AI Products · Built in-house · Run on client accounts

Craft first, AI second.

Our systems do the heavy reading: every live competitor ad, every page of a website, every offer with the date it was seen. What they are built on is the curated learning of experienced marketers who have run these accounts, not the general run of SEO-written internet content. The people make every call that touches your money. These are the tools underneath the work.

Live · Runs inside engagements

Competitor Creative Intelligence

From the ad libraries to a ranked brief directory

Reads every live competitor ad on Meta and Google, works out what the category is arguing, and hands your creative team a ranked queue of briefs with the proposed ad already mocked up.

The first page of a category readout: what the whole category is actually running, with the counts and the findings, category and brand names redacted Category readout · Real round, names redacted
50%est.more winning ads, by betting on concepts the category has already proven
~20%est.less creative-team time on research and briefing
33%est.lower cost per acquisition within two months, at the top of the range
How it works →
Live · Runs inside engagements

Website CRO

Competitor-informed, built on your outcomes

Reads what the best sites in and around your category do, checks each thing against what your site does today, and hands you one ranked sheet of changes with the screenshot beside every row.

A priority row from the Website CRO sheet: a reference pattern beside the client site today, both redrawn as schematics Priority sheet · Real round, captures redrawn
9%lower cost per order for every 10% lift in conversion, at the same ad spend
2xthe odds of a 6%+ conversion gain in a year, structured over ad hoc CRO
How it works →

The documents on these pages are from a live round, set larger for the screen, with the client, the competitors, the category and the evidence lines redacted; the site captures are redrawn as schematics so that no site is identifiable. Numbers marked est. are projections from published benchmarks applied to our first rounds, not yet measured across our clients.

House rules

Three rules every tool has to obey.

Each one exists because the opposite went wrong at least once.

01

Blind until the last step

The reading never knows who the client is until the last step, so it describes the category instead of flattering the brand.

02

Every claim carries its capture

Every finding is stored with a dated screenshot, so two months later the record can still argue back.

03

A second reader checks the work

A separate pass checks every row against the evidence before an experienced marketer reads it again.

Where they fit

They come with the engagement. They are not the engagement.

Both tools run inside our client work, at the point where they change a decision: the creative plan for the quarter, or the next set of website changes. What to do about what they find stays with the experienced professional on your account.

Start a conversation How the whole system works