A
Absent. An answer that doesn’t name your brand. See How an answer is read. Action. One piece of work meant to change the answers. See Actions. Answer. One engine’s reply to one tracked prompt. Every figure is built from answers. See How Heralded measures. Answers citing you. The share of answers that cite a page on your own site. See Answers citing you and sources mentioning you. Autonomy. The settings that decide how much Heralded does without asking you. See Autonomy.B
Brand. One website, with its own prompts, reports, competitors and connections. See Workspaces and brands. Brand access. Access to a single brand and nothing else in the workspace, as Brand manager, Brand member or Brand viewer. See Brand access. Brand-mention check. A Herald run that lists the pages on your site and your competitors’ sites and rereads the ones it watches for changes. See The team. Brand prompt. A prompt that names your brand. Left out of Mentioned, Recommended and the Heralded Score. See Which answers count.C
Cell. One prompt asked of one engine. A cell collects up to seven answers a week on a daily engine. See Prompts, engines and cells. Change. The difference in a figure between the first and last weekly report of the window, with its range and whether it is real. See Change. Channel. A social or content platform the engines name, such as YouTube or Reddit. Never ranked as a competitor. See Channels. Citation. An answer citing a page as a source. See Sources. Cited in. The share of answers that cite a site, or the answers that cite a page. See Sites and Pages. Claude add-on. Claude as a seventh engine for the whole workspace, answering weekly. See Claude. Command Center. The Heralded subscription that tracks your brands every day. See How Heralded works. Competitor. A brand you track to compare yourself with. The app also says rival. See Competitors. Competitor search. A Herald run that searches the web on the topics you track and adds new pages that mention them to your sources. See The team. Connection. A link to a service such as Search Console, WordPress or Slack. See Connections. Content credit. What one piece written in Content Studio costs. See Content credits. Content Studio. Where Heralded writes SEO and GEO pieces for you. See Content Studio.D
Daily update. The daily pass that asks every prompt of the engines due that day and records whether you are named. See Daily updates and weekly reports. Discovered brand. A brand the engines name that you don’t track. See Discovered brands.E
Engine. One AI product a buyer can ask, such as ChatGPT or Google AI Overviews. See Engines. Engineer. The team member who handles site fixes. See The team.G
GEO. Generative engine optimisation, writing so that AI engines cite you. See Content Studio.H
Herald. The team member who runs the daily update, the weekly report, the site check, the brand-mention check and the competitor search. See The team. Heralded Score. One number from 0 to 100 for how AI answers treat your brand, on a calibrated scale. See The Heralded Score.I
Identity. Your category, buyer, offer and what sets you apart, as you set them. Understood checks answers against it. See Settings. Intent. What a prompt is after: Problem, Educational, Comparison or Transactional. See Type and intent.K
Key claims. What you say about your category, buyer, offer and what sets you apart, with your facts. Each is checked against every answer that names you. See The page.L
Leading. An answer that ranks your brand as the single first choice. Counts toward Recommended. See How an answer is read.M
Managed account. A client account an agency manages through a management link. See Managed accounts. MCP. The Model Context Protocol, which AI assistants use to connect to tools such as Heralded. See Connect an AI assistant. Mentioned. As a figure, the share of answers that name your brand at any strength. As a rung, an answer that describes your brand without advising anyone to choose it. See Mentioned and Recommended.N
Needs your OK. What the chip on a task reads under the Ask first preset: Heralded prepares the work and waits for your approval. See Moving work along. Not measured. An answer the engine didn’t give, or that couldn’t be read. A gap, never a zero. See Which answers count.O
Organization. What agency plans call a workspace. The API and the assistant sign-in use the word on every plan. See Workspace.P
Perception. The page and the figures about what the engines say when they name your brand: Understood and Praised. See Perception. Period. The weekly reports a figure covers and a change compares with the reports before: the latest, the last four, the last twelve or all of them. Home reads the latest. See The period. Persona. The buyer the engines are asked as in a segment. See Segments. Praised. The share of answers naming you that frame your brand positively. See Perception. Prompt. One buyer question, tracked in one segment and asked of every engine. See Prompts. Prompt slot. What one active prompt in one segment uses of your plan. See Prompts. Publicist. The team member who handles outreach and correction requests. See The team.Q
Question type. How a suggested prompt is written: short, conversational, written as a buyer, a comparison or an alternatives question. See The Prompts page.R
Range. The 95% range around Mentioned, Recommended or the score, showing where it would land if the engines were asked the same prompts again. See Ranges. Raw score. The average of Mentioned and Recommended, before calibration turns it into the Heralded Score. See The raw score. Read. What the API calls a weekly report. The app says weekly report. See Figures in the API. Real. A change the paired test says is more than the engines’ usual variation. See Change. Recommended. As a figure, the share of answers on the Leading or Recommended rung. As a rung, an answer that tells the reader to choose, try or consider your brand. See Mentioned and Recommended. Roster. The competitors you track for a brand, up to 10. See Your roster. Rung. Where an answer lands for your brand: Leading, Recommended, Mentioned, Weak or Absent. See How an answer is read.S
Score band. One of five named ranges of the Heralded Score: Invisible, Under-visible, Emerging, Strong and Leader. See The score bands. Score version. The version of the calibrated scale a report was scored on. Change never spans two. See Score versions. Searches/mo. The monthly searches for a prompt’s topic, shown as an estimate such as “~1,000” and left empty when nobody has a number. See The Prompts page. Segment. A market, a language and a persona, read with their own prompts. See Segments. Share of voice. Your part of all the brand names in the answers. See Share of voice. Site check. A Herald run that checks the pages of your site and adds them to your inventory. See The team. Snapshot. A one-off report on one brand, which asks each prompt of each engine twice. See Snapshot. Source. A site the engines cite in their answers. See Sources. Sources mentioning you. The share of pages on other sites, cited in the answers and read by Heralded, that name you. See Answers citing you and sources mentioning you. Start. The button that begins a proposed task. The API calls its verbdispatch. See Proposed.
Strong. The score band from 50 to 69. Not the Recommended figure. See The score bands.
Suggested positioning line. A line from your latest weekly report that suggests how to position your brand. See The page.