The AI intelligence layer
Every answer comes from your context. Nothing else.
One shared context built from your own data. Every module reads from it, every agent reasons inside it.
Every answer starts from
Your brandapproved content
Your marketpages read this week
Your pipelinewhat Sales is closing
Your numbersrankings, campaigns, spend
Your budgetthe one you set
Same question. The product knew who was asking.
Ask a generic assistant about a competitor and you get the answer it would give anyone.
A generic assistant
Starts from
Here are five general strategies when a competitor lowers prices. First, consider matching the price. Second...
Intelligence
Starts from
Same question. Completely different answer, because the product knew who was asking.
The intelligence is not in a clever prompt. It is in the context, and the context is yours.
The design rule, not a slogan.
By the context
Produced from what it actually knows.
Every output is produced from what the product actually knows about your business: your brand profile, your approved content, your competitors, your numbers. Never a generic outline.
In the context
Every module sees the same facts.
Every module runs inside the same workspace. Marketing knows what Sales is closing, and a watcher hands its finding to a drafter that already knows your voice. Nothing is re-explained.
For the context
Shaped to your situation, not an average.
Every answer is shaped to your situation, not to an average of everyone’s. The reply is for you, or it is not given.
The context is made of what you choose to put there.
Nothing more.
Your brand
A voice profile from your approved content, scored 0 to 100 against every draft.
Your market
Competitor pages, search positions, what ChatGPT, Claude, Gemini and Perplexity say about you.
Your work
Pipeline, quotes, follow-ups, boards, content in the eight-stage pipeline.
Your numbers
Rankings, campaigns, spend, and the ledger of every run.
Your knowledge base
What you curate for the customer assistant, and only that.
Brand voice, 0 to 100. It quotes the passage it objects to.
You decide the scope
Add a brand and it gets its own context, enforced in code. Leave something out and the product does not pretend to know it.
An algorithm repeats. Intelligence reasons.
An algorithm gives everyone the same answer. A template fills in blanks. Neither knows your business.
Intelligence reads the context, works out what matters for your situation, and writes the answer for you.
When it proposes an article
the proposal carries the market-gap evidence that justified it.
When it explains why a number moved
the causes are computed in code from your own data first, so it never invents a correlation.
When Ask Intelligence answers
it quotes real figures and names its sources, or says the data does not exist.
The behavior we are proudest of
A generic assistant would rather guess than admit a gap. Ours cannot reach outside the context, so it cannot make things up about it.
One gateway carries every model call.
One governed gateway carries every model call, with no shadow routes. We pick the best model for each job, and you can override the choice per agent, tool or module. Tenancy is organization, brand and user, enforced in code.
Every run lands in the ledger: who, what, when, which model, exact cost.
The full version is on Security and data. The money side is on Cost and control.
One governed gateway
Best model for the job. Your pick, per agent, tool or module.
The ledger
Who, what, when, model, cost.
And always a person at the gate.
Nothing publishes, launches or spends without a human approval, and every price is on screen before it runs.
The intelligence proposes. You decide.
Draft · ready for review
Hold for one second. Let go early and nothing ships.
Six things you can check.
Pick the module where your context is thinnest.
The context starts building the day you add the first one.
No subscription. No seat fee. Money in a wallet, spent on tokens, storage and data, and your balance never expires.