Episode 2: AI in Finance: What changes, what stays, what’s next

In the second episode of The Thought Line, we explore how AI is reshaping the finance function; from reporting the status quo to predicting what comes next. Joined by Swatantra Bhatia, Partner, Accounting and Outsourcing at Forvis Mazars in India, the conversation examines why many organisations remain stuck in the “pilot phase” despite strong intent to scale up. From data quality and fragmented processes to questions of trust, governance, and scalability, the barriers to adoption go far beyond technology.

The episode also touches upon how organisations are leveraging integrated digital platforms such as Forvis Mazars in India’s AOS Cockpit to improve visibility, streamline compliance workflows, and enhance client engagement.

Tune in for a practical perspective on moving from experimentation to enterprise-wide impact. 

Listen on:

YouTube

Spotify

Episode transcript

Hosts: Rajita Rajan and Anukriti Jain
Guest: Swatantra Bhatia, Partner, Accounting & Outsourcing, Forvis Mazars in India

Introduction

Rajita Rajan:
Think if you could look at your financial data not just to understand what happened, but to know what’s likely to happen next. Imagine systems that can flag risks instantly, explain variances in real time, and recommend actions proactively. That’s the promise of AI in finance.

Yet while many organizations are experimenting with AI, very few have managed to scale it successfully. Challenges around fragmented data, inconsistent processes, trust, and governance continue to slow adoption.

You’re listening to The Thought Line, a Forvis Mazars podcast where we explore the ideas and insights shaping industries and economies across the globe.

Rajita Rajan:
Good morning, listeners. Welcome to The Thought Line. I'm Rajita Rajan.

Anukriti Jain: 
And I'm Anukriti Jain, your host for today. In this second episode, we'll dive into a topic that is rapidly reshaping the finance function, the role of AI in driving smarter, faster, and more forward-looking decisions. To help us explore this, we're joined by an expert who works closely with organizations navigating finance transformation.

We're delighted to welcome Swatantra Bhatia, Partner Accounting and Outsourcing, Forvis Mazar's in India. Swatantra is a seasoned financial leader with over two decades of experience and global exposure, combining CFO level insight with deep technical expertise. He has been closely involved in advising organizations on leveraging AI to transform finance functions and drive scalable digital adoption.

Swatantra, thank you for joining today. 

The Current State of AI in Finance

Swatantra Bhatia:
Thank you for having me on board. 

Anukriti Jain: 

AI in finance is no longer just about efficiency. It's increasingly shaping decision-making, risk, and competitive advantage. In this episode, we'll try to unpack where organizations really stand today, what's holding them back, and what it takes to move from experimentation to scale. And hopefully, by the end of this conversation, leaders will have a clearer sense of how to approach AI in their own finance function.

Swatantra, let's start by discussing where we are today with AI in finance. Are we still in early adoption, at an inflection point, or already seeing it go mainstream? AI in finance is no longer a futuristic conversation now. It's already there.

Swatantra Bhatia:

AI in finance is no longer a futuristic concept; it’s already here. Many companies, they are already building solutions using AI and also using it in their finance operations. But finance is not like marketing or operations, where you can do some experiment, because if you are making some mistakes, you will be caught in the audit in the end. So therefore, many finance functions in industries, they are using AI, but they are also cautious about how to implement it more carefully, rather than doing the experiments.

AI Adoption in India: Organized vs. Unorganized Businesses

Anukriti Jain:
Okay, so what about markets like India, where finance spans both highly organized enterprises and a large unorganized sector? How do you see AI adoption playing out across these two worlds? 

Swatantra Bhatia:
In a market like India, it's even more interesting, because you are dealing with organized companies, you are also dealing with semi-organized or unorganized companies. I feel that for organized companies, it will be a layer above what they are already doing, because they are using ERP, they already have data governance, they already have processes in place. For them, it will become an addition to what they are doing, whereas for non-organized or unorganized industries or companies, it will be basically how to deal with their headaches.

They will use it more as if they have a problem, they want to solve a problem and they will do it. Whereas if you also see with organized industries, since they have a long term decision process, they have a lot of governance. Therefore, the implementation will be more challenging.

I feel that for unorganized companies, it will be easier to adopt AI rather than for organized, because for them it's just a decision making and they can do it on smaller scales rather than going for big projects.

What Does “Semi-Organized” Mean?

Anukriti Jain:
You mentioned the term “semi-organized.” How is that segment different from a fully organized business? What are some challenges that you face while dealing with such players? 

Swatantra Bhatia:
Organized, I would say there are big companies, they already have ERPs, they already have BI tools, they are well structured in terms of tools governance.

Then you have unorganized, they are not into mainstream, I would say, and they are dealing with their processes on spreadsheets, they are taking decisions based on information which comes on WhatsApp. And then you have, I call it semi-organized. So for me, semi-organized is, they have ERPs, but they don't have right processes, they don't have BI tools to enable them to take a right decision based on data which is available with them.

Why Many AI Initiatives Fail to Scale

Rajita Rajan:
My next question to you is, we see that many organizations want to adopt AI, but struggle to move beyond pilots. Is the real constraint technology or deeper issues like budget constraints, data quality, fragmented processes and lack of standardization? 

Swatantra Bhatia:
I think most companies, they feel that they have AI problem, they actually have operating model problem like you are saying. It is more of a data fragmentation, non-standardised processes.

These are the main problems. So, before your data is inconsistent, it is fragmented. AI does not fix it.

It is basically act as a magnifier which exposes or which highlights it very efficiently that there is a problem. I will give you one example. We have a client where we have implemented AI-based expenses classification tool.

We were expecting, client was happy, we were expecting that this tool will work efficiently. But when we started using this tool, we realized that they do not have any standard processes. They have multiple departments and each department they were using rules differently.

AI was able to classify based on rules which we set up for them. But when we dig into the problem, we realize that since the processes are not standard, that is why the outcome was not efficient. When we started working with them to streamline the processes, the AI started giving right results.

It is a problem of right processes, it is problem of not having standard data, standard processes and fragmentary data. 

Governance, Auditability, and Trust in AI

Anukriti Jain:
You might trust AI in a control setting, but scaling it across the organization is a different story, right? And that is where trust becomes critical, especially in a function like finance, where accuracy and accountability are non-negotiable.

This brings us to our next question. How should organizations think about compliance, governance and auditability when AI starts influencing financial decisions? 

Swatantra Bhatia:
I strongly believe finance cannot operate on black box decision making, as of today at least. There are many systems which are being used in finance, no one knows what is happening behind them.

But since you can audit them, you can also check the consistency of data, that is why finance professionals, they believe in those systems. Unless there is proper traceability of what is happening behind the AI tools, there is auditability. I do not think finance function will trust the system.

As I mentioned earlier, it is not like other operational functions where you can experiment and many decisions, important decisions are being taken by companies based on the data. So there has to be proper governance, there has to be data which can be audited later on, then only it can be fully adopted in finance function. 

Will AI Replace Outsourcing?

Rajita Rajan:
As AI reduces manual work across finance, tax, and audit functions, what does this mean for the outsourcing industry?

Swatantra Bhatia:
There’s a perception that AI will reduce outsourcing demand, but I believe it will fundamentally reshape how outsourcing teams create value.

Today, outsourcing is often volume-driven. Going forward, it will become increasingly value-driven.

There will still be a strong need for human involvement because organizations are unlikely to rely entirely on AI-generated outputs for critical decisions.

The role of professionals will evolve toward validating AI outputs, interpreting insights, and supporting strategic decision-making.

So rather than eliminating outsourcing, AI will redefine how service providers support their clients.

Skills That Will Matter in an AI-Driven Finance Function

Rajita Rajan:
What skills do finance professionals need to build in this evolving landscape?

Swatantra Bhatia:
The importance of transactional skills will reduce over time because many repetitive tasks can be automated.

What will matter more is strong financial understanding, analytical thinking, and the ability to interpret AI-generated insights.

Professionals who can combine domain expertise with strategic thinking will become increasingly valuable.

The future finance professional will need to focus less on manual processing and more on decision-making.

What Will Finance Look Like in the Next Five Years?

Anukriti Jain:
There’s a lot of discussion about fully AI-driven finance functions. How do you see the next five years evolving?

Swatantra Bhatia:
There is definitely a lot of hype around the idea that finance will become fully AI-driven very quickly.

I don’t believe that will happen in the near term.

Finance functions move carefully because decisions are often based on multiple layers of validation, governance, and accountability.

What we will see instead is:

• More real-time finance operations
• Greater predictive decision-making
• Improved use of structured data
• Faster access to insights

But complete automation of finance remains a long-term possibility rather than an immediate reality.

Organizations should focus on adopting AI responsibly, with strong governance and controls in place.

At the end of the day, in finance, being right is still more important than simply being fast.

How Forvis Mazars in India Is Approaching AI

Anukriti Jain:
What has been your own experience at Forvis Mazars in India when it comes to digital transformation and AI adoption?

Swatantra Bhatia:
Since we are serving many clients and outsourcing, we also took a decision last year to use AI and to digitalize the processes, how are we dealing with our clients. 

Right now, we have full digitalization in outsourcing service line where we are able to interact with our clients digitally, we are able to provide them services which are more digital, we are able to extend the tools which are helping them to make their processes more digital like the forensic analysis, like big data transformation. 

All these things are being implemented, being also extended to the client which is appreciated by the client which is also helping us or our team to understand AI and especially the transformation and digitalization through AI.

At the same time, we are also quite cautious about protecting the data of our client. So obviously we have human in kind in process loop where our team members they always take decision after reviewing the data even if it is prepared by using AI. 

Which Industries Are Moving Fastest?

Rajita Rajan:
Which sectors are moving fastest in AI adoption, and which ones are likely to lag behind?

Swatantra Bhatia:
Leadership mindset matters more than industry.

Sectors such as e-commerce and SaaS are naturally moving faster because they already have large volumes of structured data and a clearer path to ROI.

Traditional industries like manufacturing may adopt AI more gradually.

However, even within the same industry, adoption levels can vary significantly depending on leadership priorities.

For example, one of our e-commerce clients used AI to solve large-scale order-to-cash reconciliation challenges very effectively, while another client in the same sector remained much slower in adoption.

Ultimately, leadership willingness to experiment and solve problems proactively makes the biggest difference.

Advice for Leaders Driving AI Adoption

Anukriti Jain:
How can leaders successfully drive AI adoption and ensure it delivers measurable value?

Swatantra Bhatia:
The first step is recognizing the underlying business problem.

AI should not be implemented simply because it’s trending. Leaders need to identify operational inefficiencies first and then determine where AI can genuinely create value.

I strongly recommend an agile approach:

• Break large problems into smaller initiatives
• Run focused pilot projects
• Learn from small successes
• Scale gradually

Many organizations fail because they launch large AI programs without fixing foundational process issues.

Starting small reduces risk, improves learning, and builds organizational confidence over time.

Closing thoughts

Rajita Rajan:
With this we come to the end of our discussion. Thank you Swatantra for sharing your perspectives with us. 

Swatantra Bhatia:
Thank you for having me on this podcast. It was an interesting discussion. 

Key thing to remember is, AI in finance is not just about the efficiency. It is about improving decisions and finance by design does not move fast.

It moves carefully. The goal is not to rush adoption but to adopt in a way that actually strengthens trust control and insight because at the end of the day finance being fast is good but being right is still more important. 

Anukriti Jain:
To our listeners we hope this conversation has given you a clearer view of where AI in finance stands today and what it means for the road ahead.

Rajita Rajan:

Stay tuned for more conversations on The Thought Line, available on YouTube, Spotify, and Apple Podcasts.

If there are topics you’d like us to explore, write to us at: thethoughtline@mazars.co.in

This is The Thought Line — a Forvis Mazars podcast where ideas meet insight to shape what’s next.

Credits

This episode was brought together with research and support from:

Devesh Srivastava

Kanishka Gulati

Simran Sadana

Editing and recording support by:

Taranpreet Valjot

Swati Goyal