.
How creators reach the assistant
The assistant is accessed through a chat icon inside the editor. A creator opens Edits as normal and reaches the assistant through that icon rather than a separate app or menu, keeping the question-and-answer exchange inside the same screen used for cutting and arranging a video.
The report does not say whether a creator can ask the assistant about one specific video's numbers, compare several videos against each other, or only get a general read across recent content. That level of detail sits below what the report covers, so it is a question to test directly once the assistant shows up in a given account rather than one this article can answer in advance.
Why building this into the editor matters
Putting the assistant inside Edits, reachable through a chat icon rather than a separate analytics screen, changes where in a creator's workflow they encounter their own numbers. Before this update, checking how a past video performed and deciding what to make next were two separate steps handled in two different places: the edit itself happened inside Edits, and the performance review happened wherever a creator normally checks Instagram insights, whether that is the app's own analytics tab or a spreadsheet a team keeps for that purpose. By placing the assistant inside the same app used to cut the next video, Meta is connecting those two steps directly, so a creator can ask a question about past performance and act on the answer in the same session, without switching apps or tabs to look the numbers up first.
This is a structural change in where the data sits, not a claim about how good the resulting advice is. The report describes the assistant combining performance metrics, comments and platform trends into an answer; it does not describe the reasoning behind any specific recommendation, so the workflow convenience and the quality of the advice are two separate questions that happen to arrive in the same chat window.
What the assistant combines, at a glance
The table below lists what the report says the assistant draws on and produces, limited to the details given in that report.
| Input or output | What the report says it includes |
|---|---|
| Instagram performance metrics | Follows, views, retention, likes and shares |
| Comments | Comments on the creator's own content |
| Platform trends | Broader trends on the platform, not otherwise defined |
| Output | Performance insights and content recommendations |
| Access point | A chat icon inside the Edits editor |
What this means for Thai creators and brands making Reels
Social Media Today's report does not mention Thailand and does not say anything about how the assistant performs for any specific market. Everything in this section is Relevant Audience's own analysis of what an in-editor assistant like this could change for creators and brands producing Reels content for the Thai market, not a claim from the source.
A large share of brand content aimed at Thai audiences is produced and edited as Reels, often on a weekly or even daily publishing schedule, which puts a premium on how fast a creator or an in-house social team can go from a rough cut to a decision about what to post next. An assistant that sits inside the editor and already has the creator's own follows, views, retention, likes and shares pulled up removes a step that currently happens outside Edits, in whatever analytics view or spreadsheet a team is using to review performance before planning the next video. If the recommendations hold up, that could shorten the loop between publishing a Reel, understanding how it did, and deciding what the next one should look like, particularly for teams producing enough volume that manually reviewing every clip's metrics eats real time. This is the kind of workflow shift Relevant Audience's content marketing team watches for across Reels-heavy accounts.
A second, more specific angle for the Thai market is retention. Short-form video in Thailand competes for attention across Instagram, TikTok and Facebook at the same time, and retention, meaning how long viewers actually keep watching a given Reel, is one of the five metrics the report names the assistant as using. A brand or creator asking the assistant a question about a clip's retention gets an answer built from that creator's own number, not an industry average, which matters because retention benchmarks vary a lot by content category and are not something the report claims the assistant compares against other accounts.
The honest limit sits right next to that possibility: Meta did not disclose how the assistant generates its recommendations or how accurate they are. The report describes what data goes in and what kind of output comes out, but not the method connecting the two, and not any measure of how often the recommendations are actually useful. A team relying on the assistant's suggestions for a Thai-market content calendar should treat its recommendations as a starting point to check against actual results, the same way any new analytics tool earns trust over a run of real posts rather than on the first answer it gives. Running the assistant's suggestions alongside a second opinion, whether that is a human editor's judgment or a separate look at the same metrics, is a reasonable precaution until a team has seen enough of its recommendations play out to know how much weight they deserve.
What the source did not say
Social Media Today's report does not say when the assistant finishes rolling out to all Edits users, whether it is available worldwide at launch, how the assistant's recommendations are generated, or how accurate those recommendations are. It does not mention Thailand or any specific brand or creator, does not say whether the assistant works differently for accounts of different sizes, and does not describe what happens to a creator's data once it has been combined into an answer. None of those gaps are filled in above; where the report stops, this article stops with it rather than filling the space with a guess.
FAQ
What did Meta add to Edits on September 30, 2026?
Meta added an AI assistant called Creative Partner inside its Edits video-editing app, according to Social Media Today's report that day. It answers a creator's typed questions using their own Instagram performance data, their comments, and platform trends.
What Instagram metrics does the assistant actually use?
The report names five metrics: follows, views, retention, likes and shares, all tied to the creator's own content rather than industry benchmarks. It does not say whether the assistant weighs any one of these more heavily than the others when it generates a recommendation.
How does a creator open the assistant inside Edits?
Through a chat icon inside the editor, according to the report, which keeps the question-and-answer exchange inside the same screen used for editing the video. The report does not describe the icon's exact placement on screen or whether it is available at every step of the editing process.
Does Meta explain how the assistant's recommendations are generated or how accurate they are?
No, the report does not disclose the method behind the recommendations or any accuracy figures, so this is an open question rather than a documented detail.
Is this live in Thailand?
Not stated by the source. Social Media Today's report does not mention a country-by-country rollout or confirm availability in Thailand specifically.
Brands and creators building a Reels-heavy content calendar for the Thai market can get a structured read on where this kind of in-editor assistant fits into a wider plan through Relevant Audience's content marketing team, and teams running paid alongside organic Reels may also want to check how it connects with social ads planning for the same content.
