What is new is, and what is coming

On the pricing page and in the app you now choose between cancel-monthly and an annual commitment. The prices on the site (49, 129 and 399 per month) are the annual rates; cancel-monthly costs 65, 159 or 499. Both are billed per month; the year commits you for a year and then renews per year, cancellable at the end of it.

The fourth plan was a fixed bundle of 15,000 measurements for a fixed amount. Whoever measures that much almost always needs their own arrangement, about volume, price per measurement and availability. That is why it is now Enterprise: all the benefits stay (dedicated onboarding, quarterly review, dedicated contact, first access to new assistants), the bundle and the price are tailored. Existing Schaal subscriptions continue unchanged.

Bureau goes from 4,000 to 5,000 measurements per month and Schaal from 10,000 to 15,000, for the same price. An agency with eight clients measuring weekly no longer fit in Schaal; now it does. Existing subscriptions get the larger bundle at the next renewal.

Whoever structurally measures more than 15,000 measurements per month gets their own price per measurement that goes down with volume, with arrangements about availability and billing. No fixed bundle; you request it via the pricing page.

Grok and Perplexity charged the promised double rate only on a single measurement; the weekly measurement round quietly charged 1 credit. That has been corrected, so the usage you see on the plan screen matches what the site promises.

Alongside Start, Groei and Bureau there is now Schaal, with 10,000 measurements per month. For agencies and brands that want to measure their core questions daily and are outgrowing 4,000 measurements.

Groei can now split the measurement budget across up to three brands, Start measures one brand, and Bureau and Schaal have unlimited brand workspaces. You divide the measurements yourself.

The app now has a clear distinction between your organization and the individual projects beneath it, with separate settings for users, keys, plan and billing. Switching organization is one click.

The free AI check runs on real measurements, and those sometimes take longer than a minute. The check no longer waits for that: you get the result by email, with a shareable page included.

The measurement pipeline gained Perplexity as an extra source, plus retries with a ceiling and a cleanup loop, so stalled measurements no longer pile up on their own.

Every screen in the app now opens with a short explanation of what you see and what you can do with it, so a new user does not have to guess.

Invoices with VAT are ready in the customer portal, immediately usable for the bookkeeping, without you having to ask.

Measurements that belong together now run bundled instead of one by one, so a round is noticeably finished sooner.

Measurements run across three assistants separately, each question repeated several times, and every percentage carries a 95% confidence interval. Assistants are shown separately, because the average describes none of them.

On the site you can have your own brand measured across the three assistants without an account, with real measurements from that moment instead of a simulation.

From pricing page to paid subscription without an intermediate step: checkout, customer portal and separate credit packs to buy extra. You can cancel monthly.

Agencies deliver reports under their own brand, on their own domain (rapport.jouwbureau.nl), which automatically gets a certificate. The client logs in on the own domain.

Every figure ends in a concrete action with a handling status (open, in progress, done, ignored), and per action the app measures the effect. The sources are at the top, not the technical side.

You share a report with a link with an expiry date, without the recipient having to log in. Handy to walk a client or colleague through the figures.

Per project you set a webhook that warns you of a real change. With overlapping intervals the alarm stays silent, so you do not get used to noise.

Put your brand question by question next to one competitor, and let the app suggest competitors based on who is cited in the answers.

An audit that shows how well your site is prepared to be cited by assistants, with the findings directly on screen.

Invite team members, switch between organizations, and receive a summary by email after every completed measurement round. Onboarding asks about your role, industry and competitors at the first project.

The three tables that an agency takes over into its own reporting can now be exported as CSV.

At the first project the app asks about your industry and your main competitors, so the first measurement is about something right away instead of hanging in a vacuum.

After a completed round you get the essentials in your inbox: what rose, what fell, and where your brand disappeared from the answers. So you do not have to remember yourself that something is ready.

The app reads along who is cited in the answers and suggests competitors based on that. Often names come up that you had not thought of yourself.

The most important figures to the top, the technical side to the bottom. The first glance at the screen now says something right away, without you having to search.

Under each percentage you expand the question and a sample answer, so a figure never hangs loose in the air and you see where it comes from.

Measurements got a queue with a rate limit, so a large round does not overload the sources and finishes neatly instead of stalling halfway.

The app got accounts and opened up to a first group of agencies and brands. From now on your measurements sit behind a login with their own history, no longer in a loose trial screen.

An agency rarely measures one brand. Projects now sit side by side, each with its own questions, sources and history, so you can look at each client separately.

Measurements are kept over time, so you see the movement and not just today's standing. A drop stands out that way before a client asks about it.

Client gone or brand phased out? A project can be archived, out of sight but not gone, so you can look back on it later.

Large measurement rounds moved to the night, so you have fresh figures in the morning without waiting on a running mill during the day.

Not just whether you are mentioned, but which source the assistant cites with it. The cited pages are listed, from your own site to a competitor's.

A measurement round can now be downloaded as a neat PDF, with the percentages and the sources included, ready to forward to a client or colleague.

Every measured answer is stored, so you can look back later at what an assistant exactly said and not just keep the bare figure.

Being mentioned is one thing, being up front another. The report now shows whether your brand appears first or only at the back of the answer.

Answers where the brand only recurred in the question but not in the answer itself no longer count. Only what the assistant says of its own accord counts.

Alongside ChatGPT the app now also measures Google Gemini. The two are shown separately, because a brand that one mentions need not be mentioned by the other.

An assistant does not answer the same way twice. That is why the measurement pipeline now asks every question several times and averages the outcome, so one chance answer does not determine the score.

Questions can be grouped by topic, so you see that you are strong on price but weak on reliability, instead of everything vanishing into one average.

The two sources sit side by side per question, so you see at a glance where one mentions you and the other does not.

You record your brand, your site and your most important questions in a workspace, so a measurement round can run again with one click without filling everything in anew.

It began with one question: does ChatGPT mention your brand when someone searches for it? The first measurement pipeline actually asked the model that question and counted whether the brand was in the answer, instead of estimating it.

One screen with the percentage of answers in which your brand appeared, per question. Rough, but based on real answers and not on search volumes.

The question list was written the way people actually ask an assistant, not as loose keywords, so the measurement is closer to reality.

The questions and answers are measured in Dutch, not translated from English, because a Dutch customer asks an assistant in Dutch too.

Some brand names are also everyday words. The counting learned to tell the brand from the word, so a chance sentence does not inflate your score.