EXCLUSIVE: “Less Dev, More Ops…” – Daniele Tonella, ING in ‘Discover Money20/20’

As Group CTO at ING Bank, Daniele Tonella is responsible for 20,000 engineers whose creativity is being harnessed to build an architecture that’s simple by design

Spending more on technology alone does not lead to better performance. Fact.

According to a report by McKinsey in 2024, banks have been increasing their investment in IT by nine per cent a year on average, reaching roughly the GDP of Belgium and Sweden by 2023 at $650billion. But that investment outpaced banks’ average annual revenue growth of four per cent.

IT spend can be a bottomless pit – especially if you’re trying to compensate for successive waves of upgrades that haven’t quite delivered what you hoped. But modernising the banking IT estate doesn’t have to follow the law of diminishing returns. A survey of the world’s largest banks by Bain & Co, found the most successful IT strategies delivered an average five percentage points higher total shareholder return, 10 percentage points lower cost-to-income ratio, and 12 points higher net promoter score than their peers. Those are much more positive statistics.

Bain found the most successful strategies were based on ‘sustained simplification, low cost and adaptable processes’. But that does require ‘a persistent focus by the C-suite, an excellent engineering function, commitment to transform the business and not just the technology, and consistent investment over time’, it said.

‘Simple’ doesn’t mean easy, then!

Daniele Tonella, who joined Dutch bank ING as Group CTO in August 2024, would no doubt agree. His role as head of a 20,000-strong engineering team, requires him to master three critical considerations – time, complexity and personnel.

“I’m the one that permanently has to hedge between the short-term and the long-term,” Tonella explains. “We are a very entrepreneurial bank that wants to go very fast, but in tech, if you go too fast without thinking long-term, you end up building up technical debt and future issues.

“At our size and with our history, we sit on a very large estate and complexity tends to grow by itself, so one of my focusses is to ensure that we keep it at bay.

That sometimes requires him to rein in developers’ natural enthusiasm to innovate. He clearly has huge affection for engineers, describing them as an ‘extraordinary force’.

“But they also need to be focussed and converged into the strategic direction that we are taking. That’s a human journey and it’s probably the most fascinating part of the story,” says Tonella.

It’s required a mind shift, away from the heady days of software developers taking priority in the DevOps cycle towards one that’s more balanced towards IT operations as the two teams work together to build and maintain platforms.

“DevOps has been a sort of freedom from infrastructure, but now, that journey is coming to an end,” says Tonella. “With a dev-centric DevOps, you end up creating complexity that just keeps growing, because we are engineers and, if you get three engineers in a room, you get four opinions, and that creates a fragmentation of your underlying stack.”

Platform engineering by contrast, requires a more collaborative approach, with a level of robustness that reduces complexity by design.

“If I’m a developer and the pipeline is designed by platform, if I don’t comply with some of the controls, it just comes back to me,” says Tonella. “It’s like when you compile code. If there’s a syntax error, you can compile as many times as you want, but you don’t get something you can use.”

Experts are optimistic about the efficiencies that a platform engineering approach can bring, but it’s not necessarily easy to implement.

“Platform engineering, is a strong layer and allows us to converge on complexity,” says Tonella. “That requires a huge culture change because it means we are moving away from the full stack engineer, away from the super-engineer who does everything from the front to bare metal, and instead towards a specialisation of engineers.

“You could almost say there are three families of engineers. There are the ones that do business features, who are really focussed on client impact. There are the ones that build the components for the first ones to glue together into services, and then there are the ones that manage the platform. Somebody said that culture is the operating system of a team. Well, then platform engineering is the kernel.”

“At our size and with our history, we sit on a very large estate and complexity tends to grow by itself, so one of my focusses is to ensure that we keep it at bay”

He defines ING’s approach to developer recruitment as ‘not looking for talent obsessed with toys’, but engineers who understand the framework within which they are operating, and seek creative solutions within its parameters.

The accumulation of these changes at ING has led to what Tonella describes as a ‘virtuous circle of attractiveness’ to engineers, one in which ‘the team spirit is extremely strong and you really feel, “I can solve serious problems here”.’

“There is a challenge, of course, in being a technology company with a legacy, with a history, and with a size,” he adds. “But if you’re an engineer who wants to create a technology platform of the future that is built to reduce complexity while the business is exploding, then we are the right place to be. That story is resonating.”

ING’s selective use of genAI is helping in these engine rooms of change.

“ING has been very proactive in adopting AI and genAI, but doing it in a smart way,” says Tonella, “avoiding a proliferation of small ideas all over the place just because ‘it’s cool’, and concentrating instead on five key areas: KYC, hyper-personalisation, contact centres/chat bots, a specific part of wholesale banking, and engineering.”

The transition of AI from fancy to functional is being echoed industry-wide, with a McKinsey article published late last year urging banks to ‘move beyond experimentation to transform critical business areas, including by reimagining complex workflows with multi-agent systems’. As in all areas of advanced intelligence there’s a lot of myth about the impact on the workforce, says Tonella.

“There is this hype that AI is coding faster than coders, but that’s only a part of the truth.” In reality, the amount of time they spend coding is proportionally much less than widely thought, he says. “[So] where we see AI helping engineers is in reducing what we call the cognitive load.

“At our size, complexity, and history, when an engineer touches something, they have to take into account a lot of contextual information. AI is helping by accelerating that context analysis,  essentially saving engineers’ time. It allows us to focus our best people on interesting work, and not just menial activities.”

Cloud-focussed

The Cloud plays a key role in making all of this possible, but the way that financial institutions approach the Cloud has changed. Embracing it no longer entails the wholesale shifting of legacy infrastructure to a different environment, but enables organisations to re-imagine how their applications are structured. It’s what McKinsey has termed ‘the progressive Cloud’.

“At ING, Cloud is a core pillar of the transformation from multiple axes,” says Tonella. “The bank started many years ago to migrate away from monolithic mainframe systems, but not just monolith to monolith, like I’ve seen happening in some places, but really re-architecting the application around microservices, a heavy modularisation of functions, and global platforms and services.”

In effect, it’s using Cloud infrastructure as the agent for re-engineering systems.

“Cloud was misunderstood for a while as a sourcing strategy: if it’s not inside, it’s in the Cloud,” says Tonella. “It’s somebody else doing it [but}… in the way we set it up, it represents a different set of principles around how to architect, develop, design and deploy software.

“Public Cloud has been a help in that space, but we also create a very large private Cloud that is essentially doing exactly the same thing. “When you go higher in the stack – so not infrastructure, but software-as-a-service – Cloud, of course, has made a difference, less because of the technical functionality, but more because it’s an easier way to standardise some processes and features.

“If you take CRM systems, or workflow management systems, or HR management platforms, for example, they tend to have a very simple, out-of-the-box type of service, and most of the time, that is enough. So in that space, Cloud has been a profound driver of standardisation.”

ING works with – and invests in – many technology providers, so, for Tonella, the question ‘build or buy?’ is somewhat irrelevant.

“Even when you buy, you have to build,” he says, “because you have to integrate into your ecosystem. And the movement towards Cloud also means glueware is code. Documents are code. Pipelines are code. Everything is code.

“So, build-versus-buy is an old sourcing logic in the sense of ‘I should stop building, I should buy because it’s more efficient’. Cloud has been abused a bit in that sense.

“You could say, it’s build, buy, and rent. And all three have advantages and disadvantages. There is no one that fits everything.”

This article was published in Discover Money20/20 2o25, Page 30-31