From first idea to post-launch

We predict how people respond to your content.

Upload an idea, a mockup, a rough cut, or a finished ad. Choose your audience, and see how they'll react, second by second: what grabs attention, what they feel, and what they do next. In minutes.

Grounded in real neuroscience

What do you want to know about your audience?

North AI commercial.mp4 1:20

Chat: seconds

What is your main ask today?

North AI Simulate

Attention across the ad0 to 100

Lowest point533176
Target audiencesEveryone else
Real answers from a run on our own commercial, 18 Nov 2026. Each point in the three crowds is one simulated viewer's reaction, 327 in all.

100K+

Audiences / sim

7

Platform models

9

Agent types · 3 tiers

100+

Paths / scenario

Backed & supported by

Trusted by

How it works

From upload to answer in three steps.

Upload your content, choose your audience, and read how they respond.

  1. Upload your content

    Drop in your video, or an image or text. It can be a rough idea, a mockup, or a finished ad. Add competitors if you want them in the picture.

    The demo at the top uses our own commercial.

  2. Choose your audience

    Pick who you're trying to reach. We build them from 100,000 profiles that notice, feel, and react the way real people do.

    The three crowds in the demo are your two target audiences and everyone else.

  3. See how they respond

    Watch a simulated audience react, second by second: what grabs attention, what they feel, and what they do at the end. You get the best, expected, and worst case.

    Simulate answers in about a minute. Create Report takes 12 to 45 minutes and builds the breakdown below.

Our own commercial, tested on three simulated audiences: 327 reactions in all.

Click a person or the logo in the video, or pick an object from the list.

Scene 5 of 220:09 / 1:20

Pick a question about this ad. Each answer comes from our report on it.

North AI

For everyone else, only for five seconds. They start at 70, then drop to 44 and never get back above 48.

North AI

Between 0:17 and 0:24. Everyone else falls to 31, and your targets dip to 53, their lowest point in the ad.

North AI

Between 0:31 and 0:42, when the ad explains the product. Your targets are back at 76, close to the 79 they started on.

North AI

Your targets average 73 across the ad. Everyone else averages 44 and stays below them the whole way through.

ChatSimulateCreate ReportCreate Survey

Chat: seconds

We break your video down, frame by frame.

Before we predict anything, we look at what's actually in your content.

Scenes

We split the video into scenes and read its pacing.

On screenIllustrative example

We detect the logos, products, people, and text in every shot.

Rows, top to bottom: person, product, logo, text

Sound

We read the audio: music, voice, and where they land.

MeaningIllustrative example

We read the tone and message of each moment.

WorriedCuriousConfidentTenseRelievedConfident

This is the fingerprint we use to predict how your audience reacts.

Why you can trust it

Grounded in real brain data, not opinions.

Our model was trained in the lab on real brain and eye-tracking data from hundreds of people. It reads what your content does to attention and emotion, not what people say in a survey.

When the stakes are high

Confirm it with real viewers.

Recruit a panel from your exact audience. They watch with eye-tracking, and you see how real people actually responded, right next to the prediction, including which logos and moments held their eye.

  • Real attention, second by secondHow a real audience's focus rose and fell across your video.
  • Where their eyes actually wentWhich logos, products, and moments held the gaze, and which got skipped.
  • What they said, in their wordsVerbatim reactions from people in your exact target audience.

Ask anything, at any stage

Use it from Claude, ChatGPT, or any tool your team already lives in.

  • Connects through MCP, with no new workflow to learn.
  • Ask in plain language; the simulation runs behind the scenes.
  • Answers land in the same thread as the rest of your work.

While planning

Simulate a Go Viral scenario on Instagram: US Gen-Z women, $5,000 budget, Creative A.

North AI

Running 100 scenario paths…

Instagram peaks day 4. Budget lifts reach 3.2× in week 1. Organic sustains if early saves >8%.

Boost on day 1 and 5, not continuously.

Mid-strategy

How does our audience hold attention versus Competitor X on TikTok?

North AI

Competitor X spikes harder in the first 12h but drops 2× faster than your baseline. Their audience chases trends; yours is more loyal.

You win on week-2 retention.

In a meeting

We have 3 ebooks. Which resonates best with 25–34s on LinkedIn versus Reddit?

North AI

Ebook 2 wins on sustained attention for that group.

LinkedIn: Ebook 1 gets more shares. Reddit: Ebook 3 matches the tone.

Different content for each platform.

After the campaign

Creative B had lower initial engagement, but the sim predicts more total views. How?

North AI

Creative A spiked early: paid drove it, and paid ramps fast then decays fast.

Creative B started slower but held attention across more segments, sustaining organic spread without extra budget.

Creative B wins past week 1. Creative A is the launch spike.

Works inClaudeChatGPTCursorAny MCP-compatible tool
MCP toolsanalyze_mediasimulate_contentcreate_upload_url
Simulates acrossTikTokYouTubeInstagramFacebookXLinkedInRedditPinterestSnapchatWhatsAppMessengerTelegramWeChatLINE

What the simulation tells you

Six signals you can act on before you spend.

PaidOrganiccrossover d12

Paid vs organic separation

Paid reach ramps fast, peaks high, and decays quickly. Organic spreads slower but sustains. The simulation models both independently, so you see where budget ends and content takes over.

d0peak d4d14

How long attention lasts

Every platform fades at its own speed: TikTok content drops in hours, YouTube sustains for weeks. See exactly when attention drops and which format holds longest.

18-24
68
25-34
82
35-44
45
45+
26

Audience breakdown

Who responds, by age, gender, and behaviour type. Passive scrollers, active sharers, and power users carry different inertia and thresholds, calibrated from demographic data.

YouCompetitorBot

Competitor & bot dynamics

Competitor activity competes for the same attention. Fake engagement is modelled separately, so you can tell real signal from inflated early metrics.

Cross-platform spread

Each platform has its own network model. Content moves through algorithmic feeds on TikTok, subscriber networks on YouTube, and group cascades on WhatsApp. The simulation tracks every path.

BestExpectedWorst

Range of outcomes

We run every scenario hundreds of times. You get the best, expected, and worst case, not a single number, so you know the downside before you commit budget.

The technology behind your offering

Objective neuroscience

AI models trained on neuroscience data predict how audiences respond: attention, engagement, and emotional response, frame by frame. Validate with real viewers when the stakes are high. No opinions, no bias, just brain data.

Entertainment & media DNA

Our team comes from Sony, NBC Universal, and Warner Bros. We've worked with brands, studios, and agencies at every scale. We speak your language.

Backed by leading institutions

Backed by Innovate UK, the University of Manchester, and the Alan Turing Institute. Supported by AWS, NVIDIA, and BridgeAI. Investors include Praetura Ventures and SFC Capital. Patent filed.

Inside the product

From a question to a decision, in one place.

Brand & competitor dashboard

Track share of voice, competitor moves, and live trends across every platform you compete on.

Simulated-audience report

Simulated-audience report

Best, expected, and worst case, with the full breakdown behind every chat answer.

What practitioners say

Trusted by the teams who ship creative every week.

The video insights from North AI are deep and help us invest in the right ad creative. Balancing human insight and AI analysis gives me confidence the data is accurate.

Tom White

Asset Manager, Science-inc

The insights from North AI have been really impressive, and especially helpful when approaching commissioners. Companies need insights like these to support the decision-making process.

Eline Van Der Velden

CEO, Particle6

We used to rely on gut feeling and post-launch metrics. North AI showed us what was working. The predictions matched our real-world results, something no survey panel has done for us.

James Morley

Head of Creative Strategy, Hawk Media