Tallinn University of Technology

Business Analytics Implementation in Financial Planning & Analysis: Challenges & Opportunities

A qualitative study on the current state, key challenges, and future perspectives of business analytics adoption in FP&A activities of Estonian technology sector companies.

EV

Author

Egert Väinaste

Master's Thesis, 202412,323 wordsSupervisor: Mari Avarmaa, PhD
open_in_newRead the Full Thesis – TalTech Digital Library
Financial Data Analytics

“It is critically important to be open to new challenges and opportunities, even if it means stepping out of your comfort zone.”

– CFO, thesis interview

Evolution of Analytics Maturity

Davenport & Harris (2007) classify business analytics into four progressively sophisticated levels. In the FP&A context, most organizations remain anchored at descriptive analytics – standard dashboards and backward-looking reports. The journey toward predictive and prescriptive analytics represents not just a technology upgrade, but a fundamental shift in how finance teams create value.

91%of Estonian CFOs consider automation and digitization a top priority for finance.PwC CFO Survey, 2022

Descriptive

What happened?

Diagnostic

Why did it happen?

Predictive

What will happen?

Prescriptive

How can we make it happen?

The Human Intelligence

Behind every data point is a decision-maker. This research chose depth over breadth – 13 semi-structured interviews with the people who actually build, buy, and use analytics in Estonian finance teams. A qualitative approach to bridge the gap between theoretical frameworks and organizational reality.

Interviews conducted February–March 2023. Analyzed using thematic analysis methodology (Braun & Clarke, 2006).

groups

13 Expert Interviews

In-depth qualitative analysis with senior finance and analytics professionals across the Estonian technology sector.

account_balance

7 CFOs

Omniva, Nortal, Veriff, Pipedrive, Salv Technologies, OIXIO Digital, Fibenol – strategic decision-makers leading finance transformation.

query_stats

6 BA Consultants

Trinidad Wiseman, Flowit, Intelex Insight, Datafruit, Infovara, Telema – implementation specialists across industries.

4

Thematic Blocks Identified

Braun & Clarke thematic analysis

Critical Thesis Findings

01

Prerequisites for Effective Analytics

“First you need digitally structured data, then you need data quality, then you need data analysis, then you need to automate.”

– BA Consultant

  • Single source of truth – Analytics data must flow into one unified warehouse. Multiple “truths” across departments kills trust in the numbers.
  • Data quality before everything – Every interviewee stressed this. If the accounting foundation isn't solid, no BI tool will save you.
  • Integration across sources – FP&A value multiplies when financial data is combined with sales, CRM, warehouse, and external data.
  • Excel isn't dead – It lost its role in data collection and reporting, but remains irreplaceable for fast ad-hoc financial analysis.
  • Leadership must champion data culture – Technology alone achieves nothing. Data-driven decision-making requires visible executive commitment.
“The solution itself gives you nothing except the opportunity to build a culture around it. I've seen companies where the solution exists, but no culture forms around it.” – BA Consultant

FP&A Activities

Budgeting – Transformed

BI enabled the shift from fixed annual budgets to rolling forecasts. Centralized data warehouses eliminated spreadsheet chaos.

“Many of our clients use rolling budgets, checking each month how the budget aligns with reality and making decisions based on that.”– BA Consultant

Forecasting – Still Human-Driven

Predictive ML models remain rare in FP&A. Finance leaders doubt algorithms can capture full business complexity.

“Forecasts can't give you certainty about the future based on the past. A human always has more information.” – CFO

Reporting – Biggest Wins

Self-service BI automated routine reporting and ended the eternal debate of “whose numbers are right.”

“The days when five people had five different numbers for the same thing – that's gone.” – CFO

Challenges

“Everyone wants to put all the world's data in there, but it's never that simple. The important thing is to define what you actually need.”

– CFO

  • Data fragmentation across systems
  • ROI justification for advanced analytics
  • Skills gap and limited knowledge of possibilities
  • Report overload creating noise instead of clarity
  • Culture over technology – tools gather dust without leadership
“There's no real out-of-box solution you can deploy for free. ML projects typically start at €50,000 minimum. You have to be ready to write off that entire amount if the project fails.” – BA Consultant

Opportunities & Future Technologies

AI/ML Democratization

Analytical tools are rapidly adding ML capabilities previously accessible only to data scientists.

“Within this decade, machines won't replace people – but people who use technology will replace those who don't.” – CFO

NLP & Conversational Analytics

Natural language interfaces entering every FP&A tool – querying data in plain language instead of SQL.

“Imagine a ChatGPT-like assistant inside your FP&A tool that helps build reports, gives recommendations – that could be an enormous help.”– CFO

Step-by-Step Adoption

Start with quick wins, prove value, then expand. Avoid ambitious projects that try to do everything at once.

“It's better to take small steps. Even adopting a new tool is actually a small step.” – BA Consultant

Business-Side Ownership

Analytics projects fail when owned by IT alone. Finance must lead – they understand the data's business context best.

“Looking at our projects, well over half are led by the CFO. They have a good overview of all processes – sales, production – at least from a financial perspective.”– BA Consultant

Strategic Evolution: 2024 → 2026

How do the thesis conclusions hold up against two years of explosive AI development? Each finding benchmarked against current industry data.

Analytics MaturityConfirmed
2024 Thesis Finding

Most Estonian companies are stuck at descriptive analytics. Predictive/prescriptive analytics is practically non-existent in FP&A.

2026 Reality

Still true. 53% of FP&A teams use no AI in any process. Only 10% use AI for forecasting. Most "AI adoption" means ChatGPT for text tasks, not ML models.

Source: FP&A Trends Research Paper 2025

AI in Standard ToolsConfirmed
2024 Thesis Finding

BI platforms like Power BI and Tableau are beginning to add ML/AI features natively. This will accelerate.

2026 Reality

Every major vendor shipped AI. Copilot in Power BI and Excel (Agent Mode, Jan 2026). Tableau Agent. SAP Joule (2,400+ skills). Oracle finance agents at no extra cost.

Source: Power BI May 2025; SAP Business AI Q4 2025; Oracle AI Financials

NLP ChatbotsConfirmed
2024 Thesis Finding

Natural language interfaces will replace SQL queries in FP&A tools.

2026 Reality

Happened faster than predicted. Microsoft Copilot reached 15M paid seats and 33M active users. Average user saves 108 hours per year.

Source: Forrester TEI Study; Microsoft Copilot Statistics 2026

Self-Service AnalyticsPartially Confirmed
2024 Thesis Finding

Self-service BI will become the standard. Business users will do their own analytics.

2026 Reality

Tools are ready, but adoption remains stubbornly low. Only 25% of employees actively use BI tools – barely changed in seven years.

Source: BARC BI & Analytics Survey 26

Tool ConsolidationConfirmed
2024 Thesis Finding

Too many separate tools. The market will consolidate toward integrated platforms.

2026 Reality

Major consolidation driven by PE. OneStream acquired by Hg for $6.4B (2026). Anaplan taken private for >$10B. Every vendor building unified platforms.

Source: OneStream/Hg Press Release, April 2026

Data QualityConfirmed
2024 Thesis Finding

Data fragmentation and quality issues are the biggest barrier to analytics adoption.

2026 Reality

BARC's 2026 Trend Monitor: "Data Quality Beats AI Hype" – data quality management reclaimed #1 position. 64% cite it as their top challenge.

Source: BARC Trend Monitor 2026; Precisely Data Integrity 2026

Skills GapConfirmed
2024 Thesis Finding

Lack of analytical skills is a critical barrier. Companies need data literacy at all levels.

2026 Reality

Gap has worsened – 87% of CFOs say AI will be critical, yet three-quarters of organizations lack the skills. GenAI is partially bridging the gap.

Source: Deloitte Q4 2025 CFO Signals; PwC AI Jobs Barometer 2025

Cost BarrierOutdated
2024 Thesis Finding

ML projects cost at least €50,000 – prohibitive for Estonian-sized companies.

2026 Reality

Costs dropped dramatically. No-code ML platforms offer predictive analytics from hundreds/month. GenAI tools do ad-hoc analysis for $20/month.

Source: Fortune Business Insights No-Code AI Market 2025

CFO as ChampionConfirmed
2024 Thesis Finding

CFOs are the critical drivers of analytics projects. They bridge business needs and technology.

2026 Reality

Amplified. 87% of CFOs predict AI extremely important for 2026. Dell CFO deployed AI agents for reconciliations – 11,000 headcount reduction.

Source: Deloitte Q4 2025 CFO Signals; Fortune, March 2026

Recommendations for Finance Leaders

01

Start with data foundations, not tools. Invest in data quality, cleansing, and integration before purchasing analytics software. The accounting system is your primary data source – if it's not solid, nothing built on top will be trustworthy.

02

Begin with quick wins in descriptive analytics. Automated reporting, self-service dashboards, and a single source of truth deliver the most immediate value. Build organizational confidence before moving to predictive models.

03

Appoint a finance-side analytics champion. Analytics projects fail when owned entirely by IT. The finance team has the best understanding of which data matters and how it connects to business decisions – they must lead.

04

Evaluate predictive analytics through specific, high-ROI use cases. Don't build a universal forecasting model. Start with targeted problems: accounts receivable prediction, cash flow forecasting, customer payment probability.

05

Stay current and experiment. The technology landscape is moving fast – tools that required data scientists two years ago now have natural language interfaces. Companies that adopt early build compounding advantages.

7/10

Confirmed

The thesis was remarkably prescient. Two years of explosive AI development validated the core findings.