NATI SALAS 🌱 D3FO

🌱 A SNAPSHOT IN TIME

The first thing to keep in mind: the dashboard Cowork builds for us is a snapshot of the information at the exact moment we ask for it. Cowork queries FnO through the connector, processes the data, and returns balances, transactions, and statuses exactly as they stand at that moment.

That makes it an ideal ally for one-off needs: an analysis for a meeting, an audit request, a pre-close review, or that question from the CFO that needs an answer today.

When the need becomes recurring, it’s time to think about which tool fits best and whether it’s worth the investment:

🌿 A Cowork automation, if we need the same report every week or every month and receiving it by email is enough (as we saw in Set it and forget it 🦋).

🌿 A Power BI report, if several people check it daily and need to filter, drill down, and see trends over time, with up-to-date data and a governed model as the single source of truth.

🌿 A Power Apps app, if, beyond viewing the information, we need to record, approve, or act on it.

💡 A tip: Cowork is also a great starting point for a future Power BI report. With the HTML in hand, we can validate with the business what information they need, which calculations, and in what format, so the development that follows is much more straightforward.

🌱 WHY HTML?

Cowork can deliver the result in Word, Excel, PowerPoint, PDF, or HTML. For these examples I chose HTML: a single file holds charts, maps, and interactive filters, we can preview it instantly in the conversation itself, and it opens in any browser.

🌱 WHERE DOES WORK IQ COME IN?

Work IQ is the piece that gives these dashboards their business meaning. While Cowork executes, Work IQ builds the context: it gathers the information, grounds the responses, and chooses which skills and tools to use at each step.

In our examples, we see it at three key moments:

🌿 It understands business data: with Dataverse data available in Microsoft 365 Copilot (we set this up in the installation post), Copilot interprets the relationships between tables and brings in the related records. That way, customer, address, segment, and sales all come together in the same analysis.

🌿 It finds our work context: when the prompt says «use the brand manual», Work IQ finds it in our previous conversations, without us having to tell it where it is, and Cowork applies it to the design.

🌿 It chooses the tools: it selects the html skill to build the page and the FnO connector to query the data.

💡 A tip: the quality of the result goes hand in hand with the quality of the data and how the tables are modeled. That’s why, in the customer map, we ask it to validate the addresses: on top of the dashboard, we get a diagnosis of our master data.

🌱 RECONCILING DATA FROM MULTIPLE TABLES, IN MINUTES

Here comes the part that surprised me the most: I was blown away by how fast Cowork retrieves the information and validates it 😅

In our day-to-day, a report that combines several FnO tables usually means exporting to Excel, cross-referencing data with VLOOKUP, checking that the totals reconcile, and starting over every time something is off. With Cowork, we describe the result we need, and it takes care of the whole journey:

🌿 It queries the necessary entities through the FnO connector.

🌿 It links the information across tables (for example, customers, invoices, and ledger transactions).

🌿 It reconciles the totals and flags any differences, following the validations we specify in the prompt.

🌿 It presents the result in the format we ask for.

And the best part: we can also ask it to validate its own work. To compare the detail against the balance shown in FnO, tell us which table each figure came from, and flag any data still pending confirmation.

💡 A tip: adding a validation instruction to the prompt, for example «verify that the sum of the detail matches the total balance and show me any differences», gives us a ready-to-use report and much more confidence in every number.

Let’s look at two examples so you can see it with your own eyes!

---

Example 3: How to impress the C-suite without spending three days reconciliating entities 

We’ve been asked to prepare a presentation for the C-suite, and we want to make a great impression. We could spend hours exporting customer data, cleaning up addresses (and we know that information isn’t always easy to access), and manually creating maps… or let Cowork do the heavy lifting while we focus on putting together a presentation that looks like we’ve been working on it for weeks. 😉 

With Cowork and the Dynamics 365 Finance & Operations connector, we can analyze customers from various legal entities, review their registered addresses, and turn that data into an interactive map showing where our customers are located. 

Instead of presenting yet another Excel spreadsheet, we can walk into the meeting with a visual overview of our customer base, identify areas with a higher customer concentration, and spot potential opportunities for business expansion. 

Who knows? Maybe this will finally be the presentation that gets us that promotion or pay raise! 😉 

Prompt 

Analyze the geographic distribution and commercial information of all active customers belonging to the legal entities USMF, USP2, USPI, and USPM in Dynamics 365 Finance & Operations. 

For each customer, retrieve: 

• Customer ID and name. 

• Legal entity. 

• Full address: street, city, state or province, postal code, and country. 

• Commercial information: business line, segment, and subsegment. 

• Sales information:  

• Total sales amount. 

• Number of sales orders or invoices. 

• Date of the most recent sale. 

• Sales by period, when the information is available. 

Data Quality and Geocoding 

• Validate the quality of customer addresses. 

• Classify each address as:  

• Valid. 

• Incomplete. 

• Potentially incorrect. 

• Unable to be geocoded. 

Interactive Map 

Create an interactive HTML map focused on commercial and sales analysis where: 

• Each customer is displayed as a geographic marker. 

• The size or color of each marker represents the customer’s total sales amount. 

• Selecting a customer displays their relevant customer, commercial, and sales information. 

• Users can filter the data. 

• Clusters and/or heatmaps can be displayed to identify concentrations of customers and sales. 

Analysis Objectives 

The objective is to identify: 

• Geographic concentrations of customers. 

• States, provinces, or regions with the highest customer presence. 

• Regions with the highest sales volume. 

• Commercial differences by business line, segment, and subsegment. 

• Areas with a high number of customers but low sales volume. 

• Areas with high commercial potential or low market penetration. 

• Address data quality issues. 

Everything in English.

Example 4: 💸 We’re billing more than ever… but are we actually making money? 

The sales team is celebrating: new projects, more customers, and revenue that keeps growing. Everything seems to be going great. But as finance leaders, we know that selling more is one thing, and turning those sales into profitability is quite another. 

Are we maintaining our margins? Are there projects that are losing money? Are budget variances building up, or are there outstanding intercompany transactions that still need to be reconciled? 

Before joining the celebration, we want a clear view of the financial health of the business. And no, we don’t want to spend the entire day gathering data from different modules, reconciling figures, and preparing charts. 

With Cowork and the Dynamics 365 Finance & Operations connector, we can generate a financial health report for USMF in HTML format, with five charts, key indicators, trends, and actionable recommendations, using the actual data available in the ERP, ready to be shared with the rest of the executive team. 

Prompt 

Generate a financial health report for the USMF legal entity using actual data from Dynamics 365 Finance & Operations, delivered as a self-contained HTML page with five charts. 

The goal is to obtain a clear view of the financial position of the business, identify risks, and detect opportunities for improvement. Don’t just show me the numbers: help me understand what is happening and where we, as finance leaders, should focus our attention. 

If any module, data point, or period is unavailable, explicitly state this in the report rather than silently omitting it. Do not invent data or present estimates as if they were actual figures. 

Use data from 2016. 

The 5 Ratios 

1. Current Ratio (Current Assets / Current Liabilities) 
→ Line chart showing the monthly trend over the last 12 months. 

2. EBITDA Margin (%) 
→ Line chart showing the quarterly trend. 

3. Project Margin (%) — PMA 
→ Bar chart ranking active projects from highest to lowest margin. 

4. Aging of Unreconciled Intercompany Transactions 
→ Bar chart by aging bucket: 0–30, 31–60, 61–90, and over 90 days. 

5. Budget Variance (Actual vs. Budget, %) 
→ Bar chart by department or project, depending on the information available. 

Requirements for Each Indicator 

• Highlight the current value and its unit of measurement. 

• Include a status indicator (green, yellow, or red), with consistent and explicit criteria for each indicator. If no corporate threshold has been defined, clearly state the criteria used and do not present them as an official company policy. 

• Add a brief interpretation explaining what the result means for the business. 

• Whenever possible, compare the result with the previous period, budget, or a relevant benchmark. Identify the source of each comparison and do not invent references. 

• If there is insufficient information to calculate an indicator, display it as Not Available and specify which data is missing. 

Deliverable 

Generate a self-contained HTML page that includes: 

1. Executive Summary: Provide 3 to 5 key conclusions, highlighting positive results, risks, and variances that require attention. 

2. Five Charts: One for each indicator, including its current value, status indicator, and interpretation. 

3. Actionable Recommendations: A maximum of five recommendations, prioritized according to their potential impact, with a brief explanation of the issue each recommendation addresses. 

4. Data Limitations: Identify any modules, periods, or calculations that could not be analyzed and explain why. 

Style 

Use an executive, direct, and easy-to-understand tone, avoiding unnecessary technical jargon. Use a professional visual design, with clear and readable charts and a well-defined information hierarchy. Follow the brand guidelines. 

Every number should be presented with context or a comparison whenever sufficient information is available. Clearly distinguish between actual data, estimates, and unavailable data. 

The final report should help me answer three questions: 

1. What is the financial health of the business? 

2. Which risks or variances require attention? 

3. Which actions should we prioritize?