Skip to content

Pilot by the numbers · Snapshot of 20 September 2026

What the platformhas actually done.

Pilot is a platform where agencies and advertisers talk to their campaigns. The agent reads the ad platforms, cross-references the sources, builds the dashboard and, when asked, changes the campaign.

  • 0

    connectors

  • 0.0 bn

    impressions analysed

  • 0

    dashboards generated

  • 0 min

    to produce one

  • 0 M

    rows consolidated

Every figure on this page is pulled from the production database on 20/09/2026. None is estimated.

01

What Pilot steers

The real scale of budgets and volumes

The four figures to remember. They show the budgets and volumes that actually run through the platform.

€0.0M

of ad budget tracked

Cumulative spend of the campaigns in the database, across all platforms.

0.00bn

impressions analysed

1,309,094,339 impressions exactly, across 822 campaigns.

0.0M

clicks measured

Every click tied back to its campaign, its ad and its keyword.

0.00M

conversions attributed

The end result, cross-referenced between the ad platform and analytics.

02

Before / after

The same work, in a fraction of the time

Eight everyday tasks for an acquisition team. On the left the manual method, on the right the duration measured with Pilot, and the gap between the two. Pilot durations are measured in production; manual durations are industry orders of magnitude.

Building a multi-platform dashboard

Google + Meta + TikTok

By hand

2 to 3 days

Exports, cleaning, reconciliation and formatting, all by hand.

With Pilot

5 min 18 s

One sentence in the chat. 318 seconds on average, from message to dashboard.

Measured across 353 dashboards produced in production.

Answering a numbers question

"What's my CPA on Meta this month?"

By hand

15 to 30 min

Open the platform, filter, recalculate, paste into the spreadsheet.

With Pilot

28 s

The fast path queries the warehouse directly. Best time measured: 0.3 s.

Median time measured across all production runs.

Auditing wasteful search terms

293,473 search terms in the database

By hand

3 h per account

Export the search terms, cross-reference conversions by hand.

With Pilot

5 min

The agent reads all 293,473 terms and returns a sorted list, ready to negate.

293,473 real search terms stored, not third-party estimates.

Auditing an e-commerce catalogue

Merchant Center · 116,867 products

By hand

2 days

The spreadsheet gives up beyond a few tens of thousands of rows.

With Pilot

40 min

116,867 products read from a single question, with no sampling.

116,867 unique products tracked, across 44 product types.

Cross-referencing SEO and SEA

Search Console + Google Ads

By hand

1 day

Two exports, two tools that don't talk to each other, a manual merge.

With Pilot

10 min

590M organic impressions and 1.31bn paid ones, in the same warehouse.

590,247,060 organic impressions, comparable with the paid ones.

Collecting the data every day

Every account, every platform

By hand

45 min a day

Reopening the platforms each morning to see what drifted overnight.

With Pilot

0 min

Collection runs on its own at 5 a.m. The drift is flagged before you reach the office.

15,476 collection passes already run without involving anyone.

Catching a budget drift

Pacing, costs, anomalies

By hand

3 to 7 days

The drift shows up in the monthly report, which is to say too late.

With Pilot

under 24 h

6,805 drifts already caught before anyone went looking for them.

6,805 drifts detected automatically in production.

Changing a campaign

Budgets, bids, statuses, audiences

By hand

10 min

Open the platform, find the campaign, apply the change, check it.

With Pilot

1 click

The agent prepares the change, you approve the validation card.

394 changes proposed, every one through a human approval.

03

The agent doesn't just read: it writes

The AI proposes. A human decides. Without exception.

That's the difference with a reporting tool. Every campaign change goes through an approval card: a human approves, or declines.

0%

of proposals approved

325 approved against 10 rejected, across the 335 decisions settled so far.

0

changes proposed

Budgets, bids, statuses, audiences, creatives.

0%

went through human approval

All 394 proposals produced a validation card. No write without an approval.

0

actions available

The number of tools the agent can call on ad platforms and sources.

04

What users have already done

Not promises, actual use

What real teams have produced on real accounts since the platform opened.

0

conversations with campaigns

That many times a question was put straight to the campaigns, with no spreadsheet in between.

0

dashboards produced

Plus 15 written reports. Revisions happen in the conversation, without starting over.

0

ad accounts monitored

Connected then watched every night, manager account hierarchies included.

0

automatic monitors

Scheduled alerts that flag a drift without anyone having to ask the question.

0

partitioned client spaces

No client ever sees another's data: the separation is enforced at database level.

0 years

of queryable history

Data goes back to 18 July 2024. Year-on-year comparison is immediate.

05

46 connectors, one conversation

Everything the agent can read, and mostly write

Ad platforms, analytics, SEO, CRM, e-commerce, productivity. Behind a single input field.

Google AdsMeta AdsTikTok AdsLinkedIn AdsSnapchat AdsPinterest AdsX / Twitter AdsMicrosoft AdsApple Search AdsAmazon AdsOpenAI AdsDV360Search Ads 360Merchant CenterGA4Search ConsoleGoogle BusinessTag ManagerAdobe AnalyticsPiano AnalyticsMatomoPostHogSemrushDataForSEOAdjustAppsFlyerHubSpotSalesforcePipedriveSellsyShopifyWordPressGmailOutlookGoogle WorkspaceMicrosoft 365DriveOneDriveGitHubVercelGammaAds TransparencyMeta Ads Libraryand 3 more

Method

  • Every figure is pulled from the Pilot production database on 20 September 2026. None is estimated.
  • "Before" durations are industry orders of magnitude, not measurements. "With Pilot" durations are measured on production runs.
  • Volumes include demo and test workspaces: they measure platform activity, not a number of billed clients.
Braike

Let's talk about your project.
No strings attached.

30 minutes to understand your needs, walk you through our tools and approach, and see together whether Braike could be a good partner for you.

The information you provide (name, email, message) is used solely to handle your request and is never resold.

No spam, just a useful conversation.