Skip to main content
Companion product to the Contently platform

Six AI agents. One editorial workflow.

AI Studio is a separate, opt-in product for teams that want AI-assisted workflows. The core Contently platform uses no AI by default.

Brand Voice, SEO Optimization, Quality Assurance, Fact Check, LLM Optimization, AI Detectability — six specialized agents, every one named, audited, and reviewed by a human editor before publish.

How AI Studio works

Inputs → Draft → Review. Every step audited.

Source library · loaded
  • FINRA filings · Q1 2026 (12 sources)
  • SEC commentary letters (4)
  • Bank for International Settlements working paper · Feb 2026
  • +18 more
01 · Inputs

Sources first. Every claim traceable to a regulator-approved document.

AI Studio pulls from regulator filings, peer-reviewed papers, and your firm's own published material — never the open web. Every fact in the draft links back to its source.

Draft · in progress

Following a successful 2008 stress test, regional banks faced new capital requirements that...

↑ Citation: BIS working paper, p. 42
02 · Draft

A draft your editors recognize — in your firm's voice, with citations attached.

Trained on your published library. Defaults to your firm's tone, structure, and terminology. The draft arrives with inline citations and a confidence score per paragraph.

Review queue
  • Fact Check pass · 0 contradictions
  • Brand Voice: 6 changes applied
  • Awaiting human editor sign-off
03 · Review

Editors stay in the loop. Quality Assurance never gets bypassed.

Human editors approve every claim, every citation, every voice change. Fact Check flags contradictions before legal sees the draft. Nothing ships without sign-off.

The six agents

Specialists, not generalists.

Each agent is named, versioned, and audited. Tap any tab to see what it actually returns to the editorial queue.

brand-voice v8banned-list=14 phrases · tone-vectors=loaded "mission-critical" → "load-bearing" (banned) "in today's fast-paced" → deleted (filler) 6 changes proposed · 5 accepted automatically · 1 deferred to editor

Banned-phrase list, terminology dictionary, structure rules — all enforced as the draft is written, not after.

seo v6target=primary-keyword + 4 supporting title + H2 structure scored 92/100 · intent match: informational internal-link suggestions: 3 (anchor text varied) schema hints emitted · FAQ + Article handoff ready

Keyword targeting, heading structure, internal-link suggestions, and schema hints — tuned for the search query the page is actually trying to rank for.

qa v5checks=grammar + style + readability + dead-links 22 line edits applied · readability grade 11.4 (target band) 2 broken outbound links → replaced with archive.org snapshot handoff to human editor · estimated review time: 8 min

Grammar, style, readability, dead-link sweep — the editorial pass that catches what the writer's eye stopped seeing on round three.

factcheck v6scope=numerical + named-entity claims claim "$2.1T outstanding" verified · SIFMA 2026-Q1 release claim "FOMC raised 50bps in March" verified · FOMC release claim "regional banks face $400B reserve gap" — 2 sources contradict, flagged for editor

Every numerical claim and every named-entity assertion is verified against its source document. Contradictions surface in the editor's queue, not on publish.

llmo v4targets=ChatGPT + Perplexity + Google AIO + Gemini extractable-snippet score: 87/100 (cite-ready) structured-answer block added at section 02 · Q-shaped H2 citation-anchor density: 1 per 180 words (above threshold)

Structures the page so AI search engines can quote it. Question-shaped headings, extractable answer blocks, citation-anchor density — the things that earn the citation, not just the impression.

aidetect v3detectors=GPTZero + Originality + Copyleaks paragraph 4: 78% AI-likelihood → rewritten with editor voice final pass: median 12% across detectors · in human-prose band handoff to human editor · flagged passages annotated

Runs the draft against the major AI-detectors and rewrites the paragraphs that read as machine-generated — before they ever reach a reader or a search-engine quality reviewer.

AI-search citation share

Measured weekly. Reported monthly.

Eight quarters of citation share for a typical Contently FS customer — on muni-bond commentary across ChatGPT, Perplexity, Gemini, and Google AI Overviews.

Citation share · muni-bond commentary +18pp last 8 quarters

10% = 12 of 120 measured prompts cite the firm. 92% = 110 of 120.

Customer outcome · financial services

A $35B financial-services firm shipped 252 localized articles in 24 hours — every piece routed through compliance, every one signed off by a named editor.

See the case study →

Recognition G2 Leader · Content Marketing Creation G2 Enterprise Leader 2026 Forrester Wave participant
AI governance SOC 2 Type II No model training on customer data Read the policy →

Attestation status may change. Contact security@contently.com for the current list.

See AI Studio in your queue

Six agents. One review queue. Your editors stay in charge.

Book an AI Studio walkthrough See LLM Optimization