[DOCS] Add structured output caveats to AI summary guide

[utksh1]

Summary

AI summary docs should explain how structured output expectations differ from free-form summaries.

Problem

The current documentation is either missing a critical explanation, outdated, or fragmented across too many places.

Where to update

• Suggested files: docs/AI_SUMMARY.md, backend/secuscan/ai_summary.py
• Related sections or headings: the sections implied by the title and summary

Expected outcome

A contributor should be able to complete this update with a focused documentation PR and leave the surrounding docs easier to follow.

Acceptance criteria

• The affected docs are updated
• Steps, commands, and references are accurate
• Cross-links are added where they reduce ambiguity
• A maintainer can follow the updated guidance without extra clarification

Verification

Review the changed docs against the current code paths or scripts they describe and spot-check any commands or file references.

Additional context

Prefer precise edits over broad rewrites unless the referenced section is already structurally broken.

[nensii21]

/assign

[Subramaniyajothi6]

@utksh1 can u assign this task to me under gssoc

[tmdeveloper007]

This issue has been open for an extended period with no linked pull request and no recent activity. Suggesting closure to keep the issue tracker focused. Please feel free to reopen if you are actively working on this.