AI Copilots in SaaS: The Value Multiplier

This isn't a retraction.

Last month, I challenged the recent opinion that Software-as-a-Service (SaaS) platforms would be replaced by AI agents.

I'm not rolling back on that position—for now, I believe SaaS is safe.

I don't think we are ready to see AI agents at the core of a bank's technology estate until there has been a transformative shift in data management capabilities, as well as deep regulatory understanding.

I argued that the broad functionality and complex logic of these banking platforms is the moat that protects them from being collapsed. My wider objection though is that SaaS is designed to reduce complexity. Given the lack of specialist skills at most banks, a move to agentic AI would do the opposite, increasing complexity.

However, AI embedded within SaaS has huge potential to help drive the value of modern banking platforms, making them more impactful.

What's more—this additional value can be calculated.

The baseline value of a SaaS platform

The promise of SaaS versus self-hosted or on-premise platforms is the chance to outsource the complexity of managing infrastructure to a trusted partner.

It’s about delivering value to a customer, and not just a product.

In banking terms, the value is simple. It’s delivered through reducing resource requirements, minimising headcount, streamlining process, freeing time, and enabling margin to be reinvested.

I believe the existential threat to SaaS businesses is not the risk of being replaced by agentic AI; it's not being able to deliver enough incremental value to clients. As technologies become more standardised, banking teams become better educated, and more competition enters the market, it will be important for SaaS businesses to continue delivering more value in order to maintain its customer base.

Putting a figure on the value

When trying to articulate value, it's best to do this in hard currency – particularly in financial services. Cost savings. New revenue. Return on Investment (ROI).

Something that’s easily understood in a business case.

For SaaS core banking systems, it's actually not that hard to quantify.

Let’s take a mid-tier UK bank. Somewhere between ￡10 and ￡50 billion in assets. A nationwide presence, focusing on retail and small business customers.

Metro Bank (UK) , The Co-operative Bank plc , or a TSB Bank – that sort of size.

Assume a technology division of, say, 200-250 people that have built and maintained in-house technology solutions, but also have experience in adopting newer SaaS solutions in non-critical functions.

Within the technology division, there are myriad tribes tasked with keeping the bank's systems running. Each impacted differently by the adoption of SaaS:

• Infrastructure and DevOps Team: Manages platform infrastructure, and reliability.
• Software Development Team: Develops and maintains software—obviously!
• Cybersecurity and Risk Team: Secures systems, monitors threats, ensures compliance.
• Quality Assurance (QA) and Testing Team: Ensures the quality of software.
• Business Analysts (BA) and Product Management: Manages product requirements.
• Project Management and Coordination: Oversees technology projects.
• Data and Analytics Team: Handles data management, analysis, and reporting.

It’s the reduction, removal, or repositioning of these roles that creates a value statement for adopting SaaS. Either removing roles altogether or reinvesting the margin into higher value areas—digital, data, product innovation.

Not all change is equal, and some areas are impacted more heavily than others.

DevOps Engineers, Cloud Architects, and Site Reliability Engineers are reduced or redefined as the responsibility for infrastructure and uptime shifts to the SaaS provider. Software Developers and Full-Stack Engineers can be focused elsewhere. The need for Security Engineers and Risk Managers is arguably reduced. QA and Testers can focus on integration rather than application testing.

However, Product Managers still manage products, Project Managers still manage projects, and whilst data infrastructure becomes a SaaS partner's service, the Analytics Team still has a job to do.

I tasked ChatGPT to crunch the data and present the role cost savings.

I asked it to “estimate the reduction or removal in headcount from the technology division based on moving to SaaS.”

I took this one step further. Based on expected salaries of the roles, I asked what the annual cost saving would be for our mid-tier bank.

The answer?

￡605,000.

That’s what the back of a napkin maths says a mid-tier UK bank should be able to save or reinvest in resource costs by adopting a SaaS core banking system. Add that to the infrastructure savings, and you've got the value created by a new platform.

Whilst the number might be notional, the maths is sound. There is already sizeable value in a move to SaaS.

Multiplying value

So how to deliver more value to SaaS customers?

AI Copilots—that’s how.

This isn't the same as an AI agent. AI agents are autonomous systems designed to perform tasks and make decisions on behalf of users, operating independently without constant human input.

In contrast, a Copilot is a collaborative tool that assists and augments human efforts. While the user remains in control, the Copilot offers suggestions, handles repetitive tasks, or provides guidance.

This will be the new move in SaaS value creation.

Like in all other areas of technology and business, AI can unlock the value growth SaaS is looking for.

Businesses are recognising this. IBM expects that 80% of SaaS applications are expected to incorporate AI technology. However with only 35% actually doing so there is a big market for first mover advantage.

Copilot Use Cases

Integrated into a functionally broad core banking platform, an AI Copilot presents a number of additional areas to further remove resource or squeeze efficiency.

Trained on internal and external data sources, they can access relational databases of a core banking system to retrieve customer records and transaction history. Then integrated to external systems from the core via API.

Like the SaaS value baseline, the additional value to a Copilot can be estimated:

• Product Development The bank's product management team is sat around a report on last quarter's retail loan performance. They want to stay competitive whilst promoting innovation, but identifying exactly what customers wanted has always been a challenge. Rather than pull in market research and crunch Excel sheets, they ask the core banking Copilot. AI retrieves the current loan catalogue, analyses repayments history and utilisation levels, and compares performance versus external customer surveys. Based on the insight, the Copilot makes several suggestions on new products to consider to target the bank's focused market.

Value Created: Ability to reduce Product Development team by 20-30% leveraging AI for market research, product management, and analysis.

• Compliance Checking One of the new suggested products is an offset retail loan that gives a preferential rate to regular saver customers of the bank. The product team needs to check with compliance to ensure regulations are complied with before the product launch. Rather than a manual review of regulatory guidelines and preparing a report for the product steering committee, the product manager consults the Copilot. The Copilot checks the specifications from the product engine against a database of relevant financial regulations, including local and international laws (e.g., the Consumer Protection Act, GDPR, etc.). It then generates a compliance report that outlines any potential issues and creates an email to share for sign off.

Value Created: Facilitates quicker review and straight-through check process reducing compliance team workload by 40-50%.

• Data Analytics The bank's Head of Data has been tasked with presenting the annual performance review to the Board. The board expects a comprehensive presentation that outlines the bank’s product performance, key operational metrics, and projections for the upcoming quarter. Traditionally, she would spend days manually collecting and analysing data from various sources, running models to forecast future performance, and creating visual reports for the presentation. The Copilot automatically analyses and presents data within the core. Layered across external sources, it detects upward trends in customer demographics, transaction types, product usage, and key projections.

Value Created: Simpler creation of forecast data and analytics removes the need for business analysis resources, streamlining the data analytics function by 30-40%.

If we return to our napkin maths, the additional value could be as much as the saved support costs of running SaaS in the first place.

Particularly in non-core engineering and infrastructure roles like product, business analysists, compliance, and support staff.

AI Won't Collapse SaaS, its the Next Stage of Delivering Value

You might now be saying I was wrong and AI agents will be next to collapse SaaS.

I'm still saying they won't.

For the foreseeable future of core banking, humans aren't going anywhere, which means AI agents won't be taking over.

The technology strategy should focus on simplicity and reducing complexity in delivering core systems to allow the focus to be on banking customers. This can be achieved first by moving this infrastructure to a trusted SaaS partner, and now through integrated AI solutions like Copilots.

Adoption and understanding of SaaS has already transformed financial services. The growth in generative AI has been far more disruptive though. 麦肯锡 estimated that in the 12 months following ChatGPT launched in 2022, large enterprise spending on AI solutions was already at $15 billion. This represented about 2% of all global enterprise software—it took SaaS four years to reach the same market share.

SaaS businesses need to recognise what enterprise clients have and make sure AI is central to their business model.

AI won't kill SaaS; it's going to super-charge it.