The Agentic ERP Series
How AI Agents Will Transform Microsoft Dynamics 365 Business Central
Chapter 4 of 6
Could the Sales Agent Become Your First Revenue Growth Adviser?

In previous chapters, we explored how AI Agents could transform ERP interaction, strengthen financial decision-making and help procurement teams anticipate supply chain risks.

However, there is another area where the potential impact may be even greater.

Sales.

Because whilst Finance protects profitability and Procurement protects supply, Sales is responsible for something fundamentally different.

Generating growth.

And growth rarely depends on having more information.

More often, it depends on identifying opportunities before they become obvious.

This is where the Sales Agent becomes particularly interesting.

Not because it automates selling.

But because it could help organisations recognise commercial signals that humans often overlook.

The Future of Sales Is Not More Data
It Is Better Interpretation

Most organisations already possess vast amounts of commercial information.

Microsoft Dynamics 365 Business Central contains:

• customer histories;
• sales orders;
• pricing agreements;
• margins;
• purchasing trends;
• product preferences;
• seasonal patterns.

The challenge has never been data availability.

The challenge has always been turning information into actionable insight.

Many organisations know what happened.

Far fewer understand what is about to happen.

And that distinction may become a significant competitive advantage.

Business Scenario 1
The Customer Who Is Quietly Leaving

Imagine an industrial distributor whose largest customer generates approximately £400,000 of annual revenue.

Everything appears normal.

The customer is still active.

Orders continue to arrive.

No complaints have been raised.

No opportunities have been marked as lost.

Yet something has changed.

Over recent months:

• order frequency has declined;
• average order value has decreased;
• certain recurring products have disappeared;
• purchase intervals have become longer.

Individually, none of these indicators seems alarming.

Collectively, they form a pattern.

A Sales Agent could identify that pattern and highlight:

This customer exhibits behavioural signals consistent with a gradual decline in commercial engagement over the next ninety days.

Suddenly, the conversation changes.

The priority is no longer generating new business.

The priority becomes protecting existing business.

And retaining revenue is often more valuable than acquiring it.

Business Scenario 2
The Opportunity Nobody Is Looking For

Consider a technology provider.

A customer regularly purchases:

• hardware;
• infrastructure upgrades;
• replacement equipment.

However, the same customer never purchases:

• support services;
• training;
• managed services;
• maintenance contracts.

Traditionally, this opportunity may remain invisible.

Nobody actively searches for it.

The Sales Agent, however, can compare commercial behaviours across similar customer profiles.

It may discover:

Customers with similar characteristics typically purchase 65% more associated services than this account currently consumes.

The result is not a new report.

It is a new conversation.

An opportunity existed all along.

The organisation simply lacked the visibility required to recognise it.

Business Scenario 3
Revenue Does Not Always Mean Growth

Many organisations still evaluate commercial success using a relatively simple metric:

Revenue.

The problem is that not all revenue creates the same value.

Consider two customers.

Customer A

• Generates £500,000 annually.
• Demands significant discounts.
• Requires extensive support.
• Creates operational complexity.

Customer B

• Generates £300,000 annually.
• Buys consistently.
• Maintains healthy margins.
• Requires minimal intervention.

Which customer contributes more to long-term growth?

The answer is not always obvious.

A Sales Agent can help uncover relationships between revenue, profitability, effort and risk.

The discussion shifts from:

How much are we selling?

to:

How much value are we creating?

And that is a far more strategic question.

Business Scenario 4
The Pipeline That Looks Healthy

Every sales organisation has experienced this.

The pipeline appears strong.

Forecasts are optimistic.

Targets seem achievable.

Yet quarter-end arrives and expectations are missed.

Why?

Because opportunities often remain in the pipeline long after momentum has disappeared.

A Sales Agent could detect signals such as:

• stalled activity;
• declining customer engagement;
• missed milestones;
• unusual sales cycle duration;
• behavioural deviations from successful opportunities.

The Agent might conclude:

The actual probability of closure is significantly lower than currently forecast.

This fundamentally changes commercial planning.

Hope is replaced by evidence.

Assumptions are replaced by analysis.

And decision-making becomes significantly more reliable.

Sales Is Evolving
From Managing Opportunities to Managing Signals

Perhaps the most important transformation is not technological.

It is conceptual.

For decades, sales teams focused on managing opportunities.

AI Agents introduce the possibility of managing signals.

This distinction matters.

Because opportunities rarely appear out of nowhere.

They begin as behaviours.

Behaviours become patterns.

Patterns become opportunities.

Organisations capable of interpreting those patterns early gain a significant advantage.

In many cases, they gain it before competitors even realise an opportunity exists.

What Could This Mean for Business Central?

Business Central already contains an extraordinary amount of commercial knowledge.

Customer behaviour.

Sales history.

Order patterns.

Pricing information.

Product relationships.

Margin data.

The challenge is not collecting more information.

The challenge is understanding what that information is trying to tell us.

Sales Agents could help bridge that gap.

Not by replacing sales professionals.

Not by replacing commercial judgement.

But by helping teams identify opportunities, risks and patterns that might otherwise remain hidden.

What Does This Mean for Functional Consultants?

Historically, functional consultants helped organisations define:

• pricing structures;
• discount policies;
• sales processes;
• approval workflows;
• reporting requirements.

These activities remain important.

However, Agentic ERP introduces a new responsibility.

Helping organisations define commercial intelligence.

Someone must determine:

• what constitutes a sales opportunity;
• which signals indicate customer risk;
• when recommendations should be generated;
• how opportunities should be prioritised;
• which actions require human intervention.

This is not primarily a technical challenge.

It is a business challenge.

Because AI Agents rely on business knowledge far more than technical complexity.

And business knowledge remains one of the most valuable assets any organisation possesses.

To Conclude

For years, organisations believed growth depended on finding more opportunities.

AI Agents introduce a different possibility.

Finding the right opportunities sooner.

Because growth often begins with subtle signals that remain invisible until it is too late.

Sales Agents have the potential to help organisations recognise those signals earlier, act more intelligently and focus effort where it creates the greatest impact.

Not as replacements for sales professionals.

But as specialised advisers capable of transforming commercial data into commercial intelligence.

And that may prove to be one of the most valuable applications of AI within Business Central.

Coming Next…
Chapter 5 of 6
The Customer Service Agent

Organisations once competed through products.

Then they competed through prices.

Today, they increasingly compete through customer experience.

But what happens when AI Agents begin identifying dissatisfied customers before a complaint is raised?

What if customer service teams could anticipate churn before relationships begin to deteriorate?

What if Business Central could help organisations protect loyalty before it is lost?

In the next chapter, we will explore how the Customer Service Agent could become the first digital guardian of customer trust.

Because if Sales can identify opportunity…

Customer Service may learn how to protect it.

The Agentic ERP Series

How AI Agents Will Transform Microsoft Dynamics 365 Business Central
Chapter 3 of 6
Will the Procurement Agent Become Your First Supply Chain Risk Analyst?

In the previous chapters, we explored how AI Agents could transform the way people interact with ERP systems and how Finance may become one of the first business functions to benefit from this evolution.

However, there is another area where the impact may prove equally significant.

Procurement.

Because whilst Finance focuses on understanding what has already happened, Procurement is responsible for preventing what has not happened yet.

And that distinction changes everything.

Most supply chain failures do not occur because organisations lack information. They occur because critical signals remain unnoticed until action becomes difficult, expensive or impossible.

This is where the Procurement Agent becomes particularly interesting.

Supply Chains Are Becoming More Complex

And Complexity Creates Risk

During the last decade, supply chains have experienced unprecedented disruption.

Organisations have been forced to deal with:

• raw material shortages;
• transportation bottlenecks;
• rising energy costs;
• supplier insolvencies;
• labour shortages;
• geopolitical instability;
• inflationary pressure.

As a result, procurement is no longer simply a purchasing function.

Increasingly, it is becoming a risk management function.

The challenge is that most risks do not arrive with a warning label.

They emerge gradually through dozens of small signals that individually appear harmless.

The question is no longer:

How do we purchase more efficiently?

The question has become:

How do we identify risk before it becomes operational disruption?

Business Scenario 1

The Supplier That Appeared Reliable

Imagine a manufacturing company that sources 40% of its critical components from a single supplier.

For years, everything has worked smoothly.

Purchase orders arrive on time.

Quality remains acceptable.

Business relationships appear stable.

However, over the last two months:

• delivery delays increase slightly;
• responsiveness decreases;
• order amendments become more frequent;
• service levels deteriorate incrementally.

None of these signals trigger immediate concern.

Individually, they appear insignificant.

Collectively, they form a pattern.

A procurement professional may not detect that pattern whilst reviewing dozens of suppliers and hundreds of transactions.

An Agent could.

It could identify correlations and highlight:

A progressive decline in supplier reliability has been detected across the last eight weeks. Current trends indicate an increased probability of fulfilment issues within the next month.

The value is not automation.

The value is visibility.

And visibility creates options.

Business Scenario 2
The Inventory Risk Nobody Saw Coming

One of the most common questions within supply chain operations is:

Do we have sufficient inventory?

Yet this may not be the most useful question.

A more powerful question could be:

Which products are most likely to experience stock shortages during the next three weeks?

The difference is subtle but significant.

The first question describes the present.

The second evaluates the future.

A Procurement Agent could analyse:

• current inventory levels;
• historical demand patterns;
• open purchase orders;
• supplier lead times;
• planning forecasts;
• seasonal fluctuations.

Rather than waiting for shortages to occur, the organisation gains early visibility into emerging risk.

The conversation moves from reaction to prevention.

Business Scenario 3

The Silent Cost Increase

Cost inflation rarely arrives overnight.

More often, it appears through a series of small increases.

Two percent this quarter.

Three percent the next.

A revised surcharge.

A transportation adjustment.

Individually these changes attract little attention.

Collectively they can significantly affect profitability.

Imagine asking:

Which suppliers have increased pricing beyond market benchmarks over the last six months?

A Procurement Agent could:

• identify abnormal trends;
• detect category-specific anomalies;
• highlight supplier concentration risks;
• prioritise renegotiation opportunities.

Once again, the objective is not simply information.

The objective is informed decision-making.

Business Scenario 4
The Most Dangerous Risk Is Often Invisible

Some risks remain hidden precisely because they have never caused problems.

Supplier dependency is one of them.

Many organisations discover too late that a critical material, component or service depends almost entirely on a single supplier.

Everything appears stable until disruption occurs.

Imagine asking:

Which critical materials depend on a single supplier?

And then:

What would be the operational impact if that supplier stopped delivering for thirty days?

Traditionally, this exercise could require significant analysis.

An Agent could perform the assessment in minutes.

The real benefit is not speed.

The real benefit is enabling strategic conversations before disruption occurs.

Procurement Is Evolving

From Operational Execution to Predictive Intelligence

A noticeable shift is emerging across procurement functions.

Less time is being spent on:

• processing transactions;
• reviewing purchase orders;
• locating information;
• generating reports.

More time is being spent on:

• supplier resilience;
• risk management;
• strategic sourcing;
• operational continuity;
• scenario planning.

AI Agents accelerate this evolution.

By continuously analysing information across Business Central, they allow procurement teams to focus on judgement rather than administration.

This is not about replacing procurement professionals.

It is about expanding their ability to anticipate.

What Could This Mean for Business Central?

Business Central already contains much of the information required to understand supply chain health.

Organisations manage:

• suppliers;
• inventory;
• purchase orders;
• receipts;
• planning parameters;
• historical transactions;
• replenishment strategies.

The challenge is not access to data.

The challenge is transforming operational data into actionable intelligence.

Procurement Agents could help bridge that gap.

Not by replacing existing processes.

But by helping organisations recognise patterns that would otherwise remain hidden.

And in today’s environment, recognising risk early may be one of the most valuable competitive advantages available.

What Does This Mean for Functional Consultants?

Historically, functional consultants helped organisations define:

• replenishment policies;
• inventory strategies;
• purchasing processes;
• supplier management frameworks;
• planning methodologies.

These responsibilities remain essential.

However, a new responsibility is emerging.

Defining how the system should think.

Someone must determine:

• what constitutes procurement risk;
• which patterns should trigger alerts;
• what actions should be recommended;
• what decisions require human approval;
• how organisational knowledge should guide recommendations.

This is not fundamentally a technical challenge.

It is a business challenge.

And business understanding remains the foundation of every successful AI initiative.

To Conclude

For years, procurement teams focused on securing supply.

The next evolution may focus on anticipating disruption.

Because the real competitive advantage will not come from responding faster to problems.

It will come from identifying them before they materialise.

That is why Procurement Agents are so interesting.

Not because they automate purchasing.

But because they may become permanent supply chain observers, capable of analysing risk, identifying patterns and providing context at a scale that would be difficult for any human team to achieve alone.

And in an increasingly unpredictable world, that capability may prove invaluable.

Coming Next…

Chapter 4 of 6
The Sales Agent

What happens when sales teams stop looking for opportunities and opportunities start looking for them?

Imagine an Agent capable of continuously analysing:

• customer behaviour;
• buying patterns;
• churn indicators;
• cross-selling opportunities;
• revenue risk;
• pipeline performance.

Could Business Central help identify opportunities before competitors do?

Could AI Agents become trusted commercial advisers capable of guiding growth decisions?

Next week, we will explore how AI Agents could transform one of the most visible functions in any organisation:

Sales.

Because if Procurement can anticipate risk…

Sales may learn to anticipate opportunity.

The Agentic ERP Series

How AI Agents Will Transform Microsoft Dynamics 365 Business Central

Chapter 2 of 6

Will the Finance Agent Become the CFO’s First Digital Colleague?

In the previous chapter, I explored a provocative idea:

What happens when people stop navigating ERP systems and start interacting with them through intent and conversation?

If that transition becomes reality, an important question immediately follows:

Which business function will benefit first?

In my view, the answer is clear.

Finance.

Not because Finance is the most technologically advanced department.

Not because Finance has the largest amount of data.

But because Finance spends more time than almost any other business function converting information into decisions.

Every day, finance professionals are expected to answer questions that can fundamentally affect the future of an organisation:

• Are we carrying too much risk?
• Will we have sufficient liquidity next quarter?
• Which customers represent a potential credit exposure?
• Where is profitability beginning to deteriorate?
• What financial trends require immediate attention?

The challenge has never been a lack of data.

The challenge has always been turning data into actionable insight.

The CFO Does Not Need More Reports

The CFO Needs More Context

For years, organisations have invested heavily in analytical capabilities.

We built:

• ERP reports;
• financial dashboards;
• Power BI solutions;
• forecasting models;
• management reporting packs.

These investments created extraordinary visibility.

Yet many Finance Directors still face the same frustration.

They spend considerable time gathering information before they can begin analysing it.

The question is no longer:

Can we access the data?

The question has become:

Can we understand what the data is trying to tell us?

This distinction is crucial.

Reports provide information.

Context provides understanding.

And understanding drives decisions.

This is where the Finance Agent becomes particularly interesting.

Scenario One

Understanding Credit Risk Before It Becomes a Problem

Imagine a Monday morning executive committee meeting.

The Chief Executive asks a simple question:

What is our most significant financial risk over the next thirty days?

The response appears straightforward.

In reality, answering accurately may require:

• reviewing customer ledgers;
• analysing overdue balances;
• evaluating concentration risk;
• examining payment behaviour;
• reviewing historical trends;
• assessing future obligations.

The information exists inside Business Central.

The challenge is assembling it into a meaningful conclusion.

Now imagine a Finance Agent receiving the following request:

Which customers represent 80% of our projected credit exposure during the next thirty days, and how has their payment behaviour evolved over the past two quarters?

Instead of generating a collection of reports, the Agent can:

• identify priority customers;
• calculate exposure levels;
• compare payment trends;
• measure concentration risk;
• explain why specific accounts require attention.

The Finance Director remains responsible for the decision.

The Agent simply removes the effort required to build the analysis.

This leads to an important reflection.

For many organisations, finance professionals spend more time gathering evidence than evaluating it.

What happens when that balance changes?

Scenario Two

From Cash Flow Reporting to Cash Flow Foresight

Cash flow management has always been one of the most critical responsibilities of any finance department.

Every organisation wants to answer a seemingly simple question:

Will we have enough liquidity to support our commitments?

The answer is rarely simple.

To produce a credible forecast, finance teams must consider:

• expected receipts;
• supplier obligations;
• outstanding invoices;
• purchasing commitments;
• seasonal trends;
• historical behaviour.

Today, this often involves exporting data, consolidating spreadsheets and manually reviewing multiple sources of information.

A Finance Agent introduces a different approach.

Imagine asking:

Which weeks present the highest liquidity risk during the next quarter and what factors are contributing to it?

Instead of returning a spreadsheet, the Agent could explain:

• when the risk emerges;
• which customers affect cash inflows;
• which supplier payments create pressure;
• where exposure is concentrated;
• which mitigating actions may be available.

The value is no longer reporting.

The value becomes anticipation.

And in finance, anticipation is often the difference between control and reaction.

Scenario Three

Detecting What Nobody Is Looking For

This may ultimately become the most powerful capability of all.

Traditional reports answer questions.

Agents can identify questions that nobody has asked yet.

Think about how many financial risks develop gradually:

• margins decline incrementally;
• costs increase progressively;
• payment behaviour deteriorates slowly;
• cash flow pressures emerge over time.

These issues rarely appear as dramatic events.

They develop quietly.

Often until they become significant enough to demand attention.

A Finance Agent could continuously evaluate financial patterns and proactively highlight concerns such as:

A sustained reduction in profitability has been detected across a specific business unit during the past six weeks.

Or:

Customer payment behaviour has deteriorated by 18% compared with the previous quarter.

Or:

Inventory holding costs have increased above historical norms, creating additional working capital pressure.

In these situations, finance leaders receive more than information.

They receive interpretation.

And that changes the nature of financial management.

The Future Is Not About Automating Reports

One misconception appears repeatedly whenever organisations discuss AI.

Many people assume Agents are simply a new generation of dashboards.

I believe this significantly underestimates their potential.

A dashboard answers known questions.

An Agent can help uncover unknown questions.

That distinction matters.

The greatest business risks rarely originate from problems we already understand.

They usually emerge from risks we have not yet identified.

This is why Agentic Finance represents something larger than automation.

It represents the possibility of augmenting financial judgement.

Not replacing it.

Strengthening it.

What Could This Mean for Business Central?

Business Central already provides an exceptionally strong financial foundation.

Finance teams can access:

• ledgers;
• dimensions;
• budgets;
• cash flow forecasts;
• receivables;
• payables;
• fixed assets;
• financial reporting.

The challenge has never been capability.

The challenge has been accessibility and interpretation.

Finance Agents have the potential to bridge that gap.

Not by replacing Business Central.

But by amplifying its value.

By transforming financial information into meaningful business conversations.

And by making insight available when it is needed most.

What Does This Mean for Functional Consultants?

This transformation has profound implications for functional professionals.

Historically, consultants helped organisations define:

• chart of accounts structures;
• posting configurations;
• reporting requirements;
• financial processes;
• operational controls.

These responsibilities remain essential.

However, Agentic Finance introduces a new layer.

Finance Agents need guidance.

They need boundaries.

They need trusted knowledge.

Someone must define:

• which metrics matter;
• which risks should trigger alerts;
• which recommendations are appropriate;
• which actions are allowed;
• which governance principles must always be respected.

This is not purely a technical challenge.

It is a business challenge.

And business challenges are precisely where functional consultants create the greatest value.

The future consultant may spend less time designing reports.

And significantly more time designing financial intelligence.

To Conclude

For years, organisations asked:

How can we obtain more financial information?

Perhaps the next strategic question is different.

How can we obtain better financial judgement?

Because the ultimate objective is not access to data.

The ultimate objective is confidence in decisions.

Finance Agents will not replace CFOs.

They will not replace Finance Directors.

They will not replace financial expertise.

But they may become the first digital colleagues capable of continuously analysing, contextualising and prioritising financial information at a scale no human team could realistically achieve alone.

And that possibility is too important to ignore.

Coming Next…

Chapter 3 of 6

The Procurement Agent

What happens when purchasing teams stop reacting and start anticipating?

Imagine an Agent capable of continuously analysing:

• supplier risk;
• lead times;
• stock availability;
• purchasing commitments;
• cost increases;
• inventory exposure.

Could procurement teams move from operational management to predictive decision-making?

Could Business Central help identify supply chain risks before they disrupt the business?

Next week, we explore how AI Agents could transform one of the most critical functions in any ERP: Procurement and Supply Chain Management.

Because if Finance may be the first department to benefit from Agents…

Procurement could be the second.

The AI Agents Will Transform Microsoft Dynamics 365 Business Central

Chapter 1 of 6

The End of ERP Navigation: Why Business Central Is About to Change

For nearly three decades, ERP projects have pursued a familiar objective:

teaching people how to use the system.

We trained users.

We created manuals.

We documented procedures.

We delivered countless workshops and adoption sessions.

And all of that effort was ultimately focused on one thing:

Learning how to navigate.

Learning where information is stored.

Learning which page to open.

Learning which process to execute.

Learning which button to click.

For years, this made perfect sense.

But today, a question emerges that could fundamentally change how we think about ERP systems:

What if users no longer needed to learn how to navigate the ERP?

For Thirty Years We Built ERP Around Navigation

When a user needs information in Business Central, they typically need to know:

• Where the data resides.
• Which page contains the information.
• Which permissions are required.
• Which filters must be applied.
• Which report provides the answer.

Even relatively simple business questions often require a significant level of system knowledge.

A finance professional may fully understand the business process.

A purchasing manager may deeply understand supplier performance.

An operations manager may know exactly which decisions need to be taken.

Yet all of them still depend on understanding how to navigate the system.

For decades, we accepted this as normal.

Perhaps we are approaching a point where it no longer has to be.

The Problem Was Never Business Knowledge

Let us take a step back.

Most business leaders already know which questions matter.

The challenge is rarely identifying the right question.

The challenge is obtaining the answer.

A Finance Director may want to understand cash exposure.

A Purchasing Manager may need to identify supply chain risks.

A Chief Executive may need visibility into emerging operational issues.

The business knowledge already exists.

The difficulty lies in accessing and connecting the information required to support decisions.

This is where AI Agents begin to change the equation.

We Are Not Changing the Interface

We Are Changing the Mental Model

For years, ERP interaction followed a predictable sequence:

User -> Screen -> Process -> Result

To obtain answers, users needed to move through the entire chain.

AI Agents introduce a fundamentally different model:

User -> Intent -> Result

At first glance, this may appear to be a minor change.

It is not.

It could become the most significant shift in ERP interaction since the introduction of modern graphical user interfaces.

The system no longer begins with navigation.

It begins with intent.

Practical Business Scenario 1

The Finance Director and Credit Risk

Imagine a Monday morning executive meeting.

The Finance Director is asked:

«What is our biggest credit risk over the next thirty days?»

Today, answering that question often requires:

• reviewing customer ledgers;
• analysing overdue balances;
• checking payment histories;
• evaluating exposure levels;
• consolidating multiple reports.

The information is available.

The process is time-consuming.

Now imagine a different scenario.

The Finance Director asks:

«Which customers represent 80% of our projected credit risk during the next thirty days, and how does this compare with the previous quarter?»

The Agent:

• retrieves customer data;
• analyses open balances;
• reviews payment behaviour;
• identifies relevant trends;
• highlights priority risks.

The Finance Director still owns the decision.

The Agent simply removes the effort required to assemble the information.

The important question is not technological.

The important question is strategic:

How much time do organisations spend collecting information instead of making decisions?

Practical Business Scenario 2

The Purchasing Manager and Stock Shortages

Consider a Purchasing Manager responsible for maintaining product availability.

Today, identifying future shortages may involve reviewing:

• inventory levels;
• planning worksheets;
• purchase orders;
• supplier lead times;
• demand forecasts.

None of these activities are especially difficult.

However, they require time, navigation and interpretation.

Now imagine the following question:

«Which products are at risk of stock shortages before the end of the month?»

An Agent could:

• identify vulnerable products;
• analyse incoming purchase orders;
• review supplier commitments;
• estimate potential shortages;
• propose corrective actions.

The business value is not simply faster reporting.

The business value is earlier action.

And in many organisations, earlier action is the difference between preventing a problem and explaining it afterwards.

Practical Business Scenario 3

The CEO Who Never Wanted to Use an ERP

Let us be honest.

Many executives use ERP systems because they need information.

Not because they enjoy navigating software.

Most senior leaders do not want to learn transactions, reports and system structures.

They want answers.

Imagine a CEO asking:

«What are the three most significant operational risks we should review this week?»

To answer correctly, the Agent might need to evaluate:

• financial indicators;
• purchasing activities;
• sales performance;
• operational exceptions;
• service-related issues.

This goes far beyond automation.

This is about interpretation.

This is about context.

This is about transforming organisational knowledge into business intelligence.

And that may be where the greatest opportunity lies.

What Is Really Going to Change?

When professionals discuss AI Agents, the conversation often revolves around technology.

I believe the most significant change will be organisational.

For decades, organisations optimised systems so that people could learn how to use applications.

We are now entering an era where applications begin learning how organisations operate.

That shift changes where value is created.

Less importance on:

• navigation;
• menus;
• transactions;
• locating information.

More importance on:

• business rules;
• organisational knowledge;
• governance;
• trusted data;
• decision frameworks.

The conversation moves away from software usage.

It moves towards business understanding.

What Does This Mean for Functional Consultants?

This may be the most important question of all.

For years, functional consultants have acted as translators between business and technology.

We translated:

• business requirements;
• processes;
• objectives;
• policies;

into system behaviour and ERP configurations.

AI Agents do not eliminate that responsibility.

They amplify it.

Someone still needs to define:

• which knowledge should be trusted;
• which recommendations are appropriate;
• which actions may be executed;
• which governance rules apply;
• which business logic should guide decisions.

The role of the functional consultant does not become less important.

It becomes more strategic.

Less time spent explaining where to click.

More time spent teaching intelligent systems how the business truly operates.

And that may ultimately increase the strategic value of functional consulting more than any previous technology shift.

To Conclude

For decades, organisations taught people how to understand ERP systems.

The next decade may be about teaching ERP systems how to understand people.

The most significant transformation may not be technological.

It may be conceptual.

Because the future may not be about using Business Central more efficiently.

It may be about collaborating with Business Central more intelligently.

And when that happens, competitive advantage will no longer belong to those who know the most screens.

It will belong to those who understand the business best.

Coming Next…

Chapter 2 of 6

The Finance Agent

Imagine beginning every morning with a conversation rather than a collection of reports.

• Credit risk.
• Cash flow exposure.
• Payment trends.
• Profitability deviations.
• Financial alerts.

All prioritised.

All contextualised.

All available through a single interaction.

Are we approaching a future where financial decisions begin with a conversation?

Could a Finance Agent become the first truly specialised digital colleague within Microsoft Dynamics 365 Business Central?

Next week, we explore a scenario that may be much closer than many organisations realise.

Final chapter of the series «The Evolving Functional Role in Microsoft Fabric»

Throughout this series, we have explored how Copilot is transforming the way functional professionals engage with data. We have seen how it assists with data preparation, helps uncover insights and enables a deeper understanding of business information.

But all these advances ultimately lead to a much larger question:

What happens when users stop searching for information and information begins responding directly to them?

1. For Decades, We Learned to Speak the Language of Machines

The history of business analytics is, in many ways, the history of translation.

Business leaders think in questions.

Systems respond in technical language.

Between these two worlds, there has always been a layer of intermediaries:

• Business Analysts
• BI Specialists
• Data Engineers
• Data Analysts
• Solution Architects

When a Commercial Director asked:

«Why is profitability declining?»

the organisation typically translated that question into:

• SQL queries
• Data models
• Measures
• Business rules
• Visualisations
• Reports

Technology evolved.

The translation layer remained.

Despite major advances in analytics platforms, the process still required users to understand where information lived, how it was structured and how to navigate it effectively.

That dependency became the hidden cost of data-driven organisations.

The image visually represents the transition from data exploration to natural conversation.

2. The Real Revolution Is Not Artificial Intelligence

A common misconception exists across the industry.

Many professionals believe the breakthrough lies in generating queries more quickly or creating reports more efficiently.

That is not the fundamental shift.

The true transformation lies in the removal of complexity.

For the first time, business users can focus entirely on their objectives rather than the technology required to achieve them.

They no longer need to understand:

• SQL
• DAX
• Data structures
• Table relationships
• Data warehouses
• Lakehouses
• Semantic models

Instead, they focus exclusively on the business question.

The platform determines how to answer it.

The experience changes from navigating technology to expressing intent.

The complexity does not disappear.

It becomes invisible.

Functional Example

A Finance Director asks:

Which product categories have experienced the largest margin decline this quarter compared with the same period last year?

Traditionally this would involve:

• requesting a report,
• validating assumptions,
• waiting for analysis,
• reviewing multiple visualisations.

Now the interaction begins with a simple question.

The conversation replaces the process.

3. Semantic Models Become Organisational Knowledge

Traditionally, semantic models were designed to support reporting.

Their primary purpose was to provide consistency across dashboards and analytical solutions.

Today their role is becoming significantly more important.

Organisations are increasingly building semantic models not just for users but for intelligent systems.

This changes everything.

Semantic models are no longer simply:

• analytical structures,
• governance mechanisms,
• metric repositories,
• reporting foundations.

They become representations of organisational knowledge.

They provide business context.

They define meaning.

They establish trust.

Most importantly, they enable intelligent systems to understand the language of the organisation.

In this new paradigm, the semantic model is no longer the destination.

It becomes the starting point for every conversation.

Semantic Model and Data Agent architecture

4. The Final Democratisation of Data

For years, technology vendors have spoken about democratising data.

Progress has undoubtedly been made.

Yet one challenge remained.

Users still needed to:

• locate reports,
• understand metrics,
• navigate pages,
• interpret visualisations,
• connect insights.

Even with self-service analytics, the burden remained on the user.

Conversational intelligence changes the equation.

Business users interact using the only interface they have mastered since childhood:

language.

They ask questions exactly as they would during a management meeting.

Examples include:

Which suppliers have increased costs most significantly this year?

Which customers are showing early signs of churn?

Which regions are exceeding their sales targets despite market contraction?

The barrier between data and decision-making becomes significantly smaller.

For many organisations, this may be the first genuine step towards universal data accessibility.

5. The Emergence of Enterprise Conversational Intelligence

Perhaps the most profound shift is not technical.

It is organisational.

Data is no longer confined to analytical platforms.

Intelligence begins to appear wherever employees work.

Imagine a sales executive opening Microsoft Teams and asking:

Which European customers present the highest revenue risk over the next quarter?

The system analyses organisational knowledge, business definitions and historical patterns.

The answer appears within the natural flow of work.

No dashboard.

No navigation.

No training course.

No technical intervention.

The user remains focused on decision-making rather than information retrieval.

This changes not only how analytics operates.

It changes how organisations think.

Data Agents integrated with Teams, Copilot Studio and enterprise applications

6. Will Analysts Become Less Important?

Quite the opposite.

As technology reduces technical barriers, business understanding becomes more valuable.

The competitive advantage shifts from:

«I know how to access information.»

to

«I know which questions matter.»

Professionals who understand:

• processes,
• operations,
• customers,
• finance,
• strategy,

will become increasingly important.

Technology can retrieve information.

Only people can define which problems are worth solving.

7. What This Means for Functional Professionals

Functional consultants have traditionally served as translators between business and technology.

That responsibility is evolving.

Less time will be spent on:

• explaining data structures,
• validating queries,
• navigating technical complexity.

More time will be devoted to:

• process design,
• governance,
• adoption,
• strategy,
• business transformation.

The role does not diminish.

It becomes more strategic.

The ability to frame problems correctly may become more valuable than the ability to build solutions.

To Conclude

Throughout this journey we have explored how intelligent technologies are reshaping the functional landscape.

We have seen how they:

• simplify complex processes,
• accelerate analysis,
• enhance decision-making,
• strengthen business understanding.

But perhaps the most significant transformation is this:

Data is no longer something we search for.

Data becomes something we speak with.

This is not merely the evolution of reporting.

It is the evolution of the relationship between people and information.

And once organisations begin having conversations with their data,

it becomes very difficult to return to searching for answers the old way.

Final Reflection

The next generation of business professionals will not be recognised for their ability to build reports. They will be recognised for their ability to ask better questions.

Episode 4 — The Semantic Model as a Cognitive Layer in Power BI: When Copilot Stops Describing and Starts Reasoning

Copilot doesn’t transform your data. Copilot transforms how your semantic model thinks.

This episode marks a turning point: Power BI is no longer just a visualization tool. It becomes a reasoning system, where the semantic model acts as a cognitive layer capable of interpreting, inferring, and sustaining business logic.

This is not about functional intelligence. This is about structural intelligence.

1. The semantic model is no longer a container — it’s a cognitive system

Business example: A retail chain with 120 stores asks Copilot why “Electronics” sales dropped in Week 32. If the model lacks hierarchies, promotion logic, store granularity, or margin measures, Copilot can only describe.

But if the model encodes business logic, Copilot can reason:

“The decline is driven by inventory shortages across 34 stores and the absence of the active promotion in the northern region.”

Copilot doesn’t invent this. The model does.

2. Copilot doesn’t work on dashboards — it works on logic

Business example: A logistics company wants Copilot to generate a report on route efficiency. If the model only contains trips, vehicles, and drivers, Copilot can describe.

But if the model includes optimal vs actual distance, loading times, delivery windows, penalties, and cost per kilometer, Copilot can infer:

“Inefficient routes correlate with strict delivery windows and high urban density.”

This is business reasoning, not visualization.

3. The new KPI for BI: Can your model reason?

Business example: A bank asks Copilot to explain changes in SME credit risk.

If the model includes risk measures, product relationships, payment history, sector classification, and scoring rules, Copilot can generate DAX such as:

“AdjustedRisk = BaseRisk + SectorPenalty + LatePayments * 0.3”

And explain:

“Risk increased due to exposure to the construction sector and two late payments in the last 12 months.”

This is semantic reasoning.

4. The functional consultant becomes an intelligence architect

Business example: A manufacturing company wants Copilot to identify production bottlenecks.

The functional consultant must now model:

• cycle times
• machine dependencies
• production capacity
• maintenance logic
• shift constraints

This consultant becomes an intelligence architect, defining how Copilot interprets operations.

Copilot doesn’t replace the functional role. Copilot elevates it.

5. The future of BI is not visual — it’s cognitive

Business example: An insurance company wants Copilot to explain why premium vehicle claims increased.

If the model includes risk zones, weather, vehicle type, driver history, traffic density, and policy rules, Copilot can reason:

“The increase is concentrated in high‑congestion zones during rainy days, especially for vehicles older than five years.”

This is contextual reasoning, not reporting.

To conclude

Copilot doesn’t make Power BI smarter. Copilot makes your semantic model visible.

And that forces teams to build models that don’t just represent data… but reason through the business.

This is the real generational shift in enterprise analytics.

Series: “The Evolving Functional Role in Microsoft Fabric” — Part 3 of 5

1. The semantic model becomes a functional intelligence engine

The semantic model has always been the backbone of Power BI. But until now, only technical profiles could truly interpret it.

Copilot changes that.

Copilot understands:

– table relationships

• cardinalities
• hierarchies
• measures
• calculated columns
• business semantics
• filter logic
• metric dependencies

With this understanding, functional roles can ask questions as if speaking to an analyst:

“Which region is generating the highest adjusted margin?” “What trend should concern us this quarter?” “Which product is losing profitability, and why?”

Copilot analyses the model and returns:

• a suggested visual
• a narrative explanation
• a key insight
• a functional recommendation

2. Copilot generates complete report pages from functional intent

Traditionally, building a report required:

• designing visuals
• selecting fields
• configuring filters
• adjusting formatting
• validating measures

Now Copilot can generate entire report pages from a functional intention.

Example:

“Create a page with regional sales, monthly variation and an executive summary.”

Copilot produces:

• a bar chart by region
• a monthly trend line
• a detail table
• an executive narrative
• an insight on the region with the highest variation
• a functional recommendation

This does not replace the technical analyst. It augments them — and allows functional roles to progress independently.

3. Copilot writes, corrects and explains DAX with business logic

This is one of the most transformative capabilities.

Copilot does not only generate DAX. Copilot explains DAX.

Example:

“Explain how the Net Margin measure is calculated.”

Copilot returns:

• the full formula
• the business logic
• applied filters
• column impact
• optimisation suggestions
• potential misinterpretations
• more efficient alternatives 

This enables functional roles to:

• validate metrics
• detect errors
• understand business logic
• explain KPIs to executives
• propose improvements

Without writing a single line of DAX.

4. Executive narratives: the report speaks, explains and recommends

Copilot generates executive‑level narratives that explain:

• trends
• anomalies
• comparisons
• risks
• opportunities
• actionable recommendations

Example:

“Summarise last quarter’s performance.”

Copilot produces:

• a clear narrative
• a variation analysis
• an insight on the root cause
• a functional recommendation
• a potential risk 

This turns Power BI into an analytical assistant — not just a visualisation tool.

5. Functional KPI validation without technical dependency

Copilot enables validation of:

• financial KPIs
• operational KPIs
• commercial KPIs
• service KPIs
• project KPIs

Example:

“Validate whether the Operating Margin KPI is calculated correctly.”

Copilot reviews:

• the measure
• the logic
• filters
• dependencies
• model coherence
• business semantics

And returns:

• validation
• explanation
• recommendation

This is invaluable for functional consultants.

To conclude

Copilot is not an add‑on to Power BI.

It is a new layer of intelligence built on top of the semantic model, transforming how functional professionals interact with data.

For the first time, a functional consultant can:

• validate KPIs without writing DAX,
• generate complete report pages without manual design,
• obtain executive narratives without drafting them,
• uncover insights without relying on technical analysts,
• and explain complex metrics with clarity and business context. 

This does not eliminate the technical role. It enhances it.

And it opens a new space for functional professionals: the augmented analyst, capable of connecting business, data and decisions with unprecedented speed.

Copilot turns Power BI into conversation.

And it places functional roles at the centre of modern analytics.

Part 4 will explore how Copilot transforms documentation, governance and data lineage within Fabric — and how this directly impacts real Dynamics 365 and Business Central projects.

Continuation of the series “The Evolving Functional Role in Microsoft Fabric” (Part 2 of 5)

Previously, we uncovered how Copilot reshapes the functional landscape. Today, we step into the engine room of Fabric… where data engineering becomes accessible to everyone.

1. Copilot in Data Factory: Transforming Data Without Writing Code

Data Factory has always been powerful—but technically demanding. Copilot introduces a functional layer that allows professionals to:

• generate transformations,
• clean and prepare data,
• adjust column logic,
• explain each step,
• validate business rules,
• detect inconsistencies.

All without writing a single line of code.

Functional example

An analyst needs to classify customers by age. Before: they relied entirely on an engineer. Now they ask Copilot:

“Create an AgeRange column with business‑defined age categories.”

Copilot generates the transformation and explains the logic. The analyst reviews and approves it confidently.

Functional value

• Reduces dependency on technical teams.
• Accelerates decision‑making.
• Ensures business rules are applied correctly.
• Improves data quality from the functional side.

2. Copilot in Pipelines: Understanding Data Flows Without Engineering Skills

Pipelines are the backbone of data movement. Copilot enables functional roles to:

• generate activities,
• add steps,
• request summaries,
• understand source and destination,
• detect errors,
• validate flow logic.

Functional example

A consultant needs to confirm whether a pipeline is moving data correctly. Before: they waited for a technical explanation. Now they ask Copilot:

“Summarise this pipeline and explain its source and destination.”

Copilot returns:

• source,
• destination,
• table,
• activity purpose,
• dependencies.

No JSON. No logs. No engineering jargon.

Functional value

• Enables functional auditing of data processes.
• Improves communication between business and engineering.
• Reduces misunderstandings in data flows.
• Increases autonomy and clarity.

3. Copilot as a Bridge Between Business Logic and Technical Execution

Copilot does not replace engineers—it enhances collaboration.

Functional professionals can now:

• validate business rules,
• review transformations,
• understand flow logic,
• detect anomalies,
• request adjustments,
• document processes.

Functional example

A consultant notices duplicated records in a dataset. Before: they waited for engineering analysis. Now they ask Copilot:

“Explain why this activity might be producing duplicates.”

Copilot identifies potential causes:

• missing primary key,
• incorrect join,
• lack of filtering,
• transformation error.

The consultant enters the meeting informed and prepared.

4. Copilot as a Functional Documentation Tool

• Copilot can generate:
• summaries,
• descriptions,
• step‑by‑step logic,
• transformation explanations,
• dependency maps.

This turns Copilot into a functional documentation engine.

Functional example

A consultant must document a pipeline for a client. Before: they relied on engineers to describe each step. Now Copilot produces:

«This pipeline copies data from Azure SQL to Lakehouse, applying filters and transformations based on defined parameters.”

The consultant simply refines the wording for the client.

5. What This Means for Functional Roles

Copilot empowers functional professionals to:

• understand complex data processes,
• validate business logic,
• detect errors early,
• request precise changes,
• document flows clearly,
• participate in data engineering,
• reduce analysis time,
• improve data quality.

All without writing code. All without losing control of the process.

To conclude

Copilot democratises data engineering. It enables analysts and consultants to engage directly with data preparation, movement and validation. Fabric is no longer a purely technical platform—it is a collaborative environment where business roles act with autonomy, clarity and speed.

And the story deepens…

Microsoft Fabric introduces a fundamental shift in how functional teams interact with data, analytics and decision‑making. Copilot is not merely an AI assistant; it is a new operating model for business users, analysts and consultants who work with data every day.

For years, functional roles relied heavily on technical specialists to transform data, validate models or build reports. Fabric changes that dynamic. Copilot empowers functional professionals to participate directly in the data value chain — without needing deep technical expertise.

1. Functional roles now participate in data engineering

Copilot enables functional roles to:

• generate transformations,
• clean and prepare data,
• create pipelines,
• explain code,
• detect errors,
• validate flows.

All without needing to write Spark, SQL or DAX.

Functional example

An analyst notices inconsistent values in the “CustomerSegment” field. Previously: they opened a ticket and waited. Now: they ask Copilot to review the transformation and identify where the inconsistency originates. The engineer steps in only for the final correction.

2. The functional consultant becomes a designer of context

Copilot works through context, not magic. This means functional consultants must now:

• guide how questions are formulated,
• define clear instructions,
• validate Copilot’s responses,
• understand token processing and consumption,
• design prompts for Fabric Data Agents,
• translate business processes into natural‑language queries.

Functional example

A manager asks:

“How is the campaign performing?”

Copilot returns a vague answer. The functional consultant reframes the question:

“Compare conversions and cost per channel for Campaign X over the last quarter.”

The insight becomes precise and actionable.

3. Power BI evolves from reporting to conversation

Copilot transforms Power BI into a conversational analytics tool where:

• users ask questions,
• reports respond,
• functional consultants guide the interaction,
• semantic models become a shared language.

Functional example

A finance director asks:

“What is the EBITDA variance by region over the last six months?”

Copilot produces:

• a visual,
• a narrative summary,
• an explanation of the trend.

The functional consultant ensures the semantic model supports the question correctly.

4. Fabric Data Agents: the biggest functional leap

ºFabric Data Agents allow any user to:

• ask questions in natural language,
• receive contextual answers,
• without opening Power BI,
• without SQL,
• without DAX.

Functional consultants design:

• instructions,
• examples,
• constraints,
• tone,
• user experience.

Functional example

An operations manager asks:

“Which plant had the most production delays this month?”

The agent responds with:

• a table,
• a chart,
• a contextual explanation,
• an additional insight.

5. Functional roles must now understand cost, capacity and consumption

Fabric introduces a new concept: cost per token.

Functional professionals must now:

• recommend when Copilot should be used,
• avoid unnecessary prompts,
• optimise consumption,
• understand capacity strategies,
• advise on scaling decisions.

Functional example

A team uses Copilot to generate long, complex SQL queries. Consumption spikes. The functional consultant recommends:

• shorter prompts,
• splitting questions,
• using Data Agents for repetitive queries.

To conclude

Fabric does not simply modernise technology. It reshapes how functional roles work with data, insights and decision‑making.

Copilot turns functional professionals into:

• context designers,
• question architects,
• decision facilitators,
• insight enablers.

This is the new functional landscape in Microsoft Fabric — one where business users and consultants play a central role in the analytical process.

Poor functional decisions in Business Central SaaS rarely show their impact on the day they are made.

They show up later—when the system is already in production, users depend on it, and correcting the damage costs far more than doing things properly from the start.

This is not a technical post.

It is a practical reflection on how small functional choices can create large operational, financial and organisational consequences.

1. When a functional change looks small… but isn’t

Some decisions seem harmless:

• enabling a setting,
• adjusting a parameter,
• adding a field,
• modifying a workflow,
• allowing an exception “just this once”.

Business Central does not forgive improvisation.

Real case

A retail company enabled bin tracking because “the warehouse manager wanted more control”. No impact assessment. No process review. No sandbox testing.

Three weeks later:

• picking slowed down,
• warehouse staff were confused,
• orders were delayed,
• inventory became unreliable,
• finance received inconsistent valuations.

The hidden cost wasn’t the configuration. It was lost productivity, frustrated staff, and a decline in service quality.

A junior consultant sees the screen. A senior consultant sees the consequences.

2. The cost of creating data without purpose

Business Central makes it easy to create fields, tables and automations. But ease can become a trap.

Real case

A team created custom fields to “improve sales control”. No one defined:

• who would use them,
• what they were for,
• how they would be validated,
• which process supported them,
• how they affected reporting.

Result:

• inconsistent reports,
• duplicated data,
• decisions based on incomplete information,
• internal debates about “which number is correct”.

The hidden cost wasn’t the field. It was the loss of trust in the ERP.

A junior consultant creates fields. A senior consultant avoids them.

3. The cost of misunderstanding accounting impact

Functional decisions always end up in finance. Always.

Real case

A user changed VAT settings because “a supplier asked for it”. No impact review. No historical analysis. No validation with accounting.

Result:

• incorrect entries,
• inconsistent tax reports,
• blocked internal audits,
• months of corrections.

The hidden cost wasn’t the change. It was the correction, the audit, the wasted time, and the compliance exposure.

A junior consultant looks at the screen. A senior consultant thinks about month-end closing.

4. The cost of allowing functional exceptions

Exceptions feel convenient. Until they destroy the system.

Real case

A client asked to post a purchase invoice without a purchase order “just this once”. It was allowed. Then it happened again. And again. And again.

Three months later:

• purchasing control disappeared,
• cost reports didn’t match,
• margins became impossible to calculate,
• the ERP stopped reflecting reality.

The hidden cost was the loss of operational discipline.

A junior consultant says “yes”. A senior consultant says “no—and explains why”.

5. The cost of not thinking ahead

A functional decision is never for today. It is for six months from now.

Real case

A company customised its approval workflow because “that’s how we do it in Excel”. It worked. But six months later they wanted:

• automated purchasing,
• supplier integration,
• Power Automate flows,
• mobile approvals.

The customisation blocked everything. The workflow had to be rebuilt from scratch.

The hidden cost was functional debt.

A junior consultant thinks about the current need. A senior consultant thinks about the future of the business.

To Conclude

Poor functional decisions don’t break the system immediately. They break it slowly—through frustration, rework, inconsistencies, and operational noise.

Business Central SaaS doesn’t fail because of technology. It fails because of functional decisions made without criteria.

The hidden cost of a poor functional decision is:

• time,
• money,
• frustration,
• loss of trust,
• loss of control,
• loss of operational quality.

A junior consultant configures. A senior consultant is responsible for planning for the future and handling updates.