Your dashboards don’t get read
You’ve been paying for a Looker, Tableau or Metabase dashboard for a year. When you ask your team “do you open it?”, half say yes — and the other half are lying. The truth is that few people look at dashboards every day. And those who do look see the same view repeatedly and miss subtle changes.
A conversational dashboard changes that. You ask it in natural language — “how are sales this month?”, “what happened on Tuesday?”, “which campaign has the best ROAS?” — and it responds with real data, relevant charts and narrative.
How we built it
1. Integration with your sources
GA4, Meta Ads, Google Ads, CRM, Shopify, Stripe, your internal DB. We connect what you already have via native APIs. No data migration, no additional warehouse. The dashboard reads where you are.
2. Conversational interface
Web chat, Slack or WhatsApp integration depending on preference. Questions in natural language, answers with text + charts. No SQL, no menu filters. The system understands context: if you ask “what about last week?”, it knows what you were referring to before.
3. Traditional dashboard as a backup
For those who prefer the classic view, we maintain a parallel Metabase or Looker Studio with the core metrics. The conversational does not replace, it complements. Some people think in questions, others in charts.
4. User permissions
The CEO sees full revenue. The head of marketing sees marketing + leads. The sales rep sees only their deals. The AI respects the permissions system you already have in your CRM/ERP.
What types of questions work
- “What were the 5 best deals this month?”
- “Which campaigns have a ROAS below 2.5 over the last 2 weeks?”
- “Compare sales in Bilbao vs Madrid over the last 3 months”.
- “Summarise what happened this week in marketing”.
- “Why did the CTR on Meta drop last Tuesday?”
What changes when this works
- Your team checks data several times a day — not because it has to, because it finds it useful.
- Decisions that used to wait for the “Monday report” are made instantly.
- The head of marketing stops asking the data team “can you pull X for me” — they ask the dashboard instead.
- Faster onboarding for new members: they ask questions in plain language.
When we do NOT recommend this
- If your team prefers Excel and it works well for them: forcing a new tool adds no value.
- If your data lives in systems without an API or with poor quality (empty fields, inconsistencies): the dashboard inherits those problems.
- If your potential query volume is low (<3 people using it): the setup is not worth it.
What it does NOT do
The system responds with data that actually exists in your database. If you ask it something that has no source, it says “I don’t have that data” — it doesn’t invent numbers to seem useful. That’s the difference between a useful dashboard and a dangerous toy.
Stack we use
- Claude 4.5 with tool use for queries.
- Vanna.ai or custom setup depending on complexity.
- Metabase / Looker Studio for classic dashboard.
- BigQuery / ClickHouse / Postgres as data layer.
We start with a diagnostic session
Before quoting the full setup, we run a 90-minute session. We look together at your current sources, your critical metrics and who would check them day to day, and we come out with an honest recommendation: whether this service fits where you are now, or whether it makes more sense to start with something else.
We don’t charge for that session. If you’re interested, get in touch.