Hatu — find and investigate warehouse issues

Rank warehouse issues, diagnose KPI changes, spot breach risk and ask follow-up questions with the data and working in view.

Hatu · At a glance

Prioritise with Watch Tower
See off-track KPIs across sites, ranked worst first.
Understand KPI changes
Review measured contributions, assumptions and gaps behind a result.
Spot breach risk early
See work at risk and where it is waiting between stages.
Keep asking with Ask Hatu
Ask follow-ups, check the working and resume saved chats.

start with the work that needs you

Hatu brings warehouse priorities, KPI diagnosis and plain-language questions into one place. Watch Tower ranks off-track measurements across sites, worst first. Open a KPI to see its target, trend and measured contributions to a change. Early Warning shows breach risk and where open work is waiting, so the team can choose what to review next.

Ask Hatu turns a question about your warehouse data into an answer with the source, query and working to inspect. In Converse mode, ask a follow-up such as “only for that courier” or “compare with last month”. Save the chat and return to the same investigation later.

Ask HatuWhich orders were late this week?Plain-English question
Connected dataplanned_date08 Sepactual_date09 SepsiteDXB-02
SQL checkedactual_date > planned_date✓ 24 matching rows
Answer24 late ordersacross 3 sitesWorking and counted rows shown
QuestionLate ordersDataDispatch and planAnswerSample chartCheckWorking shown

Illustrative sample data

try a warehouse question

Select a question to see the kind of answer Hatu can organise. These figures are examples, not customer records or a live product connection.

Sample answer: 24 orders were late across three sites. Check the planned and actual delivery dates for every counted order.

go deeper when you need to

Open a feature group to see how Hatu supports the next step.

watch and investigate across sites

Watch Tower puts off-track KPIs in worst-first order. Open a diagnosis to inspect component contributions and dated incidents, or save a diagnosis report for a later review. Unknowns and untested explanations remain part of the evidence.

see risk and work waiting between stages

Early Warning compares work at risk with the site’s usual throughput. The process view shows where open records are waiting between stages, helping the team choose which queue to investigate.

define measures around your operation

Use the KPI builder to define a measurement between two recorded moments. The Data Directory brings table meanings, columns, business names and relationships together. Site calendars account for closures, excluded hours and holidays in time-based KPIs.

continue the investigation and see usage

Save and resume Ask Hatu chats with follow-up context. Usage views show credits and investigation activity. Through its MCP connection, Claude Desktop or Claude Code can ask Hatu questions and read KPI, Early Warning and Operations information.

take evidence into the next decision

Move from an off-track KPI to the records and contributing measurements behind it. Diagnosis shows measured contributions, dated incidents and gaps the available data cannot explain. Check the assumptions and bring in the team’s operating knowledge before deciding what to change.

check the explanation, then act

A contribution to a KPI change is not proof of its cause. Breach forecasts are estimates, and an answer still needs its source fields, date range and calculation checked. Hatu supports investigation; people decide what the evidence means for the operation.

Request a Hatu demo or read the guide to natural-language warehouse questions.

Questions people ask

Can Hatu answer follow-up questions?

Yes. Converse mode carries context into follow-up questions. You can save a chat and reopen it to continue the investigation.

Does Hatu explain why a KPI changed?

Its diagnosis shows measured contributions, assumptions and gaps in the evidence. Those contributions are not guaranteed causes; use operating context to judge the explanation.

What does Early Warning show?

It highlights breach risk, records expected to cross a target and work waiting between process stages. Forecasts help prioritise a review; they do not guarantee that a breach will happen or be prevented.

Can we define our own KPIs?

The KPI builder lets you define a measurement between two recorded moments, subject to data and relationship checks. Site working calendars can account for closures, excluded hours and holidays.

Bring us the work that slows you down.

We’ll help you choose a useful first step and keep your team in control.

Request a demo