Help centre

Every screen, number and allowance explained in plain language. Inside the platform, the small “?” next to an item opens the same explanation and links straight back here.

Using the platform

What each screen does, how to read the numbers and what to do next.

Getting started

Add a brand or project, run your first prompts, then read the report.

EGNITE measures how visible your company is inside AI answers — ChatGPT, Copilot, Gemini, Perplexity, Claude and DeepSeek — and tells you what to change so those engines cite you instead of a competitor.

A normal first week looks like this: create a brand (or a project if you only want to compare a market), add five to ten questions your buyers actually ask, run them across the engines included in your plan, then open Weekly Reports and work down the Fix Queue.

  • Command Center — the summary of everything.
  • Prompt Manager — the questions you track.
  • GEO Scanner — how ready your website is to be quoted.
  • Fix Queue — the prioritised list of changes.
  • Weekly Reports — progress over time, ready to share.

Command Center

Your dashboard: current scores, recent runs and what needs attention.

The Command Center is the home screen of your workspace. It shows your current visibility score, how it moved since the last period, the engines you ran most recently, and the highest-impact items waiting in the Fix Queue.

Everything here is a shortcut — clicking any card opens the screen with the full detail behind it.

Brands and projects

A brand is your own company. A project is any market or comparison you want to track without owning it.

A brand holds your company name, website and competitors, and everything measured against it accumulates over time so you can see progress.

A project is lighter: use it when you simply want to test a set of questions or compare several companies in a market, without declaring one of them as yours. You can switch between brands and projects at the top of the workspace.

Prompt Manager

The list of buyer questions you track. Run them across engines to see who gets cited.

A prompt is a real question a buyer would type into an AI assistant — for example "best flame retardant masterbatch supplier in Europe". EGNITE sends each prompt to the AI engines your plan unlocks and records whether you were mentioned, in which position, with what sentiment, and which sources the engine used.

Good prompts are specific and commercial. Ten sharp questions beat a hundred vague ones, and they are cheaper against your monthly allowance.

  • Mentioned — your brand appears in the answer.
  • Position — where you appear relative to the companies named before you.
  • Sentiment — whether the mention is positive, neutral or negative.
  • Sources — the pages the engine relied on to build the answer.

Bulk prompt upload

Paste or upload many prompts at once and run them in one batch.

Instead of typing prompts one by one, you can paste a list or upload an Excel/CSV file with one question per row. The batch runs against the engines you select and every result lands in the same report.

Each prompt in a batch counts against your monthly prompt allowance, so the batch stops cleanly when you reach your limit rather than failing halfway.

Choosing AI engines

Pick which AI assistants to question before you run. Your plan decides which are available.

EGNITE queries the assistants your buyers actually use: ChatGPT, Copilot, Gemini, Perplexity, Claude, DeepSeek, Grok. Engines unlock in tier order, so higher plans see more of them, and any engine without a live connection is hidden rather than shown as an error.

Different engines answer differently — one may cite your datasheet, another a distributor's catalogue. Running several is how you spot where you are invisible.

  • Spark — 3 engines.
  • Current — 4 engines.
  • Momentum — 5 engines (adds Claude).
  • Surge — 6 engines (adds DeepSeek).
  • Ascend and Beyond — the full set as connections become available.

Deep searches

A deeper, slower pass on the strongest model of an engine — used for the questions that matter most.

A normal run uses each engine's standard model. A deep search uses its strongest reasoning model, which reads more sources and produces a longer, better-grounded answer.

Deep searches have their own monthly allowance because they cost more to run. Keep them for your top commercial questions and use standard runs for breadth.

Compare competitors

Run the same questions for you and your rivals and see who the engines name first.

This screen puts your brand side by side with the competitors you list, using identical questions across identical engines. You see who is mentioned, how often, in what order and with which supporting sources.

Competitors are also grouped by what they actually are online — mineral producer, compounder, resin producer, OEM, distributor and so on — so you compare against your real peer set rather than an alphabetical list.

GEO Scanner

Scores how easy your website is for AI engines to read, trust and quote.

The GEO Scanner crawls the pages you give it and scores them from 0 to 100 across five dimensions, with page-level evidence for every point won or lost. It is the supply side of visibility: prompts show whether engines cite you, the scanner shows why they can or cannot.

Re-scanning the same pages later shows the delta, so you can prove that a change moved the score.

  • Structure — headings, schema markup and machine-readable facts.
  • Answerability — whether a page answers a question directly.
  • Evidence — data, specifications, named authors, dates.
  • Authority — external signals and citations.
  • Freshness — how recently the page was updated.

EGNITE Index

One 0–100 number combining how often engines cite you and how quotable your site is.

The Index blends your citation performance across engines with your GEO Scanner score into a single figure you can track month to month and put in front of a board.

It moves slowly on purpose: a single good answer does not change it, sustained coverage does.

Fix Queue

The prioritised list of changes, strongest impact first.

Every weakness found by the scanner or by a prompt run becomes an item here, ranked by the difference it should make against the effort it takes. Each item names the page, the problem, and a first recommended correction.

Recommendations are a starting point produced automatically. When you want them checked and rewritten by a human, send the item to Expert Verification.

Weekly Reports

Your visibility over time: KPIs, per-prompt detail, sources, authors and exports.

The report is the screen you share internally. It shows headline KPIs, how they moved against the previous period, results per prompt and per engine, the competitors named alongside you, and the sources and authors the engines relied on.

Where there is not yet enough data the report says so instead of showing a zero — a blank KPI means "not measured yet", not "you scored nothing".

Filtering a report

Narrow the report to specific prompts, engines or a period before reading or exporting.

Use the filter bar to pick the prompts and engines you care about and the period to compare against. Everything below — KPIs, charts, sources, competitor tables and exports — follows the filter, so you can build a report for one product line or one market.

Sources and authors

The pages, domains and named writers the AI engines used to build their answers.

For every answer, EGNITE records the sources the engine cited and, where the page exposes one, the author who wrote it. Aggregated across a report, this shows which publications, distributors, associations and individuals shape how your market is described.

That list is your outreach plan: those are the pages worth being present on, correcting or contributing to.

Opportunities

Openings spotted during the runs — questions where nobody owns the answer yet.

Alongside fixes, EGNITE flags openings: questions where engines gave a weak or generic answer, sources that mention your category but not you, and comparisons where a small change in your content would put you in the answer.

These are usually the fastest wins because there is no incumbent to displace.

Exports and sharing

Download the report as PDF or Excel, or share a read-only link.

PDF is for circulating internally, Excel keeps every row for your own analysis, and a share link gives a colleague or client a read-only view without an account. Exports respect whatever filter is active when you generate them.

Expert Verifications

Send a recommendation to our team; you get a corrected, publish-ready version back within 48 hours.

Automatic recommendations get you 80% of the way. An expert verification sends the item to our reviewers, who check the facts, rewrite the copy and the structured data, and return a version you can publish — normally within 48 hours.

Your plan includes a number of verifications each month, and extra single verifications can be bought when you need more.

  • Spark and Current — no verifications included.
  • Momentum — 1 per month.
  • Surge and Ascend — 2 per month.
  • Beyond — agreed with you.
  • Extra single verification — €149.

PQC Scanner

Checks how ready your public endpoints are for post-quantum cryptography.

The PQC Scanner probes the TLS configuration of your public services and scores how exposed you are to "harvest now, decrypt later" risk, with concrete remediation steps ordered by urgency.

It is included on Ascend and Beyond.

Alerts and notifications

Get told when your visibility moves or your allowance runs low.

EGNITE can notify you in the app, by email, and on Slack or Teams when a tracked prompt changes materially, when a competitor overtakes you, or when you approach a monthly limit.

Account, access and approval

Workspace access is opened by us once your plan or beta invitation is confirmed.

Creating an account is the first step; the workspace itself opens once your subscription or beta invitation is confirmed. Until then you land on an activation page with a way to reach us.

If you ever see that page while you are a paying customer, it is a temporary check that failed — use "Retry", and contact us if it persists.

Data, privacy and security

Your data is stored in the EU, access-controlled per account, and never sold.

Customer data is hosted in the European Union, isolated per account, and only reachable by the account it belongs to. We publish our subprocessors, our privacy notice and a data processing agreement.

Plans, quotas and billing

What each offer includes, how allowances work and how to get more.

What each plan includes

Six offers, from a free trial run to a fully managed programme.

Every plan includes the platform, the GEO Scanner, the Fix Queue and reporting. What changes is how much you can run each month, how many AI engines you reach, and how much expert work is included.

  • Spark — €0. 3 prompts, 3 engines, 1 GEO scan.
  • Current — €99/month. 60 prompts, 20 deep searches, 4 engines, 5 GEO scans.
  • Momentum — €249/month. 250 prompts, 50 deep searches, 5 engines, 20 GEO scans, 1 expert verification.
  • Surge — €449/month. 300 prompts, 70 deep searches, 6 engines, 40 GEO scans, 2 expert verifications.
  • Ascend — €749/month. 500 prompts, 200 deep searches, 6 engines, 100 GEO scans, 2 expert verifications, PQC Scanner included.
  • Beyond — on request. Unlimited usage, every engine we support, a managed programme and agreed service levels.

How monthly allowances work

Allowances reset each billing month, counted from the day you subscribed.

Prompts, deep searches, GEO scans and expert verifications each have their own monthly allowance. The month runs from the day you subscribed, not from the first of the calendar month, and unused allowance does not roll over.

The usage meter in the workspace always shows what is left, and you are warned before you run out rather than after.

When you reach a limit

You are notified, and can either buy extra credits or move up a plan.

Runs stop cleanly at the limit — nothing is half-charged and no data is lost. From the notification you can request extra credits for the current month, or move to the next plan, which is usually cheaper if you hit the ceiling regularly.

Billing terms and discounts

Monthly, annual (−7%) or two-year (−15%).

You can pay monthly, or commit for a year at 7% off, or two years at 15% off. Payment today is by bank transfer against an invoice; the request form in the pricing page captures what you need and we send the invoice.

Changing or cancelling a plan

Move up any time; you keep access to what you paid for until the period ends.

Upgrades take effect immediately and the new allowances apply from that moment. If you cancel, your workspace stays available until the end of the period you already paid for, and your history remains exportable.

Beta accounts

Invited beta testers get a normal account on the plan they were given — never admin rights.

Beta accounts are ordinary customer accounts with the quotas and engines of the plan they were invited on. They see exactly what a paying customer on that plan sees.

Still stuck?

Ask the assistant in the bottom-right corner, or reach a human.