For multi-brand companies
Twenty brands. One set of rules. A record of every word.
Herald reads every comment and DM across your brands, routes the risky ones to the right owners, and keeps an audit trail legal can read. Nothing is sent on its own until a brand switches it on.
What keeps this team up
One brand’s crisis, every brand’s problem
A product recall on one brand page spills into the others within an hour. Possible crises rise to the top for the owners of that brand, with an email, and are never answered on their own.
Who said that, and who let them?
When legal asks who replied to a complaint and what they were told, the answer is a row in a log that cannot be edited: who read it, who approved the words, when it went out.
Regional teams, global rules
Each brand sets its own rules within workspace defaults, each person sees only the brands they hold, and every policy change is a new version with a note, one click from restored.
What lands in the inbox
An hour across a household-goods group: a cleaning brand, a baby-care line and a pet-food label.
Hannah K.Comment thread · Facebook
My toddler swallowed some of the dishwasher gel. Is it toxic?? What do I do
Possible crisisCritical95%Escalate, no draftNever automated: always a personMarcus T.Comment thread · Instagram
Is the new formula still free of phosphates?
Draft: Yes, Marcus: the new formula is phosphate-free, like the old one. The full ingredient list is on the product page under “What’s inside”.
Content questionMedium90%Draft ahead for reviewIlse V.Direct message · Facebook
Ordered the large bag twice and was charged three times. Order 88213.
Draft: I’m sorry, Ilse, that’s not right. I’ve passed order 88213 to our support team with your note so they can check the extra charge and put it right. They’ll reply here.
Order, shipping or accountMedium93%Draft ahead for reviewTomásComment thread · Instagram
Switched our cat to the salmon recipe and she finally cleans her bowl. Thank you!
AppreciationLow96%Review, draft on requestAutomatic replies are off for this brandLaw Office of R. AmesDirect message · Facebook
We represent a customer regarding an injury from your product. Please direct us to your legal department.
Legal / press / partnershipHigh97%Escalate, no draftAlways goes to a person (locked category)deals_bot_4421Comment thread · Instagram
90% OFF all brands today only >> cheap-brand-outlet.shop
Hostile / troll / spamMedium98%HidePolicy: hide hostile and spam comments
Invented sample messages. What happens to each is worked out by Herald's own policy engine under this team's starter rules.
One message, start to finish
Rules they turn on first
Governance first: spam hidden by rule, product questions and pushback kept for each brand’s owners, and automatic replies left off until the evidence is in.
Automatic hiding
OffOn
Content question: owners review
OffOn
Disagreement / critique: owners review
OffOn
What never happens on its own
However the rules are set, Herald never answers these on its own:
- Hostile / troll / spamInsults, trolling, scams or bot spam. Never answered; Herald suggests hiding it.
- Possible crisisAbuse, self-harm, loss or crisis. Always goes to a person with crisis resources; never automated.
- Legal / press / partnershipPress, legal, sponsorship or takedown requests. Goes to the Owner; never drafted.
Every reading, decision, draft, send and hide is kept in an audit log that cannot be edited, and can be exported.
Automatic replies stay off until the brand switches them on, after Herald shows what they would have sent.
What a month costs at this size
20 brands, 60 connected accounts, 15 people, 60,000 messages a month
$582.50 a month, billed yearly
Scale plan
People’s time to answer them well by hand$134,270 (2,042 hours)
People’s time with Herald$47,194 (709 hours)
Saved after the plan’s price$87,076
Estimates from the same method and sources as the comparison page; every feature, brand and teammate is included. See the working
Where it connects
Works with
- YouTube· Coming soon
- X· Coming soon
- Your own app
Want next?
Herald answers comments and DMs. It does not publish posts: keep the scheduler you use.
What the research says
3.5%
of comments on brands’ and organisations’ posts got a reply in 2025.
Respondology, Business of Comments Report, 2026Platform data. 168.8 million comments on 7.6 million posts from about 3,400 accounts on seven platforms. Data from a company that sells comment moderation. 73%
of consumers say they expect a brand to respond on social within 24 hours or sooner.
Sprout Social Index, 2025Industry survey. 4,044 consumers in the US, UK, Canada and Australia. 55%
of consumers say they would rather message a company privately than complain publicly.
Sprout Social Pulse Survey, 2026Industry survey. 2,250 consumers, August 2026.
Questions this team asks
Can a regional team see only its own brands?
Yes. Each person holds a role (Owner, Admin, Moderator, Reviewer, Viewer) and a list of brands. Their inbox, drafts, policies and audit rows are only those brands'; the database itself refuses rows from another workspace.
What happens to a possible crisis?
Self-harm, abuse, a child in danger or a legal threat goes to the top of the inbox for the brand’s owners, with an email to each of them. It is never answered automatically, and no draft is written.
Can we show legal who said what, and when?
Every reading, decision, draft, edit, send and hide is a row with the person or system that did it and the time. Rows cannot be edited or deleted by the app, and the log exports to CSV. Herald keeps a record; it does not make anyone compliant on its own.
Do policies change without anyone noticing?
Every change to a brand’s rules is a new version with a note and the person who saved it. The history shows what each version changed, and any earlier version is one click from restored.
Where do the AI calls go, and is our data trained on?
Readings and drafts come from Anthropic's Claude through its API, which does not train on your data; the privacy policy lists every sub-processor. Message text is treated as data, never as instructions, and network credentials are encrypted per workspace.
Can we switch automation off everywhere at once?
Yes. A workspace kill switch stops every automatic reply and hide in one click; each channel has its own. Reading and sorting carry on, so nothing is missed while it is off.
Bring every brand under one set of rules
We are opening a few multi-brand workspaces at a time, with a call to map your brands and roles first.