AI Drafting and Content Assessment

Overview

StretchSocial includes two optional, administrator-enabled intelligence features that speed up content creation and reduce risk:

1. AI drafting — turns a short brief into ready-to-edit posts tailored to each network, with hashtags.
2. Content assessment — scores a draft's content quality, flags publish risk, and evaluates per-platform fit before it reaches the publishing gate.

Both are gated: your administrator turns them on for your workspace. If you don't see them, they aren't enabled yet — and, in line with governed publishing, neither one posts anything or mutates your content autonomously. The AI advises; people decide.

AI drafting: brief to posts

When AI drafting is enabled, you can generate on-brand copy for several networks from a single brief:

1. In Compose, choose the AI draft option (or the Prompt draft shortcut for a quick body + campaign).
2. Enter a short brief describing the message — for example, Announce our free webinar on scaling social ops, friendly and concise.
3. Optionally specify which platforms to draft for. If you don't, it defaults to X, Instagram, and LinkedIn.
4. The AI returns a post per platform — each with post text that respects that network's norms and length, plus relevant hashtags.
5. Review, edit, and paste the copy you want into the composer, then target networks and add your campaign and CTA as usual.

The feature is fail-safe: if the model is unavailable or the brief is empty, it simply returns no drafts rather than erroring, so it never blocks your work.

Content assessment: score before you publish

When the assessment engine is enabled, StretchSocial can evaluate a draft before it reaches the gate. The assessment is pure, read-only compute — nothing is written, and the app stays the system of record. It considers:

• Content — the caption text.
• Media count — how many images/videos are attached.
• Link count and hashtags — presence and quantity.
• Targets — the networks you're posting to (with their character limits).
• Scheduled hour — the hour of day you plan to post.

From these it produces a content-quality score, publish-risk flags, and a per-platform fit read, so you can tighten a caption, adjust hashtags, or rethink a slot before scheduling.

How AI fits the algorithm contract

StretchSocial's intelligence follows a strict contract. Read-only reviews — caption-quality review, platform-fit summaries, approval-readiness, and queue-health — are available now. Predictive features like publish-risk scoring and best-time recommendations run in dry-run only. Autonomous mutation is disabled and human approval is always required. See Understanding governed publishing for the full contract.

Concrete example

You need launch copy fast:

1. You open AI drafting and enter the brief Launch our Fall Collection, upbeat, drive to shop, choosing LinkedIn and Instagram.
2. The AI returns a longer, professional LinkedIn caption and a punchier Instagram caption with hashtags.
3. You paste the Instagram version into the composer, target Instagram, add campaign Fall Launch and CTA Shop now.
4. You run content assessment with media_count: 1, targets: instagram, scheduled_hour: 10. It flags the caption as strong but suggests one fewer hashtag; you trim it and schedule.

Tips

• Treat AI drafts as a starting point — always edit for brand voice and accuracy before scheduling.
• Give the assessment engine realistic inputs (media count, targets, hour) for the most useful risk read.
• If AI options aren't visible, they're not enabled for your workspace — ask your administrator.

Troubleshooting

• AI returns no posts. The brief may be empty, or the model is temporarily unavailable — retry with a clear brief.
• Assessment says the engine is disabled. The content-assessment engine isn't switched on for your workspace; ask your administrator to enable it.
• I can see drafting but not assessment (or vice-versa). They're separate features enabled independently — one can be on without the other.