Few leaders would dispute the fact that business today is driven by data and smart algorithms. Yet, rather than real digital transformation, many instead pursue digital incrementalism, using automation to cut costs or, worse — cut jobs. Doing so might buy you some time from impatient shareholders, but it will be short-lived unless you can face the challenge: How do you reimagine what you do for a new era of AI-powered competition?
The high unemployment numbers of the Covid-19 recession have obscured a systemic problem: the accelerating effect of automation on the workforce. We have been here before. In each of the last few recessions, there have been significant spikes in labor-replacing automation. Although salaries may fall in a crisis, shrinking revenues and the impact on the bottom line typically drive companies to invest in new technology rather than hiring people.
For economist David Autor, the 2020 employment crisis will be further exaggerated by what he calls “automation forcing.” In his view, social distancing requirements and stay-at-home orders may drive temporary labor shortages, forcing firms to leverage emerging technologies to get things done with fewer people — whether it be “fewer workers per store, fewer security guards and more cameras, more automation in warehouses, or more machinery applied to nightly scrubbing of workplaces.”
That is one grim scenario, certainly, but not an inevitable one. One way out of the dystopian cycle of automation and job loss is if more organizations can harness technology to reimagine work — rather than merely replace it. To do that, we need to consider how AI and machine intelligence can enable disruptive business ideas and customer experiences, unlock new ways of working, and augment teams to innovate and solve problems more effectively.
Reimagine the customer experience.
One of the best examples of a traditional organization leveraging technology to disrupt an adjacent market is Marcus by Goldman Sachs. Marcus, a digital consumer bank, might be an unlikely spin-off from a traditional investment firm. Not so, according to Harit Talwar, global head of Goldman Sachs’ Consumer Business, who told me that these days the business sees itself as a “150-year-old startup.” At a time when many banks are leveraging basic automation to cut their operating costs, Goldman approached the digital transformation challenge differently. Rather than patching a broken system, they asked: What do people want?
After speaking to more than 10,000 consumers, Talwar and his team identified that people had three big pain points with typical retail banks — a fragmented and confusing relationship with money, opaqueness around the borrowing process, and frustration at the lack of respect for their savings. That insight was useful, but what compelled Goldman Sachs to act was knowing that it didn’t have to replicate the old banking models to compete.
“We did not need to set up hundreds of thousands of branches, or hundreds of thousands of feet on the street doing face-to-face selling,” Talwar explained to me. “The digital technology, the programmatic data analysis, the simplicity of interface design now make it possible to acquire and serve millions of customers, including mass affluent customers, in a simple and transparent way.”
Marcus may operate more like a tech company, but don’t try and tell that to Talwar. In his view, while engineering, data, and design are potent and vital ingredients of modern business, the real focus has to be elsewhere. “We don’t call ourselves a technology business; our business is solving customer problems.” For Talwar, AI is just a capability — the real future of finance is extreme customer-centricity.
“If you want to be a successful disruptor, whether starting a new business or whether as a decades-old organization, the first lesson is to ask: What it is that you’re trying to do, and for whom? What is the customer problem? Or what is the business problem you’re trying to solve? That is real innovation.”
Reinvent how you work.
The second challenge for leaders is identifying new ways of getting things done. While repetitive workflows and routine transactions are typically the first to be automated, machine intelligence is now starting to encroach on the more complex decisions previously reserved for humans. Rather than a threat, we should see that as an opportunity to revisit how we work, and why.
At UBS, harnessing AI is the cornerstone of group CIO Mike Dargan’s overall digital transformation plan. He explained to me that in the last few years, diverse AI projects have been appearing across the bank — from fraud detection to compliance, risk management to advanced HR analytics, and a new system that facilitates foreign exchange transactions. UBS’s digital transformation objective is to reimagine the bank’s entire value chain, from how they serve customers and produce investment strategies, to middle and back-office tasks.
The common thread that joins the AI projects at UBS is a new perspective about the kind of work machines should do, and where humans add the most value. According to Dargan, “As we automate away the simple tasks, roles become more sophisticated.”
Dargan gave an example of the growing challenges of managing the firm’s complex network infrastructure, which generates thousands of logs daily. Rather than monitor these manually, they now use AI to read the system alerts, and Natural Language Processing algorithms to preemptively identify serious issues. True, that is work that people might do. But in Dargan’s estimation, it would have taken a team of at least 10,000.
As it is at many big organizations, the employment impact of automation at UBS is a nuanced one. Machines are handling more work, but arguably without a high level of automation, UBS employees would find it hard to get their jobs done. The firm now has more than 2,000 software bots operating across the business, growing steadily. During the pandemic, they even created six new bots in just three days, which were needed to assist client advisors in handling a massive flood of Swiss Covid-related loan requests. Telemetry — early warning and anomaly detection — followed by automation and self-fixing solutions, supported the firm’s stability, which was experiencing quadruple peak volumes because of volatility and volume in the market.
The digitization of financial services changes both the way people work, as well as how they interact and partner with other organizations. As in other parts of the economy, such as retail and logistics, banks will have to become platforms to grow and compete — no easy task for conventional players with creaky infrastructure and a conservative mindset. However, the prizes are substantial for those companies that can get it right.
Take Apple Card. A controversial but dramatic product introduction — described by Goldman Sachs CEO David Soloman as the “most successful credit card launch ever.” One of the factors that supported the accelerated timetable was that Apple Card was developed and released in an entirely cloud-based production environment. Think of it as the difference between traditional banking and “banking-as-a-service.”
Talwar describes their “banking-as-a-service” platform as a competitive moat and believes it is characteristic of how the firm plans to defend their position against traditional retail competitors with a differentiated technology stack that can scale, be agile, and remain relevant. Rather than build thousands of retail branches or rely on conventional marketing to acquire customers, Marcus has been able to leverage a technology stack based on API microservices architecture to build distribution partnerships with Apple, Amazon, JetBlue, and Intuit. In a sense, these are all relationships based on data sharing, intermediated by machine intelligence. In the case of Amazon, Marcus provides revolving credit lines to Amazon merchants, governed by data from their e-commerce trading activities.
From this perspective, you can arguably view the entire Marcus retail bank as merely an application running on the Goldman Sachs digital banking platform. Goldman, which has expressed an ambition to create its own “Financial Cloud,” is now looking to extend its reach into other parts of the financial ecosystem by providing customers with APIs into its transaction banking and risk management platforms.
Rethink your capabilities.
Finally, rather than using AI as a blunt tool to reduce headcount, we have to train people to use machines to change their work. After all, what is more valuable: people capable of doing their jobs, or team members who can design systems, train AI models, and build bots to do their team’s work? It’s a familiar story. Whether it be factory automation or the early days of the computer revolution, staying one step ahead of our tools has been the story of human co-evolution with technology since the beginning.
As Dargan at UBS puts it, “Banking is technology, but technology is people.” In 2019, the bank rolled out a digital learning curriculum across the entire firm, providing educational content on AI, blockchain, and cloud technologies. In the first 6 months of 2020, his tech-teams have clocked over 45,000 training hours, with 50,000 courses available. While he doesn’t necessarily see a future where everyone in the bank can code, digital literacy is now an essential skill. No job or function is immune to the coming changes, even technology roles. Over the last two years, UBS has trained 350 people in the operations space to design and manage automation bots — a work profile that even didn’t exist before.
Aside from training people with new skills, reimagining work also requires you to consider how teams collaborate. Marcus may be a digital bank, but that doesn’t necessarily avoid human interactions being stubbornly analog and siloed. At Marcus, a productivity breakthrough came when they re-organized their teams into agile pod structures. Now, regardless of their functional role as an engineer, a marketer, or a lawyer, Marcus employees are attached to workstreams focused on specific objectives — improving the customer onboarding journey, for example.
The agile structure can create challenges when people try to balance tactical goals with long term vision. In Talwar’s view, that’s where leaders can add the most value. They need to manage the trade-off between telling people what to do (like a conductor controlling an orchestra) and setting a common objective with some ground rules so that teams can find creative solutions themselves (like a self-organizing flash mob).
We are just at the beginning of a new era of AI-powered competition, and the playbooks for organizations and leaders are far from clear. One thing is sure: the successful firms of the future will be those that can leverage data, algorithms, and human talent to both sidestep industry boundaries and creatively meet customer needs.
For leaders of more established firms, this is no time for timid moves. Expect a widening gap between customer-centric organizations with a deep commitment to evolving their technology platform and those whose blind pursuit of operating efficiency leaves them defenseless against a more uncertain future. In the end, our best chance at reinvention is to answer a deceptively simple question: What is possible now in an age of smart machines that was not even conceivable before?