Next-Gen

Financial Models
Why AI is not a Panacea

SPONSORED BY

™
Table of Contents 02

I. Is This Paper Right for Me? 03

II. Overview: How We Got into This Mess 05

III. Data Quality & The Challenges with Multiple Data Sources 07

IV. Collaboration Challenges in Financial Modeling 09

V. Model Management Challenges 11

VI. The Pitfalls with Current FP&A Solutions 13

VII. AI Won’t Help You: Fixing the Data Problem Will 15

VIII. A Breakthrough Approach to Models That Might Save Your Job 17

IX. Excel: Have Your Cake and Eat It Too 19

X. Conclusion 20

XI. About the Authors 21

I. Is This Paper Right for Me? 03

Executive Summary
This white paper highlights the limitations of current financial modeling tools and
processes, particularly Excel, and proposes a new approach to model management.

Key issues include

Lack of structured processes for model management, leading to

inconsistencies and errors.

Complexity and limited capability for collaboration in large models.

Challenges with knowledge transfer and continuity when individuals

leave organizations.

Limited transparency and audit trails.

Issues with user adoption, price, and flexibility with current Financial
Planning & Analysis (FP&A) solutions.

Limitations in data complexity and scalability with Relational Databases

(RDBs).

The paper proposes a new approach to financial modeling, using Event Databases
(EDBs) and a Model Server to address these challenges. EDBs can handle high-
volume, high-velocity, and high-variety data, making them ideal for Artificial
Intelligence (AI) and Machine learning (ML) applications.

A Model Server can separate data from formulas, protect sensitive information,
enable safe collaboration, and manage model approval and usage.

The paper also highlights the importance of continuing to use Excel for financial
modeling by leveraging its universal understanding and customization capabilities,
while addressing its limitations through the proposed new approach.

Next-Gen Financial Models

I. Is This Paper Right for Me? 04

Who Should Read This

This white paper is for senior executives and financial professionals who play a
critical role in driving business decisions and strategy.

The C-Suite
in small and medium-sized businesses (SMBs) responsible for
building pro formas, forecasting revenue, cash, capacity and
profitability metrics

M&A Professionals
calculating revenue and operating efficiency metrics for potential
acquisitions

Financial Analysts & Portfolio Managers

forecasting derivative portfolio metrics

Controllers & Financial Managers

Controllers and financial managers responsible for consolidating
data and providing insights to executive decision-makers

This white paper argues that there are inherent limitations of traditional financial
modeling tools and why the industry should be adopting a modern, cloud-based
approach.

Next-Gen Financial Models

II. Overview: How We Got into This Mess 05

A Quick History of Financial Modeling

Financial modeling is the process of creating a mathematical representation of a
company's financial performance in order to forecast future financial outcomes,
analyze business scenarios, and make informed decisions.

It uses historical data, market trends, and business assumptions to build a dynamic
model that simulates various financial scenarios, such as forecasting revenue,
expenses, cash flow, and profitability.

Financial models are used to evaluate investment opportunities, assess risk, and
optimize business strategies to provide a quantitative framework for decision-
making and strategic planning.

By creating a financial model, businesses can better understand their financial

dynamics, identify areas for improvement, and make data-driven decisions to drive
growth and profitability.

The world of financial modeling has undergone a significant transformation over the
past few decades, evolving from a labor-intensive, error-prone process to a
complex, data-driven discipline.

However, despite this progress, the tools used in financial modeling have not kept
pace with the changing demands of the field. To understand how we arrived at this
juncture, it's essential to take a step back and examine the history of financial
modeling.

Prior to 1979, financial modeling was a manual process, reliant on clerks armed with
pencils and paper. The introduction of mainframe computers brought some
automation, but their adoption was limited.

The game-changer came with the invention of the electronic spreadsheet by Dan
Bricklin, a Harvard MBA student, who developed VisiCalc. This innovation gave
birth to the modern financial analyst and financial modeling as we know it today.

The subsequent introduction of Lotus 1-2-3 in 1983 and Microsoft Excel in 1985
further revolutionized financial modeling, with Excel eventually dominating the
market.

Next-Gen Financial Models

II. Overview: How We Got into This Mess 06

However, 30 years on, the question has become: can a program designed three
decades ago still meet the modern demands of financial modeling?

The answer, unfortunately, is no.

Financial modeling suffers from three (3) main issues

Entering & checking data is labor-intensive

Various studies report 88% of models contain significant errors

from data or model mistakes

Collaboration is difficult, which makes it hard to distribute the

work across the team

The elephant in the room is that Excel, despite its widespread adoption, is no longer
equipped to handle the complexities of modern financial modeling.

The advances in data and increasing business complexity that have occurred in
recent years have pushed Excel beyond its 30-year-old design, exposing its
limitations.

The buzz of Artificial Intelligence has seeped into every industry – especially
financial modeling. The suggestion from enthusiasts is that AI will eliminate these
three problem issues in financial modeling.

The garbage-in garbage-out adage is even more true for AI datasets.

AI is only as good as the data that was used to train it, plus the algorithm used for
querying that data.

In this white paper, we will lay out a groundbreaking new approach to financial
modeling, including how AI should fit into the big picture.

Next-Gen Financial Models

III. Data Quality & The Challenges with Multiple Data Sources 07

Navigating Complex Data in Financial

Modeling
In today's business landscape, data has become increasingly complex, presenting a
multitude of challenges for financial modelers. According to Forbes, nearly 9 out of
10 spreadsheets (88%) have significant errors.

This chapter exposes the complexities of data collection, integration, and

synchronization in financial modeling and proposes solutions to mitigate associated
risks.

Data Variety and Sources

Data for typical corporate models originates from various enterprise systems,
including sales forecasts, accounting, inventory, and production, among others.

These diverse data sources generate reports that must be updated for every new
period, creating a continuous cycle of data collection and integration.

Labor-Intensive Data Management

Transferring data from enterprise systems to Excel is a labor-intensive process.
Moreover, checking the accuracy of this data with each new period is time-
consuming and prone to errors, as manual verification processes may overlook
discrepancies or inconsistencies.

I spend 80% of my time importing and checking data and only

20% of my time actually modeling
Lisa Payne, CFO

Next-Gen Financial Models

III. Data Quality & The Challenges with Multiple Data Sources 08

Distributed Systems & Time Coherence

The decentralization of systems has made data synchronization across different
platforms and locations exceedingly challenging.

With systems moving to the cloud and becoming more distributed, maintaining
"time coherence" — the alignment of data streams in time — has become a major
concern. Misaligned data streams can lead to errors in financial modeling and
analysis.

Understanding Data Lineage

One of the critical aspects of data management in financial modeling is
understanding the lineage of data within a model.

Tracing data back to its source systems, or "data lineage," is essential for ensuring
transparency and accuracy.

However, in many cases, this process requires going back to legacy systems,
which can be time-consuming and inefficient, and in many case the data has been
overwritten.

Next-Gen Financial Models

IV. Collaboration Challenges in Financial Modeling 09

Model Complexity Combined with a Lack

of Collaboration Support Hinder
Teamwork
Collaborating with others when building a financial model can be challenging, often
placing the modeling burden on a few individuals. Excel, the commonly-used tool
for financial modeling, was never designed to be a team collaboration platform.

There is nothing more frustrating than opening a model in the

morning to find it broken by a team member
Jeff Bouganim, CFO

Formula and Value Ambiguity

Excel stores both formulas and values in a single cell, making it difficult to discern
at a glance whether a cell contains a formula or a hard-coded value.

This ambiguity creates the risk that a collaborator might inadvertently replace a
formula with a value or vice versa, altering the integrity of the model without the
owner's knowledge.

Unintended Model Changes

Sharing a financial model increases the risk of unintended changes. Collaborators
may unknowingly modify the model in ways that the owner cannot easily detect,
such as by breaking links, changing constants, or editing formulas.

Even a seemingly innocuous change can have significant ramifications, leading to

errors or misinterpretations of the model's outputs.

Next-Gen Financial Models

IV. Collaboration Challenges in Financial Modeling 10

Difficulty in Identifying and Fixing Issues

When a collaborator inadvertently alters the model, identifying and rectifying the
problem can be challenging.

Tracking down the source of errors or broken links within a large and complex
model can take hours or even days, delaying decision-making and potentially
affecting the accuracy of financial analysis.

Complexity Hinders Collaboration

As financial models grow in size and complexity, new contributors have an
increasingly difficult time understanding the intricacies of the model.

This complexity often discourages collaboration, as third parties may find it easier
to build their own models from scratch rather than attempting to decipher and work
within an existing framework.

Collaboration in financial modeling is essential for accuracy and efficiency, yet it is

fraught with challenges. Excel's limitations, coupled with the risk of unintended
changes and the complexity of large models, hinder effective collaboration.

Next-Gen Financial Models

V. Model Management Challenges 11

You Can’t Trust the Output if You Can’t

Manage the Models
Model management is almost nonexistent, with little to no process for usage,
approvals, or compliance. It’s like running your factory without standard operating
procedures.

Lack of Model Management Processes

One of the most pressing issues in financial modeling is the absence of structured
processes for model management. Many organizations lack clear guidelines for
model usage, approvals, and compliance, leading to inconsistencies and potential
errors in decision-making.

Peer Review and Approval Procedures

Reviewing and approving models for use is crucial to ensure accuracy and
reliability. However, without formalized processes, models may be used without
proper scrutiny, increasing the risk of errors and misinterpretations of financial
data.

Ensuring Model Accuracy and Applicability

Another challenge is ensuring that the correct model is being used for a given
purpose across the organization. Without proper oversight, different departments
or individuals may use outdated or inappropriate models, leading to inconsistencies
in analysis and decision-making.

I consolidate sales numbers across the country for quarterly

forecasts and know that every manager is using a different
forecasting model
Lisa Payne, CFO

Next-Gen Financial Models

V. Model Management Challenges 12

Managing Complexity

Large and complex models can be incomprehensible to many users, hindering

collaboration and understanding. Breaking these models into smaller, more
manageable components or atomic elements, can reduce complexity and improve
transparency, making it easier for stakeholders to interpret and utilize the model
outputs.

Knowledge Transfer and Continuity

Models are often created and understood by a few individuals within an
organization. When these individuals leave, valuable institutional knowledge goes
with them, and new teams must rebuild models from scratch, wasting time and
resources.

I have worked in multiple growth stage companies and have

to reinvent the forecasting model for fund raising every
single time
Lisa Payne, CFO

Providing Transparency through Audit Trails

To enhance accountability and transparency, organizations need to provide an

audit trail of decisions made based on specific models. This ensures that
stakeholders understand the rationale behind decisions and can trace values back
to the source data and assumptions used in the modeling process.

Model management must advance in order for the fintech sector to take full
advantage of its data.

Next-Gen Financial Models

VI. The Pitfalls with Current FP&A Solutions 13

Current FP&A Solutions Face User

Adoption, Price, and Flexibility Issues
Current Financial Planning and Analysis (FP&A) solutions are fraught with pitfalls
that hinder their effectiveness.

One major issue is user adoption, which is often hampered by the complexity, steep
learning curve, and related time investment to learn the product.

As a result, many users struggle to fully utilize the tools, leading to frustration and
limited ROI.

There must be a reason that 80% of companies still use Excel

for financial modeling versus the dozens of platforms out
there.
Jeff Bouganim, CFO

This frustration and limited ROI compound the fact that enterprise FP&A solutions
are often expensive, requiring significant upfront investments and ongoing
maintenance costs.

These solutions may also be rigid and make it difficult to represent your specific
business model. Having to push the model out of the platform and into Excel is a
step backwards in terms of model management and collaboration.

Next-Gen Financial Models

VI. The Pitfalls with Current FP&A Solutions 14

Data complexity is the third main reason that FP&A platforms struggle to solve the
three core problems of modern financial modeling at an affordable price.

Answering a simple question like, ‘What will my cash position be if we close 80%
of the funnel’ may sound trivial. In fact, the queries to multiple legacy systems,
accounting, CRM, inventory, and HR, among others, is phenomenally complex. And
changing the question to ‘What if I close 75%’ requires a whole new set of queries.

Maintaining the health of these queries explains both the expense of FP&A
platforms and their inherent inflexibility.

I’ve done financial modeling at half a dozen companies and

have yet to be able to justify the expense of an FP&A
platform. Can you imagine a sales leader saying that about a
CRM system?
Lisa Payne, CFO

Next-Gen Financial Models

VII. AI Won’t Help You: Fixing the Data Problem Will 15

AI Won’t Help You: Fixing the Data

Problem Will
The way we approach databases has not kept pace with the exponential growth in
computing power and data storage capabilities.

In 1970, computing power was limited and data storage was expensive, leading to
the development of Relational Databases (RDBs) that optimized storage efficiency.

However, this approach has become a bottleneck in today's data-driven world. As

data volumes continue to grow, RDBs face scaling problems, leading to increased
costs, decreased performance, and limited ability to handle real-time data.

Remember, RDBs were designed to minimize your data footprint by taking selected
snapshots in time. Event Databases (EDBs) are the equivalent of a movie with no
data thrown away.

AI prefers a movie over snapshots.

AI is only as good as the data inputs - and AI needs A LOT OF DATA to

be effective
Chris Strachan, CEO

In contrast, EDBs are designed to handle high-volume, high-velocity, and high-

variety data, making them ideal for modern applications, especially AI and ML
applications. By adopting an EDB approach, organizations can unlock the full
potential of AI and ML capabilities, leading to significant improvements in cost,
accuracy, and time-to-insight.

With EDBs, organizations can process large datasets in real-time, enabling faster
and more accurate decision-making and achieving actual outcomes such as
reduced costs, improved accuracy, and accelerated time-to-market.

EDBs, or ledger-based systems, let you strip away the layers of query complexity
that RBDs have imposed to overcome their limitations.

Next-Gen Financial Models

Next-Gen Financial Models 16

Combining a Ledger or Event

Database with a Model Server
uniquely solves the biggest
issues in financial modeling
Chris Strachan, CEO PowerModels™
VIII. A Breakthrough Approach to Models That Might Save Your Job 17

Models Can No Longer be a Standalone

Component of Financial Modeling
Salesforce revolutionized the enterprise sales process with a system that imported
the salesperson’s Rolodex and Daytimer and transformed them into the enterprise
IP.

The same needs to be true for financial modeling, which is arguably more important
than the CRM, accounting, or inventory systems that make up FP&As atomic
elements.

The following is a set of requirements for the next-generation FP&A solution

1. Separate Data from Formulas

Import Excel models and separate data and formulas to ensure precise access and
version control.

2. Protect Data and Formulas

Implement access controls, encryption, and backups to safeguard sensitive data
and support robust collaboration.

3. Share and Collaborate Safely

Use cloud-based platforms to share and collaborate on models, ensuring real-time
updates and version control.

John accidentally overwrote the company's financial forecast,

causing a delay in quarterly reporting. With version control, he could
have easily rolled back to a previous version.
Lisa Payne, CFO

Next-Gen Financial Models

VIII. A Breakthrough Approach to Models That Might Save Your Job 18

4. Manage and Track Model Approval and Usage

Establish a clear approval process and track model usage to ensure accuracy,
consistency, and compliance.

5. Version Control
Use version control systems to track changes, identify errors, and roll back to
previous versions if needed.

6. Creating Training
Create an application workflow that includes a framework for training and best
practices for model usage.

7. IP Ownership
Ensure that the company maintains its financial modeling IP by importing models
into a Model Server. This would have additional benefits for corporate IP, including:

Supporting model hierarchies and nested models makes it easier to understand

and reuse model components

Tracking approvals and usage ensures compliance and adherence to business

practices

Enabling new team members to quickly become contributors to modeling IP

Next-Gen Financial Models

IX. Excel: Have Your Cake and Eat It Too 19

Excel: Have Your Cake and Eat It Too

Is there a reason or even a distinct advantage to keeping Excel around?

The answer is yes, if you can overcome the associated problems. There are two
strategies that, taken together, will eliminate Excel’s limitations.

Event Database
An event database that is fed real-time business data will eliminate data
management issues including:

Copy and paste errors

Time coherence
Data lineage verification
Data checking and rechecking effort

Model Server
A properly-designed Model Server eliminates other Excel-related problems by:

Protecting formulas and data from inadvertent changes or mistakes

Breaking models into separate, reusable modules to reduce the complexity and
incomprehensible nature of multi-sheet, multi-tab models
Requiring very limited user training
Implementing a tracking and approvals process

Excel As Your User Interface

Once the core two problems are solved, the advantage of continuing to use Excel
for financial modeling include:

Improved adoption by using a universally understood interface

Easy customization to model any scenario, unlike most FP&A platforms

Next-Gen Financial Models

X. Conclusion 20

Conclusion
While Artificial Intelligence has the potential to revolutionize financial modeling, it is
not a silver bullet for the challenges faced by financial professionals.

The limitations of traditional financial modeling tools, such as Excel, and the need
for better data management, collaboration, and model management practices must
be addressed first.

By adopting a modern, cloud-based approach to financial modeling, organizations

can unlock the full potential of AI and Machine Learning capabilities, leading to
significant improvements in cost, accuracy, and time-to-insight.

Key Takeaways

Financial modeling has evolved significantly over the past few decades, but the tools
used have not kept pace with the changing demands of the field.

Excel, despite its widespread adoption, is no longer equipped to handle the complexities
of modern financial modeling.

AI is only as good as the data it relies on, and poor data quality and management can
lead to inaccurate results.

Collaboration and model management are critical components of effective financial

modeling, and traditional tools and practices are no longer sufficient.

A modern, cloud-based approach to financial modeling, combined with better data

management and collaboration practices, can help organizations achieve accurate,
reliable, and secure financial models that support informed decision-making.

This white paper demonstrates why financial modeling must move to event-driven
databases as part of a modern, cloud-based approach.

Next-Gen Financial Models

XI. About the Authors 21

Contributors
Created by industry leaders, for industry leaders

Christopher Strachan Lisa Payne

CEO, PowerModels™ CFO, INETCO Systems Limited

25+ years in SaaS software and finance CFO in multiple growth stage companies
leading financial modeling initiatives
Connect On LinkedIn Connect On LinkedIn

Jeff Bouganim Chris Condron

CFO, Purpose Investments CTO, PowerModels™

25+ years leading financial analysis 25+ years of building large scale, financial
departments at leading financial institutions services software initiatives
Connect On LinkedIn Connect On LinkedIn

Next-Gen Financial Models

Next-Gen
Financial Models
Why AI is not a Panacea

Contact Us
www.canadianlenders.org

SPONSORED BY