Making Your Data Work for You: AI & BI
Unlock the power of your business data! Join us as we dive deep into how Artificial Intelligence (AI) and Business Intelligence (BI) can revolutionize the way you access, visualize, and act on your business information. Whether you’re a manufacturer, technology leader, or just starting your data journey, this session is packed with actionable strategies and real-life demos.
Audio-only fan? Relevant Podcast Episodes:
The Executive Perspective: AI & BI | Quick ROI Use-Cases: AI & BI
+ Full Transcript
How to Make Your Business Data Work for You with AI and Business Intelligence
Moderator: Hi, welcome, and thank you for joining us today. Our team is back to discuss artificial intelligence, specifically how organizations can make their business data more useful with AI and business intelligence.
I’m Sue Brizzy, Marketing Director here at 2W Tech. With me today are Mark, our President and CEO; Tom, our CTO; and Alex, our Solutions Engineer.
If you’re not familiar with 2W Tech, we’re a technology solutions and managed services provider with expertise across Epicor ERP, Microsoft technologies, cloud, cybersecurity, business intelligence, data, and AI.
I’ll turn it over to Mark to get us started.
How Can Businesses Get More Value from Their Data?
Mark: Thanks for spending your lunch hour with us.
Today, we’re going to talk about getting access to your business data, improving visibility into that information, using AI to enhance those insights, and automating some of the work that happens around it.
We’ll also show some real-world examples and demonstrations.
The biggest takeaway is understanding how to make your business data more accessible and useful.
We’re going to use a lot of Microsoft technologies in our examples because that’s where our expertise lies, but the broader concepts apply regardless of the specific data and AI platforms you use.
What Are Businesses Asking About AI?
Mark: Alex, you talk to clients all the time who say, “I need to do something with AI, but I’m not quite sure what.”
What are you hearing?
Alex: One of the biggest things I’m seeing is confusion around how to use AI while keeping company data secure.
The other thing is the enormous variety of potential AI use cases.
That could mean querying business processes and standard operating procedures, grounding AI in product information, helping employees with quoting, analyzing business data, or supporting any number of other workflows.
The possibilities are broad, so organizations need to understand where AI actually makes sense for their business.
Why Are Business Data Silos a Problem for AI and Analytics?
Tom: Most organizations have accumulated data across many different systems over the years.
You have engineering information. Manufacturing and operational data. Time and labor information. ERP data. CRM information. Marketing data. Financial information. Excel spreadsheets. Word documents. PDFs. Emails. And plenty of other sources.
Many businesses also have information spread across multiple generations of systems.
For an Epicor customer, for example, data may have accumulated across different ERP versions, applications, reports, databases, integrations, and file repositories over many years.
Add acquisitions into the equation and you’re often bringing an entirely new collection of business systems and data into the organization.
The result is data silos.
Those silos make it harder for employees to find information, combine information from different parts of the business, build consistent reports, and give AI enough context to provide useful answers.
How Can Microsoft Fabric Help Unify Business Data?
Tom: Organizations have been integrating systems for years, but historically a lot of those integrations become one-off projects.
Every time somebody needs a new report or connection, you wind up building another custom integration.
A unified data platform creates another approach.
Within the Microsoft ecosystem, Microsoft Fabric provides an end-to-end data and analytics platform designed to bring information from different sources together for analytics, business intelligence, data engineering, and AI.
Depending on the scenario, data may be replicated into a centralized data environment or accessed through connections to existing sources.
The goal is to reduce unnecessary duplication while creating more consistent access to the business information your reporting, analytics, automation, and AI tools need.
Once those data sources are brought together, tools such as Microsoft Power BI, Microsoft Copilot, machine learning services, data engineering tools, and other applications can work from a more unified source of business information.
What Does a Modern Business Data Platform Do?
Tom: At a high level, the workflow looks something like this.
Data comes in from different business systems. Some information may be stored or replicated, while other information may remain connected to the original source.
Then the organization can prepare, analyze, transform, and model that information.
From there, the data can support machine learning, AI, dashboards, reporting, historical analysis, real-time analytics, and other business use cases.
Finally, those insights need to reach the people who can actually use them.
That could happen through Microsoft Teams, Power BI reports, dashboards, alerts, business applications, or other workflows.
Do You Need AI or Do You Really Need a Dashboard?
Mark: Here’s something we see all the time.
Somebody says, “I want AI.”
And after we dig into the business problem, it turns out what they actually need is a good dashboard.
Alex: Exactly.
Sometimes you need AI. Sometimes you simply need a better way to see and organize your data instead of working from a giant Excel spreadsheet.
Visualization is often a great place to start.
How Can Manufacturers Use Power BI?
Alex: Microsoft Power BI can help manufacturers turn ERP and other business data into visual dashboards that are easier to understand and interact with.
For example, an executive dashboard might show open receivables, open payables, open orders, invoiced amounts, shipped orders, booked orders, and other high-level financial and operational metrics.
From there, users can drill into more specific information.
Maybe you want to understand accounts receivable performance by customer, time period, company, category, or another business dimension.
Power BI gives users the ability to filter and interact with the data rather than simply viewing a static report.
Manufacturers can also visualize geographic sales data, customer concentration, margins, production performance, and many other business metrics.
How Can Power BI Help You Explore Business Data?
Alex: One visualization I particularly like is the decomposition tree.
Imagine starting with total margin.
You can break that margin down by salesperson, then customer, then production group, then individual part number.
That gives users an interactive way to explore what’s driving a business result instead of trying to manually work through a spreadsheet.
It’s essentially a choose-your-own-adventure approach to business data.
How Can Power BI Alerts Turn Data into Action?
Alex: Visualization is useful, but you don’t necessarily want employees staring at a dashboard all day waiting for something to change.
Alerts can help shift reporting from passive monitoring toward proactive action.
For example, let’s say your productivity goal is 80%.
You could configure an alert so that when productivity falls below the defined threshold, the appropriate people receive a notification.
That gives them an opportunity to investigate what changed and take action without constantly monitoring the report.
Mark: And that’s really where the value is.
The happy graphs are nice, but the exceptions are often what matter most.
You want your data to tell you when something requires attention.
How Can Copilot Let You Ask Questions About Your Business Data?
Alex: Visualization is one piece. The next step is interacting with business data using natural language.
Instead of manually creating every filter or report, you can ask questions using normal conversational language.
For example: “Give me a summary of what happened last month.”
From there, you might ask: “What trends do you see?”
AI can help identify things such as top customers, changes in sales, high-performing product groups, variations between periods, or other patterns worth investigating.
You can then continue the conversation and ask follow-up questions.
That’s what makes this powerful: you’re beginning to have a conversation with your business data.
AI doesn’t eliminate the need for human analysis. Instead, it can help employees explore large amounts of information more quickly and identify areas worth investigating further.
Why Does AI Need Clean, Connected Business Data?
Tom: The speed and quality of these experiences depend heavily on the underlying data architecture.
When business information is organized, available, and properly modeled within a modern data platform, analytics and AI tools can work much more effectively.
That becomes particularly important when you’re dealing with years of history and millions of rows of transactional data.
AI can only work with the information it can appropriately access and understand.
What Is Resolve IQ?
Alex: Resolve IQ is an AI platform designed to help organizations interact with business-specific information and workflows.
It can be connected to approved business data and content sources so employees can ask questions, retrieve information, and support business processes using natural language.
Instead of relying only on a general-purpose AI model, the goal is to provide relevant organizational context.
That might include company documentation, Microsoft 365 content, structured business data, process information, or other approved sources.
How Can Business-Specific AI Help Create Documents?
Alex: Here’s one example.
Let’s say I need to create a statement of work for a mail migration.
I can ask the AI for assistance. Because a request that simply says “I need an SOW for a mail migration” doesn’t contain enough information, the system can ask follow-up questions to gather the additional details necessary to produce a useful result.
Once those questions are answered, it can use approved company information, established formats, and the provided project details to produce a first draft.
Instead of searching through hundreds of previous documents trying to find one that looks similar, employees can work from an AI-assisted draft grounded in the organization’s own information.
That can save time while helping maintain greater consistency across documents and business processes.
How Can AI Agents Automate Business Processes?
Tom: Natural-language querying and summarization are useful, but AI becomes even more interesting when you begin combining it with automation.
Think about what happens when an AI agent can interpret an incoming request and then take predefined actions based on the information it receives.
At 2W Tech, one example involves customer service and help desk tickets.
A user can send an email describing a technical problem.
An AI-driven workflow can analyze the request, summarize the issue, identify relevant information, help determine the required skills or priority, route the ticket, notify the appropriate people, and provide the requester with an initial response.
Depending on the workflow, it can also suggest initial troubleshooting steps while the customer waits for additional help.
Multiple parts of that process can happen automatically within seconds of the original request.
That’s where AI starts moving beyond “talking to your data” and into business process automation.
Where Can Manufacturers Use AI Agents?
Tom: Our ticketing workflow is just one example.
Manufacturers should look for processes that currently involve repeatable manual steps.
Maybe something happens when a manufacturing job closes. Maybe an order comes in. Maybe a quality issue is reported. Maybe customer service receives a request. Maybe data needs to move between an ERP, CRM, MES, or another business application.
Ask: Which parts of this process require a person to repeatedly interpret information, route work, assign responsibility, create notifications, or enter data?
Those can become candidates for AI-assisted automation.
Why Are Real-Time Business Insights Important?
Mark: Real-time and near-real-time insight is one of the biggest opportunities here.
No matter how involved you are in your company, there eventually comes a point where you can’t personally know what’s happening at every desk, in every department, and across every business system.
Modern analytics and AI can help monitor large amounts of information and distill it into the context required to make a decision.
The goal isn’t more data.
The goal is getting the right information to the right people when they can still do something with it.
Why Does AI Need to Fit Your Business Processes?
Alex: Customization is incredibly important.
Every organization operates a little differently.
The goal shouldn’t be forcing your business to completely change the way it operates just to fit an AI tool.
The better approach is identifying the business process, determining what employees need, and configuring the technology around that use case when possible.
Your reports, dashboards, AI experiences, and automations should reflect how your organization actually works.
How Do You Keep Business Data Secure When Using AI?
Alex: Security starts with understanding where your information lives, which AI services can access it, and who has permission to see it.
Business data should remain subject to appropriate identity, access, security, compliance, and governance controls.
Within Microsoft environments, permissions established through technologies such as Microsoft Entra ID, SharePoint, and related Microsoft 365 security controls can play an important role in determining which users can access specific information.
Financial information shouldn’t suddenly become visible to employees who aren’t authorized to see financial information just because you’re using AI.
The same applies to HR records, customer information, intellectual property, and other sensitive business data.
How Do You Get Business Data Ready for AI?
Tom: This is an important step that organizations sometimes underestimate.
Turning on Microsoft Copilot or another AI product doesn’t automatically make your data ready for AI.
Organizations may have decades of documents sitting across shared drives, old folders, inconsistent repositories, and business applications.
Some information may be duplicated. Some may be obsolete. And some may have inappropriate permissions.
Before scaling AI, organizations should review where important information lives, who has access, which versions are authoritative, and how sensitive information is classified and protected.
SharePoint can provide a modern environment for document management within Microsoft 365, while Microsoft Purview provides capabilities around information protection, data governance, classification, and data loss prevention.
The goal is to create data discipline before allowing AI to surface information faster than employees could find it manually.
Why Are Permissions So Important Before Deploying AI?
Tom: Here’s a simple example.
Imagine your company has an old public shared drive that everybody can access.
Buried somewhere inside that drive is a sensitive document that employees technically have permission to open, even though they probably shouldn’t.
Before AI, that document might have been difficult enough to find that nobody noticed the permissions problem.
AI can make information dramatically easier to discover.
That’s why AI readiness has to include permissions, identity, data governance, information protection, and access reviews.
AI may expose an existing permissions problem. It doesn’t necessarily create that problem.
How Should a Business Start Using AI?
Alex: AI without a purpose can turn into a party trick pretty quickly.
Don’t start with “AI everywhere.”
Start with a small, well-defined business task.
Ask yourself: What process is too manual? Where are employees losing time? Where is information difficult to find? What repeatable task could be faster or more consistent?
Those are better starting points than choosing a large AI platform first and then trying to figure out what to do with it.
AI, Business Intelligence, and Microsoft Fabric: Webinar Q&A
What Manufacturing AI Use Cases Can Deliver Quick ROI?
Alex: One relatively straightforward opportunity is querying standard operating procedures and process documentation.
Imagine an employee encounters a nonconforming part and needs to know your organization’s approved process for handling it.
Traditionally, they may ask another employee or search through PDFs, Word documents, binders, or shared folders.
An AI tool grounded in your approved SOPs can help employees find that information more quickly while providing an answer based on your organization’s documented processes rather than a generic answer from the internet.
Another useful example is product documentation.
Customer service employees could ask questions about a machine or product and retrieve relevant information without manually searching through long product manuals and technical documents.
Tom: Supply chain analytics is another area where manufacturers can potentially identify relatively quick value.
Many organizations already have useful purchasing, inventory, and production data available. Improving the analytics around inventory levels, purchasing behavior, production performance, and related processes can reveal opportunities for improvement.
Mark: And don’t overlook complicated spreadsheets.
Almost every organization has spreadsheets that have become unofficial business systems because the information doesn’t fit neatly into an existing ERP or application.
Those spreadsheets can be strong candidates for better business intelligence, data integration, and eventually AI-assisted analysis.
How Can Businesses Use a Secure AI Chat Experience?
Alex: Organizations often want the convenience of a conversational AI experience while maintaining stronger control over business data.
The key is choosing an enterprise AI architecture where your organization’s security, identity, data-handling, and governance requirements can be properly applied.
Different AI models may be useful for different tasks, but the model is only one piece of the architecture.
Organizations should evaluate where data is processed, whether prompts or outputs are retained, whether data is used for model training, which users have access, and what organizational security policies apply before sensitive business information is introduced.
Can AI Make Epicor BAQ Creation Easier?
Tom: AI can be useful when building queries and helping users work through data structures.
For Epicor users, that could include helping developers or analysts better understand tables, relationships, query logic, SQL concepts, and the structure needed to create Business Activity Queries, or BAQs.
As with any AI-generated technical output, the resulting query should still be reviewed and tested before being used in production.
Why Do I Need a Unified Data Platform If I Already Have ERP and CRM?
Tom: The fact that you already have an ERP and CRM is actually part of the reason.
You already have multiple important business systems producing different types of data.
Add finance, manufacturing systems, quality applications, spreadsheets, Microsoft 365, customer service platforms, and other sources, and the challenge becomes bringing that information together.
A unified data platform can give analytics, automation, and AI tools a more consistent way to access information across those different systems.
That’s what creates the opportunity to move from simply reporting on individual applications toward asking broader questions about the business.
How Does Power BI Work with Epicor ERP?
Alex: Power BI can be connected to Epicor data to create dashboards and reports covering financial, sales, production, inventory, customer, and other ERP information.
The exact integration depends on your Epicor environment, data architecture, security requirements, and reporting goals.
Once the connection and data model are established, organizations can build reusable dashboards while still creating custom reports around business-specific requirements.
Final Takeaway: Make Your Data Ready Before You Scale AI
The path to better AI and business intelligence starts with making your data useful.
Break down unnecessary data silos. Establish trusted sources. Organize and secure your information. Connect business systems where needed. Give employees useful visualizations. Then introduce AI where it solves a clearly defined problem.
The goal isn’t AI for the sake of AI.
It’s creating an environment where employees can find answers faster, identify important changes, automate repetitive work, and make better decisions using the information your business already has.
Moderator: Thanks, everybody, for joining us today. If you have additional questions about AI, business intelligence, Microsoft Fabric, Power BI, or your broader data strategy, reach out to the 2W Tech team.