Move forward with clarity.
Collaborative AI for the people turning data into decisions.
Work through complex analyses and build plans together, grounded in your metrics and your business context.
Good morning, Maria
Wednesday, August 13 · your briefingAida is AI-native business intelligence.
Analyses, forecasts and plans, built on your metrics and shaped by your team.
Runs on the stack you already have
Connected to your warehouse and your dbt models. Answers come back where you already work: Slack, Teams, Gmail, the app, a document, or through Claude, ChatGPT and your own tools over MCP.
Aida has transformed the way we run our food delivery business. From accounting to identifying new business opportunities.
One task. One analysis. A decision that is yours to make.
Aida does the investigating, comparing and forecasting. The judgement stays with the people who own the outcome. Change the priority in any example and the plan changes with it.
Three things carry it: the new checkout (+€310k), the Outdoor price change (+€190k) and returning customers (+€140k). Paid social is flat.
Two of three. Checkout and pricing are structural. Returning customers depend on the spring cohort’s second order, due in May. Plan on +5%, not +8%.
Revenue vs plan, Q3 · by driver
+€640k above plan, of which €500k is structural. The returning-customer share is the one to watch in May.
Order 1,200 units across four products. Headlamps carry most of the stock-out risk; a second, smaller order in week 5 keeps capital down. That ties up €180k.
Then order 1,600 now. It costs €60k more in stock and takes every product below 5%.
Units to order · next two months
| Product | Order | Stock-out risk |
|---|---|---|
| Tents | 180 | 2% |
| Sleeping bags | 420 | 6% |
| Camp stoves | 260 | 9% |
| Headlamps | 340 | 11% |
Plan at 9% category stock-out risk. Order lead time 4 weeks, sales history 18 months.
Weight search and partnerships. Blended payback drops from 10 to 7 months and growth still clears the target at +12%.
Lean into social: +18% new customers, at 10 months payback and a CAC of €78 instead of €62.
Budget split, Q4 · faster payback
Blended CAC €62, payback 7 months, new customers +12% quarter on quarter.
One zone. DE-NRW is 186% above its 7-day average since Tuesday; every other zone is in range. It matches the courier change on Monday, and it is confined to first orders.
Done. Daily at 08:00 in #ops, with the zone breakdown. I’ll stop when NRW has been under 3% for three days.
Cancellation rate, last 14 days · DE zones
The step on Tuesday coincides with the courier switch in NRW. Refunds in the zone are up 2.3 pts.
41 of 2,180. Twenty-eight wait for headlamps that are back-ordered, nine sit with the NRW courier, four have address problems.
Done. late_delivery_risk is available over MCP, on the same definition the team uses here. Your app gets the order list and the reason.
Orders at risk of late delivery · this week, by reason
Same rule, same definition of “late”, whether a person asks in chat or your app asks over MCP.
Aida turns processes that used to take weeks into minutes and gives us significantly more control over our business.
On demand. And unprompted.
No ticket. No backlog. Just @Aida.
Answers where the question comes up. Mention Aida in a thread and the answer is there before you switch tabs: the number, a chart when it helps, and the sources if anyone wants them. Everyone in the channel sees the same answer.
Aida watches, and tells you first.
Every morning it checks your metrics against your business events and reports what changed, with a verdict.
Aida Discoveries
A chatbot answers you. Aida answers your whole company.
Conversational AI is brilliant for personal, exploratory work. But becoming a data-driven company needs everyone reading the same numbers. That's conversational analytics.
Conversational AI
Conversational Analytics
Your business. One shared understanding.
Every analysis starts from what your company has already agreed on: what the metrics mean, which ways you cut the business, and what happened when. That is why the same question gets a consistent calculation, and why the answer shows its working.
The warehouse you already have. Aida reads it in place; nothing is copied out.
Your models and tests, imported as tables with columns, tags and lineage preserved.
One definition per number, with an owner and a status. Audited ones are what Trusted mode answers from.
The ways you cut the business — market, category, segment — under the names you actually use.
What happened, and when, so a campaign is never mistaken for a trend.
Rules every answer has to pass. Standard ones out of the box; semantic ones you write yourself.
Context Studio
| Table | Source | Status |
|---|---|---|
| orders.fct | Snowflake | Connected |
| customers.dim | Snowflake | Connected |
| shipments.fct | Postgres | Connected |
| Model | Materialization | Last synced | Status |
|---|---|---|---|
| fct_orders | table | 2 min ago | In sync |
| dim_customers | view | 2 min ago | In sync |
| fct_shipments | incremental | 2 min ago | In sync |
| Name | Identifier | Status | Team |
|---|---|---|---|
| Gross margin | gross_margin | Approved | Finance |
| CAC | cac | Approved | Marketing |
| Net revenue | net_revenue | In review | Finance |
| Name | Table | Type | Values |
|---|---|---|---|
| Market | customers.dim | categorical | Nordics · DACH · UK |
| Category | products.dim | categorical | Outdoor · Home · Kids |
| Customer segment | customers.dim | categorical | New · Returning · B2B |
| Start date | End date | What happened? |
|---|---|---|
| 1 Mar 2026 | — | Price increase, 4%, Outdoor |
| 3 Apr 2026 | 17 Apr 2026 | Spring campaign, paid social |
| 12 May 2026 | — | New freight supplier, Nordics |
| Name | Kind | Status |
|---|---|---|
| Audited metrics only in Trusted mode | Standard | Active |
| Never sum a ratio across periods | Semantic | Active |
| Revenue excludes VAT and refunds | Semantic | Active |
Start with one team. Keep your definitions.
It connects to the warehouse and the definitions you already have, so a pilot is a week of setup, not a migration.
Pilot
One teamA subset of your data, the questions that team already asks.
Validate
DepartmentProve the accuracy against your analysts, then expand.
Roll out
OrganizationTop down on metrics, so definitions stay aligned.
Continue with SSO
Your organization uses single sign-on.
You’ll be redirected to your identity provider to finish signing in.
- Single sign-on (SSO)
- Access controlled per person
- Multi or single tenant — you choose
- Encrypted at rest and in transit
- Hosted in the EEA
- Your data is never used to train models