Signal:
what the world tells you.
The cognitive foundation of everything else you do. Unify fragmented data into a single, AI-readable source of truth that defines your brand's unique DNA.
Every module
Eight ways of listening
Each one is a page. Social Listening and the AEO/GEO Monitor carry the pillar, so start there.
Brand DNA
Identity, voice, vocabulary and channel tone as a specification software can follow, plus your catalogue and documents inside it.
Read moreBrand AI Assistant
Works out what is already answered, asks only about the gaps, and writes it back for the rest of the platform.
Read moreAudience Intelligence
Eight groups from your own orders, with what each is worth, a health flag, written portraits, and a push straight to the ad platforms.
Read moreCompetitor Radar
Suggests the rival set, then tracks messaging, channel mix, share of voice, whitespace and a strengths and weaknesses read, continuously.
Read moreSocial ListeningFlagship
A broad set of sources watched by an agent that runs on its own, with sentiment, topics, share of voice, a written digest, and a filter that shows what it set aside.
Read moreTime & Event Intelligence
Every moment scored against your brand and your region, with the angle, the captions, the shifting festival dates and your own dates too.
Read moreAEO/GEO MonitorFlagship
Whether AI names your brand, why not, what to publish about it, and whether the fix worked.
Read moreTag RadarSupporting
Which tags are rising in your category, read as timing rather than reach. A focused utility, not more.
Read moreSIGNAL module · Brand DNA
Your brand voice, written down properly
Not a PDF brand book that nobody opens. A specification precise enough that software can follow it, and that survives the people who wrote it leaving.
The specification, not the brand book
Illustrative
Voice: plain over clever
Say "we found four problems". Not "we unlocked key learnings".
Never: unlock, seamless, leverage
Words this brand has spent years not using.
Tone: warm 70, formal 30
Set on a scale, with writing guidelines generated from where you put it.
Catalogue: 240 products
So the copy names what you actually sell.
Documents: 12 files read
Guidelines, decks and research kept, rather than re-explained each time.
A rule a person can follow is a rule software can follow. That is the whole test.
Voice, with worked examples
Not adjectives. Each voice attribute carries a do and a do-not written out, so the rule is demonstrated rather than described.
The words you never use
Preferred and forbidden vocabulary, applied to everything the platform writes. If your brand has never once said "unlock", that is not an accident and it should not start now.
Your catalogue and your documents
Upload your product range and the material you already have, guidelines, decks, research. Read once and kept, so nobody re-explains the brand every time.
Checked before it publishes
Every caption, ad and article is screened against regulatory, platform and brand risk before it can go out, and anything doubtful is held for a person.
SIGNAL module · Brand AI Assistant
It reads everything we know before it asks you anything
The opposite of a forty-question brand brief. It works out what is already answered, then asks only about the gaps, and writes the answers back where the rest of the platform can use them.
What it worked out before asking you anything
Illustrative
The two lowest are what it asks about. The rest it already knew, so nobody types it twice.
No question script
It works out what is already known and asks only about the gaps, which is why it finishes instead of being abandoned like a forty-question brief.
Writes back automatically
Answers land in your brand profile where the rest of the platform can use them, rather than in a transcript nobody reads again.
Bring documents mid-conversation
Drop in a deck or a research file part way through and it reads them, rather than making you summarise on its behalf.
Researches when asked
It can go and look something up about your market rather than only asking you, which is the difference between an interview and an assistant.
SIGNAL module · Audience Intelligence
The segments are already in your orders table
Most brands buy an audience tool before using the data they already have. This groups your customers by how they actually buy, and writes a portrait of each group that the creative work is briefed from.
Eight groups from your own orders
Illustrative
Sized by value, not headcount. Each carries a health flag, a written portrait, and a push out to the ad platforms.
The groups, from best to gone
Your customers grouped by how recently, how often and how much they buy, from your strongest through to the ones who have lapsed, each with a written portrait of who they are and what.
What each group is worth
Value and spend per group, so the argument about where the budget goes starts from a number rather than a preference.
A health flag per group
Which groups are healthy and which are draining, so the one quietly leaking revenue does not need somebody to notice it first.
Straight out to the ad platforms
Push a group to your paid and programmatic channels as an audience, with the match rate shown, instead of rebuilding it by hand in each one.
SIGNAL module · Competitor Radar
A competitive deck is out of date the day it lands
Research your competitors continuously instead of once a year, across the places your rivals are actually active, and get told where the gap is rather than just what they did.
Share of the category conversation, this month
Illustrative
Competitor names invented. Tracked run over run, with whitespace and a strengths read attached.
Who your rivals actually are
It proposes the competitor set from your own brand rather than making you guess, then you keep, cut or add to it.
Share of the conversation
How much of the category each of you owns, tracked run over run rather than at a single point, so you see movement and not a snapshot.
Whitespace
The topics and positions nobody in the set is occupying. Usually the most useful page in the report, because it is cheaper to win an empty question than a contested one.
Strengths and weaknesses, both ways
A worked read on each rival and on you against them, rather than a list of what they posted last week.
SIGNAL module · Social Listening
Most listening tools report volume. Volume is the wrong number
A broad set of sources, watched continuously by an agent that runs on its own schedule rather than waiting to be asked.
What the filter kept, and what it set aside
Illustrative
1,240 kept
Actually about your brand, with sentiment and topic attached.
312 set aside: shares your name
A different company with the same word in its name.
96 set aside: wrong industry
Judged against your category before it was ever stored.
54 set aside: bot pattern
The same text repeated across new accounts.
Every one reversible
The reason is shown, and you can overrule it.
Most tools show you the number. The number is only worth something if you can see what it excluded.
Sentiment and topics over time
Movement week by week with the spike explained rather than just marked, and what people actually talk about grouped rather than listed.
Share of the conversation
How much of the category conversation is yours against the competitors the rest of the platform is already tracking.
The words your customers use
Rarely the words in your brand book, and the earliest signal that a market has moved. Available to the people writing, not quoted in a monthly deck.
Gaps against your own voice
Where how people describe you diverges from how your brand says it should sound. Nobody else joins those two up.
Underneath all eight
Your data, from wherever it lives
Nothing here assumes you sell online. Direct connectors for the common sources, a supported file and warehouse route for everything else, and a new integration built where a brand needs one. Every connection is checked against the source the moment you make it.
Connections, checked at the source
Illustrative
The last row is the point. A broken connection is caught while you can still fix it, not found later as a gap in a chart.
Orders and customers
Shopify, Google Analytics 4 and BigQuery directly. A billing export, a distributor file or a spreadsheet of phone numbers works just as well.
Anything else you run
Scheduled file drops over cloud storage or SFTP, or a warehouse table. Supported as a first-class route rather than a workaround.
Checked when you connect it
Credentials tested against the source while you are still on the screen that caused the problem, not three weeks later as a gap in a chart.
Signal module · Time & Event Intelligence
The moment you missed was already on the calendar
Cultural, seasonal, religious, industry, sports and entertainment moments, scored against your brand and your region, with the words to use and the dates tracked for you.
Fit against this brand, this region
Illustrative
23 days out
Peak season scores highest of anything this quarter. Angle and captions drafted, ready to schedule from Studio.
A low score is a useful output. The sporting final is a moment for some brands and not for this one, and the reasoning says why.
Scored 0 to 100, with the reasoning
Against your values, audience, content pillars and voice, and against what you have already run and what competitors are doing. Brand-level, not category-level.
Regional, not national
A moment that matters in one region and not the next is scored that way. Geography is part of the moment rather than a footnote on it.
The dates that move
Most of the calendar that matters lands on a different date every year. Resolved and carried forward rather than re-entered each January.
An angle, and the words for it
A way into the moment that suits your brand rather than the obvious one everyone takes, with captions drafted in your own voice.
A year written out, not shaded in
A narrative twelve-month calendar ranked by fit, plus your own dates, product drops and category weeks sitting alongside the curated ones.
Alerts, then straight into the work
Told while there is still time to make something, then one step to a scheduled post or a paid campaign with the date already set.
Signal module · AEO/GEO Monitor
When someone asks an assistant what to buy
The Monitor tracks how your brand is cited, or missing, in ChatGPT, Perplexity and Gemini answers to the questions your buyers actually ask. Answer Engine Optimization and Generative Engine Optimization are how that visibility gets measured.
The work behind it is content engineering: structuring what you publish so a model can extract it, attribute it and quote it. Princeton's GEO benchmark put the lift from expert quotations at 41% and from cited sources at 40%, which is why this is engineering rather than copywriting.
Share of the answer: how often you are named against competitors
Illustrative
Gap found
You are not cited for "best mid-range skincare for sensitive skin" on Perplexity. Two rivals are. Brief created, ready to publish from Studio.
Whether you are named
Your category questions asked on a schedule across the major assistants and tracked over time, so you have a trend rather than somebody's screenshot.
Why you were not
Which sources the answer drew on, what the brands named instead have published, and how the assistants characterise you against them.
The questions nobody owns
The queries in your category where no brand has a hold yet, which is a cheaper place to win than the ones a rival already occupies.
What to publish about it
A specific brief for the gap, structured to be extracted and attributed rather than merely read, and written against your own brand specification.
Whether the fix worked
The same monitor re-checks the question afterwards, so the loop closes on evidence rather than on somebody saying it is done.
Running on its own
The whole loop runs to a schedule behind one configuration, rather than waiting for someone to remember to check it this month.
Get started today
Ready to know your
brand like never before?
See the product in action. No pitch deck. Just the platform. See exactly how Brand Velocity would work for your brand.
No hard sell. If it's not a fit, we'll tell you.