Specialist worker agents
Assign agents to locations, business units or priorities such as slow stock and rising costs.
EZDash EZDash AI
YOUR BUSINESS CONTEXT
EZDASH AIWork through
Work through
business decisions
with AI.
Explore ideas, compare performance and investigate problems using your data and business context.
YOUR BUSINESS CONTEXTGoals and business rules
Data and comparisons
Findings and supporting evidence
AI ASSISTANT
Discuss ideas and investigate your data.
Explore a sales decline, compare customer groups or test a hunch. Keep refining the question in conversation. The Assistant handles SQL, charts and dashboards.
BRANCH SALES REVIEW
Show me weekly net sales across our three depots.
Define the comparison.
Ask for weekly net sales across three depots, specifying the period and what to check.
Review the chart and explanation.
The Assistant flags part-weeks and holiday weeks that could look like a sales decline.
Save the dashboard.
Approve the chart, then return to Branch Sales Review for the next trading meeting or keep refining it.
AI COMMAND CENTRE
Schedule investigations across your business.
Worker agents investigate their areas. A manager agent combines the findings into a business review with supporting evidence.
Scheduled investigations
Agents examine performance and prepare findings automatically on your schedule.
Follow-up conversations
Discuss the findings, explore another explanation and create files to share with colleagues.
WORKER AGENTSDepot performance Detailed reportStock & demand Detailed report
Costs & operations Detailed report
Depot performance Detailed reportStock & demand Detailed reportCosts & operations Detailed reportMANAGER AGENT
Combined business review
Connects the workers’ findings to identify shared patterns and priorities.
Example: sales missing from product reporting.
SMITHFIELD / BRANCH REVIEW
Which invoice lines are missing from product reports?
Review the dashboard
Review discounts, receivables and unmapped invoice lines in the Branch Control dashboard.
Trace the discrepancy
Compare three depots and trace Smithfield’s unmapped sales to five line descriptions.
Identify the correction
Map the descriptions to product codes, then rerun the margin report. Line names and figures accompany the recommendation.
SALES OUTSIDE PRODUCT REPORTING$31,509
5 line descriptions · 581 invoice lines
Invoices awaiting product codes · data through 30 September 2026CHAT WITH THE REPORT
Create the files your team needs.
Discuss the findings, then ask for a checklist, briefing or presentation to help your team act.
Excel workbook
Product-code checklist
Five line descriptions, sales totals and fields to assign the work.
Download workbookPDF briefing
Manager briefing
The gap, evidence and next steps on one page.
Download briefingPowerPoint deck
Team presentation
Three slides for reviewing the finding and agreeing next steps.
Download slidesComplete file pack
Cleanup pack
Workbook, PDF, slides and CSV checklist in one download.
Download all filesSmithfield’s cleanup pack, created in chat.
Review the checks behind the analysis.
Scheduled reports can include verification of declared figures, data snapshots and reporting periods.
Recalculate figures
Replay declared calculations against extracted data to identify mismatched figures.
Check the inputs
Check for altered analysis files and data outside the reporting period.
Review flagged results
Review flagged calculations and explanations. Failed final checks can warn the reader or block publication, as configured.
Set access and operating limits.
Data permissions
Investigations use the running identity’s dashboard permissions and configured row-level data rules.
Isolated workspace
Calculations and files use a separate workspace. Supported deployments can enforce sandbox restrictions.
Schedules and budgets
Configure investigation schedules and AI spending limits.
Data stored in Sydney
Australian client data is stored in AWS Sydney. Choose AI processing in Australia, or run EZDash and local models on your own infrastructure.
Choose your AI provider and deployment.
OpenAI
Choose Claude or OpenAI. We configure the connection around your data requirements and budget.
Locally hosted models
Kimi K3
GLM 5.3
QwenWe match the model and hardware to your private deployment.
Bring your own API keys
We configure supported providers and validate compatibility. Local inference requires a hosted model and endpoint.
Interactive dashboards
Compare locations, products and customers in views shared by your team and agents.
Explore dashboards & BIOur analysts help you get started.
We set up EZDash around your business priorities and help your team use it.
Discuss your use caseAgree the priorities
Review your goals, data sources and existing Excel reporting.
Configure EZDash
Set up dashboards, agents, business context, schedules and verification.
Train your team
Review the first findings together and practise your own business questions.
First agreed rollout within 30 days. Custom extensions are scoped separately.