Product / Opportunity Engine

Every gap becomes one action, with a score and a reason.

This is where measurement turns into a backlog. Nothing here is a score you are left to interpret: each finding arrives as a specific thing to do, with the questions it should move, the evidence behind it, and a decision for a person to make.

Quick answer

The Opportunity Engine turns each AI visibility gap into one action, scored on impact, confidence, effort and urgency, with the evidence behind it. A person approves, postpones or rejects it. Brand Brain checks the facts you state against what engines say about you and flags the differences.

Four scores, and what each one is for

Four numbers rather than one, because a single priority score hides the trade-off you are actually making.

  • Impact

    How much the questions this action targets are worth to you, based on their intent and how often they come up in your set. A comparison question a buyer asks before a shortlist is not worth the same as a definition question.

  • Confidence

    How strong the evidence behind the action is. An action drawn from many answers that agree scores higher than one drawn from a single run, and the score moves when the next refresh either confirms it or does not.

  • Effort

    What it will take from you. A missing FAQ answer and a new comparison page are both actions, and treating them as the same size is how backlogs die.

  • Urgency

    Whether waiting makes it worse. A competitor gaining ground on a question, or a fact about you that engines have wrong, does not improve while it sits in a queue.

Each action also carries the type of work it is, from content and website or entity fixes through to outreach and technical changes, so it goes to the right person without being re-read first.

Brand Brain: what engines believe about you

An action is only as good as the picture behind it. Brand Brain is that picture, held as facts rather than as copy.

  • Your facts, as you state them

    Pricing, products, capabilities, people, locations and the industries you serve, held as entities and claims rather than as a paragraph, so each one can be checked separately.

  • What engines say instead

    The same claims as they come back in answers. Where the two disagree you get a flag: pricing that is out of date, a person who has moved on, a capability you shipped that no engine has picked up.

  • The sources shaping the answers

    Which pages the answers about you are citing, yours and other people's. A wrong fact usually has an address, and this is where you find it.

  • AI Readiness audit of your site

    robots.txt, llms.txt, sitemap, structured data and whether your pages answer the question they target. Unglamorous, and the cheapest set of actions on the list.

From gap to measured

The path every approved action takes.

  1. A gap becomes one action

    Not a score and not a theme. One thing to do, written as a sentence, with the questions it is meant to move and the evidence it came from attached.

  2. A person decides

    Approve it, postpone it or reject it. Rejecting asks for a reason, which keeps the backlog from filling up with the same suggestion every month.

  3. The work gets drafted and reviewed

    Approved actions go to the Content Workspace as briefs, where the agents draft and a person reviews before anything leaves the building.

  4. Publishing closes the action

    Publishing the asset marks the action done and schedules its questions to be asked again about a week later.

  5. The before and after sit together

    When the re-ask comes back you see what those questions said before the work and what they say after, on the action itself rather than in a separate report.

What the free plan sees

Enough to judge whether the reasoning holds up before you pay for anything.

  • The free baseline starts by itself at signup and needs no card.
  • Free shows the top three actions, read only: you can read the reasoning and the evidence, and you cannot approve or work them.
  • Paid plans open the full list, the approve, postpone and reject decisions, and the scheduled refresh that keeps the scores current.

Planned, not available yet

These are on the roadmap and are not part of the product today. They are listed so nothing above reads as more than it is.

  • Ranking actions from your own past results is planned. Scoring today weighs the evidence behind an action, and is not trained on your history.
  • Multiple named users, each with their own permissions, is planned. An account is one workspace today, so an approval is recorded against the account rather than against a person.

What these results are, and are not

Everything the product reports is a snapshot of the questions and engines that were tested, on the day they ran. AI answers change between runs, accounts and regions, so results vary and nothing here is a ranking. We do not guarantee that any engine will mention, recommend, cite or include you.

ModernTechLap measures and influences the evidence around a business; it does not control AI model outputs.

Common questions

Who executes the actions?
Your team does. Agents prepare drafts from the stored evidence and the workspace tracks them through review, but the approval and the publishing decision belong to a person. Nothing external happens on its own.
Can I reject an action?
Yes, with a reason. The action leaves your list and the reason is kept with it, so the same suggestion does not come back on the next refresh as if nothing had been said.
What does Brand Brain actually check?
It holds the facts you state about your company as entities and claims, compares them with the claims that appear in AI answers about you, and flags the differences. It also shows the sources those answers cite, which is usually where a wrong fact is coming from.
How do you know an action worked?
Publishing the asset marks the action done and re-asks that action's questions about a week later, so the before and after are on the same screen. It is still a snapshot of tested questions on the engines that were asked, and answers vary between runs, so treat a single move as one sample.

Find out what AI says about you. Then change it.

Free baseline, 25 buyer questions, every AI system you connect. About ten minutes, no card.

Run your free baseline

25 buyer questions across the AI systems you connect, for .

Checking …