The Content Advisor

The Content Advisor is an AI assistant that lives inside Nolorem. It knows your content, your usage stats, the documents in your media library, and the page you are on, so it gives relevant, actionable answers without leaving the platform. It is available to Pro, Premium, and Trial subscribers.

Opening It

Click the sparkle button in the bottom-right corner of any platform page to open a slide-over chat panel. You can also press Ctrl+Shift+H to toggle the panel.

Answering With Clicks, Voice, or a File

The message box does more than take typed text.

Choice pills. When the advisor offers you a few options, they appear as pills you can click. If the advisor offers a single choice and your message box is empty, a click sends that choice at once. If you have already typed something, attached a file, or the question allows more than one answer, a click selects the pill instead, so you can pick several, add a sentence of your own and send it all together. A line under the message box says what is ready to send. Click a selected pill again to drop it.

Dictation. The microphone button next to the send button turns your speech into text. What you say lands in the message box, where you can read and edit it before sending, so a misheard word never goes out on its own. It follows your interface language, so a Dutch interface listens for Dutch. The button only appears in browsers that support this; where it is missing, your browser does not offer it.

Attaching a document. The paperclip adds a document from your computer to the conversation. It is uploaded to your media library, and the advisor reads its text when you refer to it, so you can ask questions about a brief, a report, or a transcript without pasting it in. Mention the file by name in your message and the advisor will look it up.

Two things worth knowing. An image can be attached, but the advisor receives its name, not the picture itself: it reads documents, it does not look at images. And having the advisor read your own documents is part of the Pro plan.

Context-Aware Assistance

The advisor detects which part of the platform you are on and adjusts its focus:

Current page	Assistant focus
Blog pages	Topics, keywords, posts
Social pages	Channels, scheduling
Settings	Configuration, billing
Dashboard	Stats, activity
Media library	Images, presets
The post list, the calendar and the blog settings	The blog you are working in
An open post in the editor	That post: its title and its sections

When you have a post open in the editor, the advisor knows which one it is. Ask it what you have open and it names the post by its title, and it can quote the section headings from inside it. It reads that post; it does not write it for you.

On the post list, the calendar and the blog settings the advisor knows which blog you mean. The blog named in the address wins, otherwise the blog selected in the blog switcher, otherwise your default blog. It only looks in the blogs of your own organisation. The panel subtitle shows where you are, and each of these three screens has its own starter questions.

Knowing Which Screen You Are On

The advisor does not only know which part of the platform you are in, it also knows which screen you have open. On an open blog post or social post it knows which one. On a campaign it knows the campaign, and when you have one of its items open, that item. On the post list it knows which posts you have ticked, up to 20. Before it uses any of this, Nolorem checks it against your own organisation: a post or campaign that is not yours, or that has been deleted, is ignored, as if you had opened nothing.

If you have more than one blog, the advisor asks which blog you mean before it does something that belongs to one blog. It asks this on the dashboard and on screens that belong to no particular blog; on a screen that names a blog, such as the post list or the blog settings, it does not ask. If you only have one blog, it never asks. On a screen that belongs to no particular blog it may also offer the post or campaign you worked on last and ask you to confirm it, so you do not have to explain which one you mean.

The advisor is meant to name a button the way it reads on your screen, in your interface language. Nolorem does not block a wrong button name automatically yet. If the advisor names a button you cannot find, tell it which screen you are on and what you want to do.

Asking About Your Own Documents

Ask the advisor about a document you uploaded to the media library and it answers from what is in that document, not from a general impression of your business.

You do not select anything first. Name the file or the subject and it finds the right one: "What does the annual report say about our costs?" works, and so does "What do you know about our sales figures?" when the numbers sit in a spreadsheet whose name you have forgotten. The document it used is named in the answer, and that name is a link, so one click takes you to the document's own row in the media library.

It can draw on several documents in one answer, up to a limit set by your plan: 3 documents per question on Pro, 6 on Premium, and 2 during a trial. Basic does not include the Content Advisor, so it has no document answers either.

When it finds nothing it says which of three things happened, instead of filling the gap with something that merely sounds right:

• You have not uploaded any documents yet.
• You have documents, but none of them bear on what you asked.
• The document you named holds no readable text. That is usually a scanned PDF with no text layer, and it is the one case you can fix yourself: run the file through an OCR tool and upload it again.

The Use for post suggestions switch does not silence it. That switch decides whether a document is used to propose posts to you on its own. Turn it off and the advisor still reads the document when you ask for it by name. The two are deliberately separate: not wanting a document to generate ideas is not the same as not wanting to ask about it.

Objectives That Name Their Source

When you ask the advisor to set up a campaign from something in a document, the objective it proposes carries the document it came from and the sentence it took it from. You see both on the card before you apply it, and they stay on the campaign afterwards.

The quote is checked against the document before anything is saved. If a sentence does not actually appear in the document it claims, the objective is still kept but the source is dropped, so an attribution you read is one that held up. Delete the document later and the objective keeps its quote, with a line saying the document is no longer in your library.

Documents in a Campaign Conversation

When you set up a campaign with the advisor, it reads only the documents chosen for that campaign and any document you ask for by name. It asks once whether you want to use documents, and it never decides by itself that a document is the plan. If you say no, it works without documents until you name one.

Setting Up and Adjusting a Campaign

On a campaign page the advisor proposes the audience, the channels and the cadence, and applies them when you confirm. Changing the plan of a campaign works through it as well: the campaign page no longer has a separate Adjust field, so you ask the advisor to change the plan, or you change a post directly on its card. When you ask the advisor to adjust the plan, a short block on the campaign page counts what actually changed. When you ask for a whole new plan, you are asked to confirm first, because it replaces the one on screen.

The advisor follows what Nolorem finds about the campaign template. When the goals of a campaign no longer match its template, it does not present the template as a fact and does not hold the campaign to the template's length. See Communication Campaigns.

Having Existing Posts Changed

You can ask the advisor to change posts that are already in a campaign plan, in plain words: "Make all the titles shorter", "Move the Friday posts to Thursday" or "Mention the new opening hours in the posts about the sale". It proposes the posts it finds for your request and shows a card with, for each post, what it says now, what it would become and why. It can miss a post that words the subject differently than you did. You see the proposal before anything changes, so check that every post you meant is on it. If one is missing, ask again with the word that appears in that post.

• Everything is ticked, and you untick what should stay. Nothing changes before you click Apply.
• Some posts are listed as not possible, with the reason: the post is locked, it is already published, it is being written right now, or the new date falls outside the campaign period. Changing a date also requires publishing rights.
• A post that is already written takes over the new title, but its text stays as it is, unless you tick Rewrite on that row. Then the post is written again with the new briefing.
• A blog post that moves to a later date takes its promotion posts along, so a promotion never comes before its blog post. This also goes for a promotion post that is locked, and the card says so.
• A post that changed in the meantime is not overwritten. If you or someone else edited it after the card was made, it is reported as changed in the meantime and left as it is.
• If nothing in the plan is about your request, the advisor says so and nothing changes.

When you change a default setting after the plan was made, the advisor answers in a fixed way. If you change the documents, the Voice DNA or the writing tone, it offers to adjust the posts that already exist to it. If you accept, the card proposes only the posts the new setting really affects, and if none is affected it says that the new setting asks for no change. If you change only the post length, there is no card: the advisor says that new posts follow the new length, and that posts already written keep their length and can be rewritten one by one. It does not make the whole plan again. A locked post stays as it is.

What the Assistant Knows About Your Account

When you ask, the assistant looks things up in your own account instead of guessing. It can read:

• Your plan. Whether you are on a trial, have a paid subscription, have complimentary access or have paused, and how much you have used this month. When there is no active subscription it says so.
• Your campaigns. Running, paused and draft campaigns, and finished ones, plus the campaigns in the trash. Finished and deleted campaigns it only reads; it never changes them. It sees a campaign's period, its goals and its posts.
• Your posts. It can search your blog posts and social posts and open one to read it.
• Your company profile and strategy. The company profile of the blog you are working in, and the goals, audiences and pillars of your strategy.
• The screen you are on. If you have a campaign or a post open, the assistant knows which one.

It only reads your own organisation. If a lookup fails, it tells you it could not look that up instead of saying there is nothing there. While Nolorem support is viewing your account on your behalf, the assistant cannot look things up.

Changing a Campaign Through the Assistant

On a campaign, ask the assistant in your own words ("make the campaign two weeks longer", "post on LinkedIn too", "pause it"). It can propose changes to the briefing (it adds a goal), the period, the rhythm, the channels, the audiences and the content defaults. It can add posts to the plan or remove them, pause or resume the campaign, and start the writing of posts you name.

It never changes anything straight away. It answers with a card called Proposed change that shows what is true now next to what would change, and nothing happens until you press Apply. If you do not press it, your campaign stays exactly as it was.

• A change your role does not allow. Usually the assistant tells you in its answer that your role does not allow it and which roles can, and there is no card. If a part that cannot be done does end up in a proposal, the card marks it Not possible, with the roles that can do it and your own role. As long as such a part is on the card, you cannot apply the card: ask the assistant for a proposal without that part, or ask someone with that role. Pausing, resuming and changing the audiences need an Admin or an Editor, and removing a post needs the right to delete.
• Other reasons appear there too: the campaign is generating right now, its status does not allow the change, the part is locked because the campaign is already running, a post is already published, your monthly allowance of posts is too small (the card says how many posts it needs and how many are left), or the daily limit for AI generation is reached. When the assistant adds a post to a running campaign, the card also says beforehand when that post cannot be placed: the day is a weekend, has already passed or falls outside the campaign period, or the channel is not connected to the campaign. Each reason appears once, however many posts it applies to. The same goes for a period that ends before it starts or falls outside the length this type of campaign allows (the card names the range in weeks), a channel name that is not connected to your organisation, and a value that is not one of the options of a setting.
• Writing posts uses your monthly allowance. The card says how many posts it would write and how many are left this month, or "no limit" when your plan has no monthly limit. Afterwards it reports how many were started, which is not the same as finished.
• Resuming a paused campaign shows the date from which the posts continue: the date of the first post that is not published yet, or today if that date has passed. A post that would land on a weekend moves to the next working day.
• Channels are shown by their name, for example "linkedin-page: Your company", so you can see exactly which channel is added or removed.
• The channel of posts already in the plan. The assistant changes the channels of the whole campaign, not the channel of one existing post. Asked for that, it says so and points you to the post in the plan of the campaign, where you choose its channel.
• A rule for every future post, such as "from now on, end every post with a question", is often not a setting. The assistant then says so plainly. If the rule is about content, for example a standpoint or a subject to leave out, it offers to save it as a guideline of the campaign, see "What the Assistant Shows and What It Says Is Saved" below. It may also offer a card that rewords the posts that already exist, if you want that; nothing changes unless you apply it. A rule that is a setting, such as the tone, gets its own card.
• Changing the period of a running campaign asks one more confirmation after you press Apply, showing how many posts would move.
• If the campaign changed while the card was open, the card is not applied and a fresh proposal is made.
• The assistant saves through the same road as the campaign page. When it adds posts, only the new posts are added: the posts already in the plan are not touched and keep what you did to them. The same checks apply as on the page: your monthly allowance, locked posts that stay as they are, and a channel that is not connected, which is refused instead of being dropped without a word.
• If a step fails, the card says which one. The steps before it were applied; the failing one and those after it were not.
• Removing posts. Ask the assistant to remove posts from the plan and it answers with a card that asks you to confirm; the posts are removed only when you confirm. It never rewrites a post to say "removed". Only an admin or an editor can have posts deleted this way, and a locked post stays where it is until you unlock it.
• Reading the outline first. Tell the assistant that you want to read the outline of every blog post before it is written, and it saves that wish on the campaign. During setup it is a row on the setup card; later it is a Proposed change card that you apply. See Communication Campaigns.
• What happens when the assistant cannot finish its answer? If it runs out of steps before it has an answer, it no longer shows a half sentence such as "I will look that up". It tells you that it could not finish its answer this time. Ask your question again, or split it into smaller questions.

What the Assistant Shows and What It Says Is Saved

Plain words on every card. A card shows what you can recognise: the name of a choice (such as "Blog post" or an image style) instead of a code, the title of a document instead of a number, and the date in your language. A document that was removed in the meantime appears as "a document that no longer exists". A part that cannot be done names the setting it is about.

It only says something is saved when it is. When the assistant says that a document, a guideline, a default setting, an agreed post or a goal has been saved or linked, Nolorem checks that sentence against your campaign as it is now. If it is not there, the sentence becomes: "That is not in your campaign yet." If the assistant has only prepared it on a card, you read: "That is ready to be saved: the card below shows it, and it is only saved once you apply it." If Nolorem cannot check at that moment, you read that too: "I could not check whether this is already in your campaign." The check recognises the usual ways of saying this; when in doubt, the campaign page itself shows what is saved.

Guidelines: a rule for the whole campaign. A wish that fits no setting, for example "we never compare ourselves to competitors by name", gets no card for a setting. The assistant offers to save it as a guideline instead. You can also write or change it yourself under Guidelines in the campaign briefing (the pencil on the campaign card, the Goals tab, up to 1000 characters). The planner, the filling in of missing posts and the writer all receive the guideline, and in their instructions it wins over a goal when the two conflict. It applies to what you generate from then on; posts that are already written stay as they are, and a plan made earlier shows the notice that the briefing has changed. How often you post does not belong in a guideline: that is set under Period and rhythm. A guideline is an instruction to an AI model, so read the posts before you approve them.

Topics you approve in the chat. When the assistant proposes a list of topics and you approve them, it gives a card with the row Topics to agree on, which shows the day of each topic before you apply. Nolorem chooses those days, not the assistant: working days inside the period that follow the rhythm of the campaign. When not everything fits in the period, the assistant says so in plain words and the card holds only what fits. After you press Apply, the topics are saved as agreed posts. On a campaign that is not running yet they appear in the plan the next time you generate it. Before you apply, the assistant does not say that they are added or saved. See "Agreeing on Posts with the Assistant" in Communication Campaigns.

Titles with a colon. The assistant does not propose titles with a colon between two parts, because they often look as if an AI wrote them. If you bring such a title yourself, the assistant does not block it: it explains briefly why such titles are usually avoided and suggests an alternative for each. Nolorem's checks are built not to rewrite a title that you typed yourself. If you keep your own version, the assistant is meant to use it exactly as you wrote it; look at the card before you apply it.

A card that is out of date is not applied. If the plan changed after the card was made (you changed it, a colleague did, or an earlier card already did), pressing Apply writes nothing. The assistant reads the plan again, tells you the situation changed since the card, and shows a fresh proposal. Nothing changes until you apply that one. When a figure it gave earlier has changed, it says so. If the assistant named a number earlier in the conversation, for example how many posts a campaign has, and that number is different now, you read: "Earlier I said 12, now it is 14." It does not repeat the old number as if it were still true.

Changing One Post Through the Assistant

On the post list, the calendar and an open post you can ask the assistant to move, un-schedule, publish, approve or delete a blog post or social post that is not part of a campaign. "Move the post about the opening to Thursday at 10:00" and "Delete this draft" are examples.

• Always with a confirmation. The assistant shows a card with the post and what will happen. Nothing changes until you confirm it.
• Your role decides. Moving, un-scheduling, publishing and approving need the right to publish; deleting needs the right to delete. When your role does not allow it, you read a sentence saying so and nothing changes.
• The post must be in the right state. A post that is waiting for approval cannot be moved until it is approved, a published post is not moved or published again, and approving a social post needs a time. The assistant says what is in the way.
• Publishing names the destination on the card: the blog and its CMS, or the social channels. When there is no channel to publish to, the card says so and has no confirm button.
• Deleting moves the post to the trash, where it stays for 30 days; a social post also leaves its networks. A blog post that is live on your site is taken offline, and the card says so. Restoring an item and emptying the trash are done on the Trash page.
• What you read afterwards comes from the saved post. The assistant says "published" only when the post is published; a social post that is queued is called queued.
• If the post changed while the card was open, the card is replaced by a fresh one and nothing is written. If something goes wrong you read why, and the next answer of the assistant starts with that reason. Confirming again after a failure first checks the post again.
• A post in a campaign is changed through the campaign. The assistant does not offer this card for it.

Text and an Image in an Open Blog Post

In an open blog post you can ask the assistant for new text for one part of the post, or for a new image.

• Text for one part. Ask for example "write a shorter introduction" or "rewrite the second section more formally". The assistant proposes text for one section, the title, the introduction or the meta description, and shows it next to the current text. Nothing changes in your post until you choose Put it in the editor and then Take over and save in the editor. A whole post is not replaced at once: ask per section.
• The panel makes room. On a screen at least 1280 pixels wide, the page moves aside while the assistant is open, so you see the proposal and its buttons in the editor next to the panel. On a narrower screen the panel closes when you choose Put it in the editor, and also when an image you asked for is ready, so you can accept or discard it.
• Your unsaved work comes first. If you changed that part yourself and have not saved it yet, the editor shows both texts and you choose. With Keep my text nothing of yours is overwritten.
• Formatting counts as your change. If you made a word bold or added a link in that part, that is also your own change: the editor shows both texts and you choose.
• Changed your mind? With Do not take over nothing changes, and the card in the assistant can send the proposal to the editor again.
• At most two proposals per message. Ask for more parts at once and the assistant says which ones are still waiting; ask for them in your next message.
• On a language tab. The proposal is for the source text of the post. On the tab of another language the card says so; open the source tab and put it in the editor there.
• An image only after your yes. Ask for example "make an image for the second section". The card says beforehand that it uses one image of your monthly allowance and how many you have left. The image is only made when you click Make the image, and it appears in the images of the post, where you accept or discard it as after the Generate button. An image for a section is placed in the text; a featured image is made from the topic of the post.
• A version on every save. When you save in the editor, with or without text from the assistant, Nolorem keeps a version of the post, at most one every ten minutes. You cannot browse these versions in the editor yet; they are included when you export your data. Before text from the assistant replaces a part, Nolorem first keeps a version of the post as it was, so the text from before the assistant is always kept.

What the Assistant Does Not Do

The assistant never does these for you, even if you ask. It tells you where to do them yourself:

• Publish a post that is part of a campaign. That goes through the campaign. For a post outside a campaign, see the section above.
• Cancel a campaign. You cancel on the campaign page.
• Restore something from the trash or empty the trash. You do that on the Trash page. Moving a post outside a campaign to the trash, the assistant does after your confirmation.
• Start a draft campaign. You start it on the campaign page. The assistant can pause and resume a campaign that is running.

It also cannot reliably change the tone over a whole range of posts. A request like "make posts 3 and 4 more formal" does not reliably produce a change. Asking for one specific field works better, for example "rewrite the summary of posts 3 and 4 in a more formal tone".

When the Assistant Can Only Point You Elsewhere

The advisor offers a card only for a change that is possible in the situation you are in. If it prepared a card for something that cannot be done at that moment, the card is not shown. You read a sentence that says what is not possible and, where there is one, the button that does it:

• Starting a draft campaign. The advisor cannot do it. Use the Generate button on the campaign page.
• Changing the audiences of a campaign, or adding a post to a running one, while its page is not open. Open the campaign and use the briefing dialog or the add-post control in the plan.
• Writing a whole new blog post in one go. The advisor opens the generator for that. In an open post it proposes text per part, see "Text and an Image in an Open Blog Post".
• An unsaved draft in the social composer. The advisor reads a social post you have saved. A draft that only exists in the open composer is not sent to it, so save it first or tell the advisor what it says.

When a card was dropped for another reason, the advisor says: "I had promised something here that I cannot do, so I took it back. Nothing has been changed." When a card it prepared was broken, it says so and you can ask again; when only some parts were broken, the rest is shown and the sentence says how many parts were left out. If the advisor names a number of posts that does not match the plan, Nolorem replaces it with the right number. None of these sentences changes anything.

When the Advisor Does Not Answer

When a question gets no answer, the panel says why instead of staying empty, and what you can do: send the message again, split it into smaller parts, or try again later when the daily limit for the assistant has been reached. The panel never shows technical text about the failure.

In a conversation you reopen later, a message that was cut off shows that it was cut off, and a message that has waited more than 10 minutes without an answer shows "This message was never answered, because the connection broke off. Send it again." A message that was answered shows nothing.

Approving Only Some of the Topics

When a campaign has fewer posts than its rhythm promises, the advisor can propose topics for the empty places. When you approve some of them, Nolorem puts the approved topics into the empty places first, per type: a blog topic into a blog place, a social topic into a social place. Only topics left over once the places are full become agreed posts.

• Not everything fits. The card says how many posts are added and why the others are not: there is no free working day left in the period, that type of post is already at the number it has in the plan, or the day would fall outside the period.
• The filled places stay. A post that fills an empty place is locked in the plan, so Generate a new plan keeps it.
• Nothing happens before Apply.

Asking for No Duplicate Content

If you ask for no duplicate, repeated or overlapping content, the advisor looks up the pairs of posts that look like each other, the same pairs the campaign page marks with Looks like, and names them by title. If no pairs are found, it says that the campaign page shows no post that looks like another one, and if some posts could not be measured yet it says that too instead of calling the plan free of duplicates. Once you have asked, a card that adds a post is checked: if the new post overlaps with a post already in the plan, it is not added and the card says so. See Communication Campaigns.

Read Is Not the Same as Added

Reading a document and adding it to a campaign are two different things, and Nolorem checks the advisor's sentence against what really happened. If the advisor says a document is added to the campaign but it is not, the sentence is replaced: "I have read the document [title], but it has not been added to your campaign.", or "A document only becomes part of the campaign through a card that you apply." If it says it has read a document but nothing was read in that answer, you read that it could not read one and that you can attach it again or name it. Questions, plans for later and denials are left as they are. This check is not a guarantee for every possible wording.

Remembering a Fact About Your Company

Tell the assistant something lasting about your business, for example "we only publish non-fiction" or "we only work in Flanders", and ask it to remember that. It answers with a card called Remember this about your company. The card shows the company profile as it is now and as it will be after saving, with the new fact added on its own line. Your existing text stays as it is. Nothing is saved until an admin presses Save to profile.

• Where does it go? The fact is added to the company profile of the blog you are working in. You can read and edit it at any time in the blog settings, under Company. See Blog Settings Overview.
• Who can save it? Only an admin. With another role the assistant does not offer the card and says that an admin changes the company profile.
• Length. The profile holds at most 2,000 characters. If the fact does not fit, the card says so and nothing is saved.
• What is it for? The campaign planner and the writing of your posts read the company profile. Use it for lasting facts, not for a wish about one post.
• Not the same as the assistant's memory. What the assistant remembers about you (see Memory below) is not read by the planner or the writing. A fact only counts for your campaigns once it is in the company profile, so the assistant offers the card even when it already remembers the fact.

Starter and Proactive Prompts

When you open a new conversation, the panel shows suggested questions tailored to the current context, such as "What should I write about next?" or "How are my social posts performing?". It may also show proactive suggestions based on your data, for example when your keyword bank is empty or you are running low on image generations. Click any suggestion to send it.

Rewriting One Post

Ask the assistant to rewrite a post that already has text, and name the post, for example: "Rewrite the post "Our spring open day" completely." The assistant answers with a Proposed change card for that one post. The card:

• warns that the existing text is replaced and that any edits you made to the post yourself are lost;
• says what it costs: "Regenerating counts as a new post in this month's quota.";
• says that a post that was approved and scheduled has to be approved again after the rewrite;
• asks you to tick the box "I understand that the current text of this post is replaced". The button Rewrite this post works only after you tick it.

Nothing is rewritten before you press that button. If the post has no text yet, the assistant offers the card that writes it instead. A post that was published in the meantime is not rewritten. If a monthly limit of your plan is already reached when the writing starts, the post shows that limit and keeps its old text.

Updating a Post or Making a New One From a Draft

When the assistant writes a social draft and you have the post composer open, the draft card carries buttons under the text:

• Update the [channel] post, for example "Update the Acme Marketing post", one for each channel of the composer that the draft is meant for. It first asks "Replace the text of the Acme Marketing post?" with Cancel and Replace text. Only after you confirm is the text of that channel replaced by the draft; the other channels stay as they are. Afterwards the card says that the post now holds this text.
• Make a new social post from this asks you to confirm with Make a new post. The post you have open stays saved as a draft.

If the post is not open in the composer, the card says so and you can make a new post from the draft instead. If something goes wrong, the card says that your post has not changed.

Confirming a Card in Words

When one card is open and you answer in words, for example "yes, do it", the advisor carries out that card exactly as if you had pressed its button. Nothing else is applied, and a few things are never guessed:

• Two cards open. The advisor asks "Which card do you mean?" and does nothing until you say.
• A card with a checkbox per post. The card Proposed changes to the plan is only applied with its own button. A typed "yes" changes nothing there; the advisor answers "You apply this card with the button on the card itself."
• The campaign changed in the meantime. The card is read again and shown to you again ("The campaign has changed in the meantime; here is the card again."), so you never confirm an old picture.
• No card open. A "yes" on its own changes nothing.
• A yes to steps the advisor suggested. When the advisor suggested steps in its answer without a card, a bare "yes" carries out only the steps it suggested, never a different change. When it cannot tell which steps it suggested, it asks: "Tell me which of the steps I suggested I should carry out."
• A no never changes anything. "No" closes the card. "No, but remove post 3" only removes post 3: the rest of the card stays untouched.

Several Wishes in One Message

Ask for several things at once, for example "move the Easter post to Friday, drop the post about the fair and add one about our new opening hours", and the advisor answers with one card that has a checkbox per wish. Nothing is checked yet: you tick what you want and press Apply the checked ones. The button stays off until at least one wish is ticked.

• A wish that needs another one. The card says "Depends on:" and ticking it ticks the wish it rests on as well.
• A wish that is not possible has no checkbox and says why, for example that the post cannot be found or that your role does not allow it.
• A new post gets the day you name. Without a day it gets the first free posting day in the campaign period. Each new post also says whether it looks like a post that is already in the plan (Looks like: with the title), or that this could not be checked. You can still apply it.
• One result per wish. After applying, each line says whether it was applied or why it was not. If the campaign changed while the card was open, nothing of that group is applied and the card asks you to look again.

When the Advisor Offers to Fill the Plan

When a campaign has empty places and you ask the advisor to fill them, it answers with a card of proposed topics for those places, each with its day. That card is an offer: nothing is checked, so you choose which topics go in. Only when you said so yourself, for example "fill everything", are all of them checked for you. A topic you said no to before is left out, and the card says so (see the next section).

A No That Stays

When you say no to a topic, the advisor remembers it, also in a new conversation. When that topic would come up again, it is left out and the advisor says when you said no and who said it: "You said no to [topic] earlier (by you, [date])." A no from a colleague is named the same way.

• The campaign changed since. Then the advisor asks once whether you want the topic now after all.
• You change your mind. Ask for the topic explicitly, for example "add the topic about the fair after all", and the earlier no no longer counts.
• Your role. A contributor's no counts in the current conversation, but is not saved for later; the advisor says so.

When You Say "This One"

When you write "this one" or "that post", the advisor uses the post you have open on screen, but only after checking that it belongs to your organization. When there is no such post on screen, it never picks one silently: it asks in one line "Do you mean [title]?", and only a yes makes it carry on. When it has no candidate at all, it asks you for the title or the date.

Your Own Text as a Post

Give the advisor a text of your own and ask it to post it, for example "post this on LinkedIn next Tuesday: ...". It answers with a card that shows your text letter for letter, the day and the channels, and a Make this post button. Nothing is rewritten, and nothing is posted before you press the button. Afterwards the card says whether the post is scheduled or saved as a draft.

• A day outside the campaign period. The card says "This day falls outside the campaign period." and gives you two options: As a separate post, outside the campaign or Extend the campaign to this day.
• Too long for the channel. The advisor says so and names the limit; it never cuts your text off.

A Plan Rule

You can give a campaign a rule for its plan, for example "never more than two blog posts a month, and at least a week between them". If something is missing, the advisor asks which posts the rule applies to or how many days apart, with a recommendation. The rule is saved with the campaign, and from then on:

• A new plan follows it. When you generate a plan, Nolorem moves posts so the rule holds, keeping their time of day and only using the days your campaign posts on. When the rule cannot be met, the plan is saved as generated and you are told.
• A card follows it. A card that adds or moves a post checks the rule and proposes another day when the rule would break.
• Your own move is never refused. When you move a post by hand on the campaign page and that breaks the rule, the move is carried out and a notice says "This move breaks your plan rule ([rule]); it was carried out anyway."

Ask "remove the plan rule" to drop it.

Smart Cards and Social Drafts

The advisor can render smart cards inline: usage stats with progress bars, stat highlights, and key-value lists. When you ask it to repurpose blog content, it can generate social draft cards with platform-specific text. The buttons on a social draft card are described under "Updating a Post or Making a New One From a Draft" below.

Conversations

The advisor saves your conversations. A conversation is stored for you alone, in the workspace where you had it: other members of your workspace cannot see it, and neither can an admin of your workspace. It is kept for 12 months after its last message and then deleted. You can delete it yourself earlier. A Nolorem administrator can look at a conversation only to check a complaint or when you ask us to, and every such view is recorded with its reason. See Your Data and Privacy.

The list. The Earlier conversations button in the header of the panel opens the list of your conversations in the current workspace. Each entry shows a title, which is the first thing you wrote, shortened; where the conversation began (a campaign, a blog, a post or General); and the date of its last message. On a campaign page the conversations about that campaign come first, under This campaign, and the rest follow under Other conversations.

Search. The Search your conversations field above the list looks in the titles and in the text of your own conversations.

Opening and continuing. Click a conversation to open it and keep talking in it. A conversation about another campaign opens that campaign for you. Cards in an earlier conversation are read only: they show what was proposed and, where it was recorded, what was applied, but they have no buttons. When you continue, the advisor reads the campaign as it is now, and if something still needs doing it makes a new card.

Coming back to a campaign. When you return to a campaign, the panel opens an empty new conversation. The previous conversation about that campaign is at the top of the list.

A deleted campaign. A conversation about a campaign that has since been deleted stays in the list, labelled Deleted campaign. You can read it but not continue it, and opening it goes nowhere.

New chat. The New chat button ends the current conversation and starts a new one. It deletes nothing: the conversation you leave stays in the list.

Deleting. Click the bin icon next to the conversation in the list (Delete conversation) and confirm with Delete. The conversation is removed for good and this cannot be undone.

Memory

The Content Advisor remembers facts about you across conversations, organized into four categories: Niche, Preference, Workflow, and Style. It stores extracted facts only, not conversation text. Your saved conversations, described above under Conversations, are a separate thing, and deleting one does not delete the other. The memory is personal to you and kept per workspace: colleagues in your workspace never see it, and a fact the advisor learned in one workspace is not used in another. While you are chatting about a post you have open in the editor, nothing new is added to it; start a new conversation if you want the advisor to remember something again. Manage it in Settings then Assistant: each fact shows when it was learned, you can delete individual facts with the trash icon, or use Clear all memory to erase everything and start fresh. The memory keeps preferences, not house rules: a fixed rule about what may or may not appear in your texts, such as a word to avoid, is not stored there. Set such a rule with the house rules card the assistant offers. Every user, whatever their role, can delete their own memory facts under Settings then Assistant. A conversation in which the advisor refused something, for example a change your role does not allow, adds nothing to the memory.

What the Assistant Reads on Each Screen

On the screens below the assistant reads what is saved, from the same place the screen itself gets it. When it cannot read something it says it cannot see it, instead of guessing.

• The dashboard. The post figures (total, drafts, scheduled, published, recent) are the figures the dashboard shows. With more than one blog it first asks which blog you mean.
• An open blog post. The title, the SEO title, the SEO description, the slug, the status and the planned date and time, as they are saved.
• A saved social post. The text, the channels, the language and the planning of the post as saved. A draft that is not saved is not seen.
• Blog settings. The current value of a setting such as languages, region, tone, audience, website or publication time, with a link to the settings page of that blog. Passwords, colours and logos are not read.
• Account settings. The assistant cannot see or change your subscription, your payment card or your invoices. It says so and points you to Settings and then Account.

Dates and times are given in your organisation's time zone.

What the Assistant Remembers per Blog

When you start a conversation, the assistant gets a short summary of what happened in your earlier conversations about the same blog: the changes you applied, the proposals it made and the decisions you took. It is made only from what Nolorem saved about those cards and decisions, never from the text of what was said, so it cannot bring back something the assistant merely claimed. The summary is short; when there is more, the oldest lines drop out first.

The summary belongs to you and to the blog: colleagues do not see it, and a conversation about another blog does not use it. It stays the same for the whole conversation.

To make the assistant forget it, delete the earlier conversations concerned (the bin icon in the Earlier conversations list); their lines leave the summary at once, and conversations disappear by themselves 12 months after their last message. Clear all memory under Settings and then Assistant clears the facts the advisor learned, not this summary.