Opportunities are prompts where at least one competitor was mentioned by an AI model, but your brand was not. Each opportunity represents a concrete gap where you could gain visibility.

Opportunity data is generated for every plan. On FREE, the page shows a preview of the first few items; full access is available on STARTER and above. Generating content from an opportunity happens in the Content Engine studio and is metered by your Agent-credits pool (available on every plan).


A page organised by question

On opening, the page first answers "where do I stand?" with four figures, then "what should I do first?":

FigureContents
To doQuestions where AI assistants cite other brands than yours, no action taken yet.
In progressContent is ready or deposited; the next analysis will tell whether it changed the AI answers.
WonAfter publishing, the re-measurement found the AI assistants now name you.
ResistingPublished, but still absent after re-measurement. This is not a verdict on the publication: two calls to the same model share only ~15% of the brands they cite.
  • Do this first: the open question with the highest priority score, showing who is cited instead of you, what the AI assistants read and a single button for the next step. The ranking is recomputed after each analysis; each card shows its raw evidence rather than a generated summary.
  • Where to act: the sites AI assistants read, grouped by what you can do there: on your own site, where to sign up or publish, competitors' ground (a link leads to Competitors › Their sources), to check yourself.
  • All gaps: the list of every question with a gap, with tabs (To do, In progress, Resisting, Won, Dismissed) and a search. Each row shows who is cited instead of you, how many answers name you, where to act, and carries its own button; a click opens the detail.

The filter bar keeps only market and topic. There is no period filter: an open gap stays visible however old it is, hiding it would have implied it had disappeared. On FREE the page keeps the same layout: the figures are real, but only the first to do gaps are readable; the following ones appear as text-free placeholders.


Priority scoring

Each opportunity has a priority score from 0 to 100, made of four weighted signals:

SignalWeightWhat it measures
Gap45The gap between competitors' appearance rate and yours on that prompt.
Spread20The share of your active prompts touched by this kind of gap.
Accessibility15How easy it is to get referenced on this type of source (a directory is more accessible than a long-form article).
Influence20The real weight of the sources involved, when a benchmark exists for them.

When none of the gap's sources have an influence benchmark, that signal is excluded from the calculation and the other three share the 100 points — it is never replaced by an invented value. The score's breakdown is shown on the card, not just the total.


Opportunity detail

Every opportunity has its own URL. It shows:

  1. The tracking steps: to do, content deposited, published, re-measured, then won or resisting.
  2. Four figures: answers that name you, the brand cited instead of you, sites read by AI assistants, priority. "—" when the measurement does not exist (a gap that predates the survey).
  3. What the AI assistants answer: the answers of the last 90 days to this question, all analyses combined (a different scope from the gap's, measured at detection, and the page says so).
  4. The sites AI assistants read, and whether your own site is cited.
  5. The numbered action plan, with a single button dictated by the gap's real state: create content, open the draft, see the publish date, or ask Mentova why published content did not move the measurement. You can also dismiss the gap.

When a gap is closed, its sites are no longer tracked and the page says so rather than claiming no site was read.


Lifecycle

The lifecycle is never set manually: it is derived from facts already on record.

StateCondition
ActiveNo draft, no publication.
Content draftedA draft is linked to this opportunity.
PublishedThe draft was published, not yet re-measured.
ResistingPublished, then re-measured, and your brand still isn't appearing.
ResolvedYour brand reappeared in AI answers — the gap is closed.

Acting on an opportunity

  1. Click Create content — the Content Studio opens with the opportunity's context pre-seeded in the chat composer, one send away from a scored draft.
  2. Once a draft exists, the button becomes Open draft.
  3. Once published, the publish date is shown; check your publishing connectors at a glance in the "Where to publish" block, which leads to the Publications page (Automation › Publications).
  4. If the content is published but the gap resists re-measurement, ask Mentova to analyze why and suggest a follow-up.

Automate

On brands with at least one connector (WordPress, Webflow, GitHub — or Zapier, with the caveat described below), an Automate button chains the first three steps: it generates the content, saves it as a draft on your site, then reads that draft back straight from your site to show you what it actually contains. The click first opens a launch dialog — even with a single connector — naming the target, the format that will be written, the cost (5 Agent credits, or 15 with an illustration) and your balance. That is where you choose, for this launch only, between Save as draft (the default: nothing goes live) and Write and publish live, whose consequence is written under the button: once live, the article is immediately visible to your visitors and to Google. That choice is never remembered — not on the brand, not on the connector — and scheduled publications never see it: they always save a draft. The same dialog carries an Add an illustration checkbox (+10 Agent credits, unchecked by default, likewise chosen per launch and never remembered): the image is created after the text and travels with the article to your site — featured image on WordPress, image field on Webflow, a file in the Pull Request on GitHub, imageUrl in the Zapier payload. If the illustration fails, the article is saved without it, the arrival screen says so and the image's 10 credits are returned to you; if the text itself fails, both charges are. If, however, the processing is interrupted midway (a restart of the server running the chain, for instance), the article may not be saved and those credits are not returned. Even when you ask to go live, Mentova saves the draft first, reads it back, then publishes it: if that last step fails, the draft stays untouched on your site and the screen tells you so, with its edit link. Without a go-live request, a Publish live button appears once the draft is confirmed, and nothing is published without that click. On GitHub, the draft is a Pull Request opened on a dedicated branch, never on your production branch, and Publish live merges it.

A Zapier connector is also a possible target, but it cannot confirm anything: Mentova sends the content to your Zap and gets an acknowledgement, nothing more. The launch dialog tells you so as soon as you select it, your choice between draft and live goes out once in the payload (status) — your Zap does what it wants with it — and the arrival reads "Sent to", never "saved as draft". When a brand has several connectors, the dialog preselects the one that can read the draft back.

When the target is a social network connected via Zapier — LinkedIn, Facebook, X or Reddit, see Content Engine — the launch dialog changes shape: there is no draft on a network, so there is no choice to make. A single button, Write and publish to LinkedIn (or the network in question), with its consequence written underneath: the text will be posted immediately, visible to everyone, and no review will be possible, since Mentova cannot read back what Zapier sends. The format announced before the click is the network's own — LinkedIn post, X thread, forum answer for Reddit, Facebook post — and that is the format the 5 credits produce, not the gap's. As with any Zapier connector, the arrival reads "Sent to", never "published". One destination per launch: covering three networks means three launches, and three times 5 credits.

The dialog stays open and becomes the progress view — each step named with its target ("Saving on mysite.com…"), the elapsed time, and a notice when it is taking longer than usual — then the arrival. The arrival hands you the draft's edit link (on WordPress, the wp-admin edit page; on GitHub, the Pull Request) rather than its public URL, which returns a 404 to visitors as long as the article is not live; Webflow provides no such link, and the screen says so. An Automate's state is rebuilt from the server: closing the dialog, refreshing the page, opening another tab or another device loses nothing, and a second launch on the same gap is refused while the first one is running.

Automate costs the same 5 Agent credits as a regular generation (15 with the illustration, charged separately: 5 for the text, 10 for the image) and runs one gap at a time — there is no batch version for automating several opportunities at once: a Starter plan has 75 Agent credits for the whole month (chat included), and automating a dozen gaps in one click would burn through most of it.

If no connector is connected, the button stays visible and clickable: it opens a dialog that names the reason and lets you connect WordPress, Webflow or GitHub right there, without leaving the page (Zapier is connected from its own wizard on the Publications page). Once the connector is saved the button becomes usable — generation itself only starts on your next click. Below the required plan (Pro, which carries the connectors), though, the button stays disabled: only a plan change lifts that one, not a connector. During the 14-day trial the connectors are unlocked so you can watch the chain work — the launch dialog reminds you of it on your first Automate. When the trial ends, a draft already saved stays on your site: you publish it from your CMS.