Interview Quote Permissions and Fact-Checking Workflow
Before publishing interview quotes, set the attribution level at the start of the call, track every quote in a clearance tracker, verify the facts inside quotes against primary sources, and get sign-off on exact wording where you promised it. Copy the tracker.
Permission to publish interview quotes is something you establish, record, and verify — not something you assume because the expert agreed to talk. The working rule: set the attribution default at minute zero of the call, log every quote you plan to use in a clearance tracker with its permission status, fact-check the factual claims inside quotes before they ship, and give the expert a bounded review of their own words where you promised one. An article is not publish-ready while any quote in the tracker is pending.
This matters because interview quotes fail in three distinct ways — a permission failure (the expert never agreed to be quoted that way), an attribution failure (the wording or label misrepresents who said what on what basis), and an accuracy failure (the quote is faithful, but the fact inside it is wrong). Each failure has a different fix, and the tracker exists to catch all three before publication instead of in a correction.
One boundary up front: this is an editorial workflow, not legal advice. Consent norms for interviews, defamation exposure, and copyright in quoted material vary by jurisdiction and situation; when the stakes are commercial or adversarial, involve counsel.
Who this is for
This is for editors, writers, and content leads publishing interview-led articles on company blogs and independent sites — usually with cooperative sources: customers, in-house experts, practitioners who want the exposure. You are not running an investigative newsroom, but you are publishing real people's words next to your brand, and an unapproved or wrong quote costs you the relationship and the article's credibility at the same time.
It assumes the interview itself already happened, ideally recorded with consent and run against an interview brief that flagged source-sensitive claims in advance.
What attribution levels should you agree on before the call?
Agree on one of five levels before the interview produces anything quotable: on record named, on record role-only, paraphrase-only, background, or approval-before-publish. Attribution levels are only meaningful when both sides agree on them before the material exists. Reuters' interview guidance is built on that premise: ground rules are set in advance, and quotes are handled exactly as agreed. Adapt the newsroom ladder to content work:
| Level | What it means here | When to use it |
|---|---|---|
| On record, named | Quote verbatim, with name, role, and company | The default for expert-led marketing and educational content |
| On record, role only | Quote verbatim, attributed to "a platform engineer at a mid-size SaaS company" | Expert is willing but employer approval is uncertain |
| Paraphrase only | Ideas usable, no quotation marks, named or role-based attribution as agreed | Expert speaks carefully but writes nothing off the cuff |
| Background | Informs your analysis; not quoted or attributed | Context-setting conversations, competitive detail |
| Approval-before-publish | Any of the above, plus the expert sees their quotes in context first | Common courtesy in content marketing; near-universal with customers |
Two practical notes. First, in content marketing — unlike news — quote approval is usually cheap to grant and buys goodwill; the tradeoff is that approval invites softening, so bound it (below). Second, when the expert speaks as an employee, ask directly at minute zero: "does your comms team need to see quotes before we publish?" Discovering the answer after layout is the expensive version.
How do you track quote permissions before publishing?
Track them in a quote clearance tracker: a per-quote table kept next to the draft, recording each quote's speaker and attribution level, where it appears, its permission status, the claim inside it, and its fact-check status. One row per quote or attributed paraphrase you intend to publish:
| Quote (short form) | Speaker + attribution level | Where it appears | Permission status | Claim inside the quote | Fact-check status |
|---|---|---|---|---|---|
| "We cut the form from eleven fields to four" | J.D., named, on record | Section 3 | approved 2026-09-13 | Product change described from first-hand work | speaker's own work — confirmed with speaker |
| "Vendor X's API drops webhooks under load" | J.D., named, on record | Section 4 | pending | Third-party product behavior | needs primary source or independent confirmation — do not publish on the quote alone |
| Paraphrase: rollout took most of a quarter | J.D., role only | Section 2 | approved (paraphrase-only) | Timeline from speaker's own project | confirmed with speaker |
| "This is the only approach that works" | J.D., named | Intro | approved | Universal claim | reframe as opinion or cut — cannot be verified as fact |
The two right-hand columns are the point of the artifact: permission and accuracy are tracked separately because they fail separately. A publish-ready article is one where every row reads approved and every claim column reads confirmed, sourced, reframed as opinion, or cut.
The second row is the pattern that burns blogs most often: an expert confidently describing a third party's product, pricing, or policy. The quote being real does not make the claim true. Claims about third parties get a primary source or independent confirmation, or the quote gets reframed as clearly-labeled single-source experience ("in J.D.'s deployment, they saw...") — or cut.
How do you get from transcript to claims you can publish?
1. Harvest and log
Pull every quote you might use into the tracker with timestamps. Log attributed paraphrases too — attribution failures happen just as easily in paraphrase, and a paraphrase that misstates the expert's view is harder for them to catch later.
2. Classify each claim inside each quote
Three bins: the speaker's own experience (confirm details with the speaker), verifiable external facts (verify against primary sources exactly as you would an unquoted claim — the sourcing workflow applies inside quotation marks), and opinion (publishable as opinion, framed as such).
3. Clean up wording without changing meaning
Removing filler and false starts is standard; anything more gets the paraphrase treatment or a re-ask ("can I tighten this to X?"). Never move a quote to a context that changes what the speaker was answering — that is the subtle attribution failure, and it is why the tracker records where each quote appears.
4. Run a bounded approval pass
Send the expert their quotes in context — the surrounding paragraph, not the whole draft — with a reply-by date and clear scope: they are checking accuracy of their words and facts, not editing your analysis. Accept corrections of substance, push back on hedging that deletes the specific ("we improved the form" is not a correction of "eleven fields to four", it is a retreat — ask which number is wrong).
5. Freeze and archive
When rows go green, mark the article publish-ready and archive the tracker, the recording, and the approval emails together. That bundle is your answer if a quote is challenged after publication — and your starting point when a derived article reuses the material, which requires re-confirming, not assuming, clearance.
Mistakes and edge cases
- Treating "they agreed to the interview" as permission for everything. Consent to talk, consent to record, and consent to be quoted by name are three different agreements. The recording consent is also jurisdiction-sensitive — another reason to ask explicitly, on tape, at minute zero.
- Granting unbounded draft approval. Sending the full draft with "let me know your thoughts" converts a quote check into a negotiation over your article. Scope the review to their quotes in context, with a deadline and a default ("no reply by Friday means the quotes run as sent" — only where that default was agreed up front).
- Publishing an expert's error because the quote was accurate. Faithful transcription of a wrong fact is still a wrong fact on your page. Fact-checking applies inside quotation marks.
- Quoting from other people's published interviews as if you conducted them. Attribute to the original publication, keep the excerpt short, and remember that quoting published material is a copyright question — the U.S. Copyright Office's fair use materials outline the four factors courts weigh, and none of them is "it was only a sentence." When in doubt, link and paraphrase instead.
- Anonymizing so vaguely the quote loses its evidence value. "A senior engineer at a well-known company" carries no weight. Negotiate the most specific label the expert will accept — role, company size, industry — because the specificity is what makes the quote worth publishing.
- Losing the trail. Approval that lives in a deleted Slack thread does not exist. Approvals live with the tracker.
Claim ledger
| Claim | Source | Status | Review rule |
|---|---|---|---|
| Interview ground rules and attribution terms should be agreed in advance and quotes handled exactly as agreed. | Reuters Handbook interview guidance | verified on 2026-09-13 | Recheck within 180 days |
| Fair use of quoted published material is assessed on multiple statutory factors, not a fixed word count. | U.S. Copyright Office fair use materials | verified on 2026-09-13 | Recheck within 180 days |
| People-first content guidance favors demonstrable accuracy and trustworthiness in published claims. | Google Search Central helpful content guidance | verified on 2026-09-13 | Recheck within 90 days or if guidance changes |
| Recording-consent and defamation rules vary by jurisdiction. | General legal variation, flagged as out of scope | disclosed | This guide gives editorial workflow, not legal advice |
| The clearance tracker and approval-pass rules are our own artifacts. | Our own, dated 2026-09-13 | disclosed | Present as workflow advice only |
FAQ
Do I need permission to quote someone in a blog post?
For an interview you conducted: yes in practice — quote on the terms you agreed at the start of the call, and record those terms. For quoting someone's published writing or a talk, attribution plus a short excerpt is the norm and copyright's fair use factors apply; for anything commercial or adversarial, get legal advice rather than relying on a blog's rule of thumb — including this one.
Should I let interviewees approve their quotes before publishing?
In content marketing, usually yes — it is cheap goodwill and catches real errors. Bound it: quotes in context only, a reply-by date, and an agreed default if they go silent. Do not extend approval to your surrounding analysis.
Can I edit an interview quote for clarity?
Removing filler words and false starts without changing meaning is standard practice. Anything that alters substance, compresses two answers into one, or moves the quote to a different question's context requires either a paraphrase (no quotation marks) or the expert's sign-off on the new wording.
What if the expert says something factually wrong in a great quote?
Do not publish it as fact. Check the claim against a primary source; if it fails, either drop the quote, correct the frame around it, or go back to the expert — often they sharpen it into something accurate and still quotable. A published quote transfers the error to your page.
How do I attribute a quote when the expert cannot be named?
Agree on the most specific honest label the speaker accepts — role, company size, industry — and note the reason for anonymity in your tracker. If the only acceptable label is so vague the quote proves nothing, treat the material as background and make the point with sourced evidence instead.
Where this fits in the interview pipeline
Set permission defaults before the call using the expert interview content brief, and capture attribution agreements on tape using the expert interview run sheet. If one conversation feeds several pieces, re-clear quotes per article as described in the interview-to-article splitting map. Cleared material then flows into drafting through the voice-preservation drafting protocol, which enforces this workflow's quote boundaries inside AI-assisted assembly. The fact-checking half of this workflow is the interview-specific case of how to source claims in blog posts without making them boring, and the escalation rules belong in your content governance framework.
Sources and last-reviewed notes
Last reviewed: 2026-09-14. Sources checked: the Reuters Handbook guidance on interviews (ground rules, quotations, attribution), the U.S. Copyright Office fair use materials, and Google Search Central guidance on helpful, people-first content. The quote clearance tracker and all example rows are our own; the quotes shown in the tracker are illustrative phrasing, not real quotes from a real interview. This guide is editorial workflow, not legal advice.