How can we
help?

Sign-up walks you through it in one go: your account, the brand and its domain, the market you are measured in, the competitors you want to follow, and the plan. Your first numbers appear on the “Overview” screen once the first measurement has been analysed.

1. Create the account with your work email.
2. Enter the brand name and its domain.
3. Choose the market: the country and language your buyers ask in.
4. Confirm the competitors you want to follow. You can add more later on the “Competitors” screen.
5. Choose the plan and add a card in Stripe.

Every Monday at 07:00 UTC, every active question runs on every engine in your plan.

New numbers appear next to last week’s, alerts fire if something moved, and the weekly email summary goes out at the end of the day to the addresses listed in “Settings › Notifications”.

At the next Monday run.

A question you add or mark as naming your brand is measured from then on, and its past weeks are never recalculated. Competitors are the exception: archiving, rejecting or reclassifying one changes the ranking for past weeks too.

There is no button to run a measurement on demand: the run is weekly for every account, so the weeks stay comparable. New recommendations in the action plan also arrive with the next weekly analysis.

The trial is open on the Lite, Pro and Business plans and needs a work email. You add a card when you sign up, and nothing is charged until the trial ends.

The trial measures a shorter set of questions and engines, shown at sign-up; on the trial, answers are collected through the engines’ APIs, and reading in the apps starts with the paid plan.

When the trial ends, the plan you chose starts and the first invoice is issued. To stop before then, cancel in “Settings › Subscription”.

The menu follows four questions. The “Overview” screen is the week at a glance.

The “How am I doing?” group holds the “Competitors” and “Engines” screens. The “Why?” group holds “Questions”, “Sources”, “What AI says” and the other screens that explain a number. The “What do I do?” group holds “Plan”, your action plan.

The “More” group, which opens on a click, holds “Traffic”, “Reports”, “Studio” and “Topics and journey”. “Settings” sits at the bottom.

For a guided reading of each screen, open the avatar menu and choose “How to read”.

Phrase questions as a buyer would ask an assistant, not as keywords: “best invoicing software for freelancers” rather than “invoicing software”. Keep them stable so weeks stay comparable.

Add them one at a time or in bulk from a CSV. A new question is active straight away and is measured at the next Monday run. Each one counts against your plan’s question limit, except questions that name your brand.

1. Open the “Questions” screen and click “Add question”.
2. Write the question (8 to 500 characters). The category and the buying stage are suggested for you; change them if they are wrong.
3. Tick “This question names your brand” if it does, then save.
4. For many at once, use “Import or generate › Import CSV”, with the columns prompt_text, category, intent_stage, topic and branded. Up to 500 rows per file, each under 500 characters, or the file is refused; duplicates and rows under 8 characters are skipped, and rows beyond your limit are not added.

Not once it has been measured.

A question’s series is tied to what was asked: a reworded question is a different question, and comparing it with the old weeks would mislead you.

To change the wording, archive the question and add the new version, which starts its own series. That is why the wording is worth getting right before the first run.

Questions are never deleted for good, so their history stays.

The “Pause” button moves a question to the “Archived” tab and frees a place in your limit; the × on the row does the same, with a message to undo it straight away.

To measure it again, open the “Archived” tab and click “Reactivate”, which needs a free place in your limit.

Questions we suggest arrive in the “Proposed” tab and are not measured until you accept them.

The “Accept” button moves a question to the “Active” tab; the “Reject” button archives it. The “Accept the N in this list” button accepts every question in the filtered list, up to the places left in your limit.

A question that names your brand cites you almost by construction, so counting it would inflate your citation rate.

Marked as naming your brand, it stays out of the citation rate, share of voice, position, topics and your question limit, and still counts for sentiment and sources.

Set it when adding the question, or in the question’s detail; the change applies from the next run.

Competitors you add count against your plan’s competitor limit; brands the engines bring up by themselves are listed too, and take no place in it.

Only direct competitors enter your share of voice. Confirmed competitors also get a weekly check of their moves and prices.

1. Open the “Competitors” screen and click “Add competitor”.
2. Enter the name, the domain (it sets the logo) and the classification. The default is direct competitor.
3. To change a classification, use the selector on the row. The ✓ button confirms a competitor; the ✕ button rejects one, with a message to undo it.
4. If the limit is reached, archive one or upgrade in “Settings › Subscription”.

Click “Archive” on the competitor’s row and confirm: the competitor is no longer tracked and leaves the list. The ranking is recalculated from the current list, so past weeks change too.

A message lets you undo straight away; after that there is no list of archived competitors, so write to us to bring one back.

A persona changes who is asking: the engine receives the question framed as that buyer, while the question’s text stays the same. The pricing page shows which plans include personas.

Our team defines them with you; to use one, open a question’s detail and choose it under “Who is asking”. To create or change a persona, write to us.

An extra market is set up by our team as a measurement of its own, with its country, language and competitors.

The pricing page shows which plans include more than one market and where it is available as an add-on. Write to us with the country and the language you want measured.

Write to us: the measurement language and market are set by our team, because changing them changes the next measurement and breaks comparison with past weeks.

The “Language” option in the avatar menu only changes the interface you see; reports and alerts follow the brand’s language.

A mention means your brand appears in the answer. A recommendation means the engine picks you. We track both, plus position, separately.

Engines answer differently inside their apps. Measured on Claude in August 2026, only 24% of the brands named overlapped between the API and the app.

That is why, on paid plans, the web-search answers of up to seven engines (the ones in your plan) are read in the app, in a live session.

How many answers it was measured on. A ratio over five answers moves for reasons that are not about your brand, and the product always shows the denominator next to the number.

Four weeks of comparable data. The first measurement is a photograph, not a trend, and reading it as one is the most common mistake in the first month.

The “Plan” screen lists what to do next, split into “To do”, “Waiting on you”, “Done” and “Dismissed”. It is rewritten with each weekly analysis, so there is no button to generate it.

Click “Mark as done” and the date is recorded, which is what lets the product read the effect before and after. The “Dismiss” button removes an action, with a message to undo it, and the “Reopen” button brings it back.

You can also send playbooks to the plan from the “How to grow in each” tab of the “Engines” screen, with the “Add to plan” button.

Critical and notable changes appear in a banner at the top of every screen until someone who can edit marks them as read, one by one or with the “Mark all as read” button.

The thresholds that trigger an alert are set per brand by our team; write to us if you want them tighter or looser.

The “Answers” tab of the “Questions” screen lists every answer the engines gave to your questions in one week, one row per answer, ordered by question and then by engine.

The filters for engine, sentiment, citation, position and week combine, and the week filter goes back 12 weeks. Other screens open this tab already filtered, for example on one engine.

Open a row to read the full text, the competitors named and the sources the engine cited, with your own site flagged. The “Open the question” link goes to the question’s detail.

Everyone on the account can read it, viewers included. It changes once a week, after the Monday run; there is no export of the answers.

The “What changed” tab of the “Questions” screen compares the latest measured week with the one before, question by question, in four columns: “Lost”, “Newly cited”, “Solid” and “Never cited”.

Each column shows three cards, and the “See all N” button shows the rest.

On “Lost” and “Newly cited” cards, the engines that changed; on “Solid” ones, how many weeks in a row the question has been cited; on “Never cited” ones, when it was last cited.

Outside “Solid”, hovering a card shows the competitors in those answers. A click opens the question.

In the first measured week there is nothing to compare, and the tab says so. Streaks look back six weeks. Everyone on the account can read it.

The “Prompts x sources” screen, in the “Why?” group, shows for each question which domains the engines cited when they answered with web search, and the path for your brand to get into that answer. It is included in every plan.

There are seven paths: “You're in”, “Answered from memory”, “Rewrite page”, “Entity”, “Community”, “PR and lists” and “Missing page”. The chips at the top filter the table by path.

The screen reads only the latest week and only answers with web search, so it stays empty until the first measurement with search is done. A competitor’s page is never proposed as a target.

1. Open “Prompts x sources” and, if you want, choose a path in the chips at the top.
2. Click a question to open its detail: the path, “What to do”, the table “Where the answer comes from” and the brands named.
3. Click “Add to plan” to send up to three of the steps to “Plan”, tagged with the question. Only an “Owner” or an editor can do this; viewers only read.

Before answering with web search, an engine splits your question into searches of its own. The “Query fan-out” screen, in the “Why?” group, shows those searches for the latest week: how many there were, how many per answer, how many name your brand and how many domains were read.

The table lists each search with its engine and how often it came up. The “What it always goes looking for” section looks for patterns over eight weeks, and the readings by question link to “Plan”.

For engines read in their app, the screen shows the subject of the search rather than its exact words. An empty field for an engine means we do not collect its searches yet, not that it did not search. The screen is read-only.

The “Engines” screen, in the “How am I doing?” group, opens on the “Rate by engine” tab: your citation rate on each engine in the latest measured week.

The “By engine” table has one row per engine and mode (with or without search, read in the app or through the API), with rate, position and number of answers. A row opens that engine’s detail: “Where this engine draws its answers from”, “How to grow here” and “This week's answers”.

The “Cards by engine” section shows an eight-week trend and the competitors that appear most, with “All” or “Direct”. The “see answers →” link opens the “Answers” tab filtered on that engine.

An “Owner” or an editor can drag the cards to change their order. Only the engines in your plan appear. To measure another engine, see the article on getting more questions, competitors or engines.

The “How to grow in each” tab of the “Engines” screen answers, engine by engine, where you stand, through which route you are cited, where you can get in and what to do.

At the top, “Sources engines cite most” and “Where, inside the platforms” show the sites, subreddits and channels the engines lean on.

Then each engine has its own section: the change against about four weeks earlier, the open plan actions aimed at it under “What to do now”, and the playbook “How to appear here”.

Next to each playbook item, the “Add to plan” button creates an action in “Plan”. Only an “Owner” or an editor can use it.

The “Topics and journey” screen, in the “More” group, splits your citation rate by question category and by stage of the buying journey.

The “Topics” table gives, for each category, the questions, the rate, the leading brand and the change over four weeks, with the state “Ahead” (15% or more), “Competitive” (5% to 15%) or “Behind” (under 5%). A row opens that category in “Questions”.

The “Funnel” reads the same answers by stage: “Learn (TOFU)”, “Consider (MOFU)” and “Buy (BOFU)”.

The topic is the question’s category and the stage is where it sits in the buying journey, both chosen when the question is added. Questions that name your brand do not count. The screen is read-only.

The “What AI says” screen, in the “Why?” group, shows how the engines describe your brand, in three tabs.

The “Brand portrait” tab sums up what AI understands your brand to be, with three panels: “What AI never says”, “What they say against you” and “Differs from the facts”. The last one compares the answers with your brand profile, so it stays empty while the profile is empty.

The “Sentiment and claims” tab counts sentiment only in answers that cite you, and lists the negatives one by one. The “Against competitors” tab places each brand on a map of presence and narrative.

The text comes from the weekly analysis of the Monday run. Everyone on the account can read it; the brand profile is filled in “Settings › Client”.

The “Sources” screen, in the “Why?” group, lists the domains the engines cite in your category over the last eight measured weeks, and whether your brand appears next to them. Only answers with web search return sources.

The “Sources · Authority map” table shows the most cited domains and their relation to you: “Your own”, “Social”, “Competitor”, “Adjacent” or “Editorial or other”. The relation comes from your domain and your competitor list, so a competitor with no domain is not recognised.

“Authority targets” are editorial or social domains cited without you. The “see analysis →” link opens one domain: the pages cited and “Which questions it appears in”.

The screen is read-only and open to everyone on the account.

The “Territory” screen, in the “Why?” group, classifies each question of the latest week by who wins it: “Open territory” (no brand is recommended), “Contested, no leader”, “Owned” (a competitor dominates) or “Your territory”.

The “Territory by question” list marks open questions at the bottom of the funnel as “gold”. Another table shows the category’s sources, who dominates them and whether you have a page there; that presence is checked in the weekly collection.

On an open or contested question, the “Attack” button asks for a content brief, which the analysis writes into “Plan”.

Until then the row says the brief is being prepared; once it is ready, the row links to it and offers “Attack again”. Only an “Owner” or an editor can use it.

The “Community and social” screen shows what is said about your brand in forums, social networks and reviews, and whether an AI engine cited that thread.

The “Community sources” cards show what each source brought in over 30 days, with the “cited by AI” tag when an engine cited it; a click filters the list.

Under “What was said”, the filters “All”, “Cited by AI” and “About the brand” filter the mentions. A new mention reads “Not yet analysed” until the weekly analysis.

Further down are reputation themes, “Improvements your customers are pointing at” and “Everything that was said”.

Collection is weekly. LinkedIn, X, Instagram, YouTube and TikTok are searched for every brand; the forums and subreddits are confirmed one by one by our team, so to add or remove a source, write to us.

The screen is a module our team switches on per brand. Without it, the item is not in the menu; if it fits your brand, write to us.

The “Press” screen gathers the press pieces that mention your brand, collected every day, with the pieces cited by AI counted apart.

The “What was published” table gives the outlet, the headline, the tone, whether it is about the brand or names it only “in passing”, and the date. A piece republished elsewhere sits on one row, with the other outlets under “Also in”. The “In the press and on podcasts” section adds the podcasts.

The search terms are set by our team; to change them, write to us. The screen depends on the “Community and social” module: without it, the screen only says the module is off.

The “Products in AI” screen shows what AI answers when someone asks which product to buy: the “Recommended products” and how often each came up, the “Named shops”, and under “Who decides”, the sources behind each recommendation.

It is fed by observation runs inside the apps, not by the Monday run, and each figure is a point observation. The screen needs your product catalogue: without it, it only says the catalogue is missing. In that case, write to us.

The screen is a module our team switches on per brand, for brands that sell products. Without it, the item is not in the menu; if it fits your brand, write to us.

The “Site and entity” screen, in the “More” group, opens on the “Site health” tab. It looks at your site in two parts: the foundation, which exists once per site (robots.txt, llms.txt, sitemap, the homepage Schema, Wikidata and Wikipedia), and each page, with a technical score and an AEO score.

The tab has the “Checks”, “Pages”, “What to do” and “Signals outside the score” sections. Click a page and then “Open the page” to see its detail: “What to do on this page”, the issues found and the “Fix text”, which the “Copy the fix text” button copies.

The site is read again every Monday. The tab needs the brand’s domain, which our team sets.

1. Open “Site and entity” in the “More” group.
2. To read the site again without waiting for Monday, click “Scan pages”; to read only the foundation, click “Scan foundation” in the “Checks” section. Only an “Owner” or an editor can use them.
3. Click a page in the list, then “Open the page”, to see what to fix on it.

The “Entity” tab of the “Site and entity” screen answers whether AI assistants know who your brand is. It shows where the brand exists (Wikidata, Wikipedia, the profiles linked from your site, Google Business and sector directories) and what is missing.

When we are not sure of a profile, or found none, the tab asks you, with “Yes” and “No” buttons. An “Owner” or an editor can answer; viewers see the question only.

The entity is checked by our team and again in the weekly collection. If the tab says the check has not run yet for your brand, write to us.

The agent answers questions about your own data, written in plain words: the latest measured week, site health and your active competitors. The answer is written as you wait.

Open it from the “Agent” item, in the “Ask” group, or from any screen with the “Ask the agent” button at the top (Ctrl K, or ⌘K on a Mac), which opens a side panel. Past conversations stay on the “Agent” screen.

Only an “Owner” or an editor can ask; a viewer who asks gets the read-only account message. There is a ceiling on questions over the last seven days, and the agent says when it is reached.

The pricing page shows which plans include the agent. If yours does not, the menu item shows the name of the plan that does, and an “Owner” can upgrade in “Settings › Subscription”; on agency plans, the item is not shown.

1. Click “Ask the agent”, or open the “Agent” screen and click “New conversation”.
2. Write the question, or choose one of the suggestions, and click “Ask”.
3. Under the answer, “What to do next” lists the actions it suggests. The “Add to the action plan” button sends them to “Plan”.

On the “Inception” screen, in the “More” group, you declare what you want AI answers to say about you, as value propositions that cover several questions, and you follow whether the engines start saying it.

Each proposition shows how many questions it covers and in how many answers AI says it. The “Declared targets” table gives each question’s target and its state: “Reached”, “Not yet” or “Not measured yet”.

A target is set in the question’s detail, under “What you want to change in this answer”, with the “Set target” button.

Only an “Owner” or an editor can edit; viewers can only read. The pricing page shows which plans include Inception; if yours does not, the menu item shows the name of the plan that does.

1. Click “New value proposition”, fill in the short name and what you want the answer to say, and, if you have them, the subject and the address where it is publicly proven. Then click “Create proposition”.
2. Open the proposition and add questions with “Apply to every question containing” and “Apply”, or with “Pick questions”.
3. Click “Request the work plan”. The plan arrives in “Plan”, with the proposition on each action, and is listed under “The work this produced”.
4. The “Deactivate” button pauses a proposition and “Delete” removes it, after confirmation.

The “Questions › Import or generate › Export CSV” option gives the questions in every state, and the “Export CSV” button on the “Competitors” screen gives the active competitors.

On the “Reports” screen, the “PDF for the month” button gives the monthly report, and invoices download as PDF in “Settings › Subscription”.

There is no single export of everything; if you need data the screens do not export, write to us.

Anyone on the account, including viewers, can download the “PDF for the month” on the “Reports” screen.

A report link for people outside the account is published by our team; once it exists, the “Copy link” button on the “Reports” screen copies it and shows when it expires and how often it was opened.

To renew or revoke a link, write to us.

The weekly email summary goes out on Monday at 16:00 UTC, to a list of up to 10 addresses per brand. Being a member of the account does not add you to it.

In “Settings › Notifications”, type an address and click “Add”; the “Remove” button takes one out, with a message to undo it. With no address on the list, nobody receives it.

Connecting brings your search and site traffic next to your AI visibility.

The property is chosen automatically: the one that matches your brand’s domain, or the first one in the Google account. The “Disconnect” button asks for confirmation and keeps the data already collected.

Viewers cannot connect; a service marked as unavailable is enabled by our team on request.

1. Open “Settings › Integrations”.
2. Click “Connect Google” for Search Console, or “Connect Google Analytics”.
3. Sign in with the Google account that has access to the property, and accept.
4. Back in Periscopy, the “Sync now” button fetches the data straight away; after that, the data syncs every day. If the connection expires, the “Reconnect” button restores it.

Periscopy has an MCP endpoint, so an assistant can read your week, your action plan and your competitors.

There are no API keys: when the assistant connects, you sign in to Periscopy and authorise the assistant. Access is read-only, follows your own role, and lasts 30 days, after which you authorise again.

To revoke access sooner, write to us. The developer docs have the details.

The “Traffic” screen shows the AI crawlers that read your site and the visits that arrive from assistants.

To start collecting, open “Traffic”, then “How to connect”, and pick one of three routes: a log drain from Vercel or Cloudflare, a short snippet in your site’s middleware, or the WordPress route.

The connection token is visible to editors and owners; to rotate it, write to us.

On the “Reports” screen, the “Week by week” tab lists every measured week. A row opens that week’s page: its state, the citation rate, share of voice, share of recommendation, average position and the number of answers that cite you.

When the weekly analysis has written one, the “Editorial summary” explains the week. The “Open this week's N answers” button opens the “Answers” tab filtered on that week.

Everyone on the account can open past weeks.

There are three possible reasons.

If the menu item shows a plan name next to it, such as Pro or Business, your plan does not include it, and the owner can upgrade in “Settings › Subscription”.

The “Products in AI” and “Community and social” modules depend on the kind of business and are switched on per brand by our team; without the “Community and social” module, the “Press” item opens a screen saying the module is off.

An integration marked as unavailable is also enabled by us. In the last two cases, write to us.

The “Settings › Subscription” tab shows your usage against your plan’s limits. Questions that name your brand, and competitors the engines found by themselves, do not count.

To get more, you can upgrade, which applies straight away, or add an add-on to your current plan: more questions, an extra engine, an extra market, three runs per question, or personas on Lite.

The pricing page lists each add-on and the plans it applies to.

Add-ons are added to your subscription by our team, not from inside the product.

Write to hello@periscopy.ai with the add-on you want; the account must be on a paid plan, not on the trial.

The add-on is charged in proportion to what is left of the current period. Once an account has an add-on, plan changes also go through us.

Only an “Owner” sees “Settings › Subscription”.

An upgrade applies now: you pay the difference until renewal, and the next Monday run uses the new plan.

A downgrade applies at the end of the paid period and nothing changes today.

If you have more questions than the new plan allows, you choose which stay; otherwise the oldest stay, and the rest go back to the “Proposed” tab, with nothing deleted. Over the competitor limit, the most cited ones stay.

You get an email three days before, and the “Cancel downgrade” button undoes it until then.

1. Open “Settings › Subscription”.
2. Under “Change plan”, choose the plan.
3. For a downgrade, confirm first; then, until the date, the “Choose prompts” button in the scheduled downgrade notice lets you pick which questions stay.
4. Accounts with an add-on, on an agency plan, or whose plan is managed by our team change plan through us.

There is no notice period and no minimum commitment on monthly billing.

The “Cancel subscription” button, in “Settings › Subscription”, asks for a reason and ends the plan at the end of the paid period; until then, the “Resume subscription” button restores it without losing anything.

After the end, the Monday runs, the weekly email and the alerts stop, and other members lose access.

Your data is kept for three months, during which the owner can resume with a new payment; after that the account and its data are deleted, and you are emailed 30 and 7 days before.

The “Switch to yearly” button, in “Settings › Subscription”, moves you to yearly billing with the discount shown there. It starts at your next renewal and nothing is charged today.

Yearly billing is a 12-month commitment; you can cancel the following renewal at any time. It cannot be combined with a scheduled downgrade.

To go back to monthly, write to us.

Invoices are listed in “Settings › Subscription”, each with a PDF. Prices exclude VAT, which Stripe calculates for your country.

The VAT number is optional; without one the invoice is issued to a final consumer, and the “Edit” button, under the billing details, changes the name, VAT number and address.

The “Change card” button opens Stripe, so your card is never stored on our servers.

Enterprise is on request: the form in “Settings › Subscription” sends us what you need, and your current plan does not change until you accept a proposal.

Agency plans have no “Subscription” tab; changes, more brands and white label are arranged with our team.

There is a monthly allowance of texts per brand in “Studio”. When it runs out, “Studio” says so and shows the date it resets, the first day of the next month. Texts already written stay in the history.

Two add-ons are services from our team rather than features in the product: the “Human monthly report”, written by us every month, and the “45-minute read-through session”, a one-off purchase you can repeat as often as you like.

Neither changes anything in your account, and neither replaces the automatic “PDF for the month” on the “Reports” screen. The pricing page lists the add-ons on offer.

To get one, write to hello@periscopy.ai; as with every add-on, the account must be on a paid plan, not on the trial.

There is no limit on the number of people.

The account owner invites them; anyone else sees a note saying only an account owner can invite. The invitation goes by email in the brand’s language, and the new member chooses a password on first sign-in.

1. Open “Settings › Account and appearance”.
2. Under “Invite someone”, enter the email.
3. Choose “Viewer” or “Owner”, then click “Invite”.
4. If the email already has a Periscopy account, write to us and we give it access to this brand.

A “Viewer” sees everything and changes nothing: any change returns “Read-only account, no permission to change this.”

An “Owner” also changes questions, competitors, settings and integrations, and manages the subscription.

An editor, set by our team, changes the same things as an “Owner” except the subscription.

Inviting people is for the account owner.

Write to us: removing a member, changing a role or resending an invitation are done by our team. The team list in “Settings › Account and appearance”, visible to users with the “Owner” role, shows who has access and when each person last signed in.

Sign in with your email and password; the “Forgot?” option sends a link to choose a new one. Where the sign-in page offers it, Google, Microsoft or Slack sign-in works for an email that already has access.

To change your password, go to “Settings › Account and appearance › Change password”; it needs at least 8 characters.

Single sign-on (SAML) and two-factor authentication are not available.

Both live in “Settings › Client”.

The “Brand profile” section holds the facts about the brand (description, founding date, address, social profiles). It saves when you leave a field, and emptying a field deletes that fact. Addresses must be full, starting with https://.

The profile feeds the identity file in “Plan › Ready to use” and the check of what AI claims about you.

The “Brand kit” section holds your voice, what you sell and what is never said, and “Studio” writes from it; click “Save” after editing.

The brand name and domain are changed by our team.

If you have access to more than one brand, the brand selector at the top of the sidebar moves between them. Each brand keeps its own questions, competitors, settings and weekly email. New brands are added by our team.

Only an “Owner” can do it, and it cannot be undone.

Deletion is immediate: an active subscription ends today, with no refund for the current period. It deletes this brand and its data; the account goes when it has no brands left, and a login is removed only if it has no other access. Invoices stay with Stripe.

If you only want to stop paying, cancel instead: your data stays for three months.

1. Open “Settings › Subscription”.
2. Click “Delete account and data”.
3. Type the brand name exactly as shown.
4. Click “Delete forever”.