Most intelligent automation projects that are underway or in the pipeline will fail, according to a new report published this morning by professional services firm, KPMG.
The report, Ready, Set, Fail? Avoiding Setbacks in the Intelligent Automation Race, says that this is because business and technology leaders are not yet ready to implement intelligent automation – the set of technologies that includes artificial intelligence (AI) and robotic process automation (RPA) – from the top down and at scale.
KPMG’s aim isn’t to depress the market or halt the uptake of the new technologies, but to force leaders to recognise two critical issues before attempting to drive a hard return from their technology investments.
First, deployment of intelligent automation (IA) needs to be a C-level strategy imperative, and not a tactical operational decision. And second, IA is about business and operating model transformation, and not departmental technology deployments.
“This means shifting the business and operating model from one of people supported by technology to one of technology supported by people,” claims the report. “It’s a digital-first operating model.”
But to realise these “transformational benefits”, implementations should be part of a strategy aligned with overarching business goals and pursued across the whole organisation, rather than in piecemeal improvements, it says.
Long lists, but little detail
Strategic implementation of intelligent automation yields important advantages, continues the report, including: improved customer service; empowered employees; better innovation; lowered costs; faster projects; and upgraded, standardised, and higher-quality operations. An impressive list, but lacking in critical detail and supporting evidence. As such, it reads like marketing collateral, not cogent strategic analysis.
“Piecemeal efforts that focus mainly on cutting the cost of legacy processes and reducing headcount – with, for example, siloed efforts to automate payroll, invoice processing and customer service inquiries – will not move the needle in this new world,” says the report.
Focusing exclusively on these types of internal upgrades can waste time and resources that companies could better spend on wider investment to help the organisation thrive and compete going forward, it suggests.
“Growing evidence shows that taking a strategic approach to IA by focusing early on creating new business and operating models can yield 5X to 10X dividends”, it says, and yet – again – doesn’t back this up with supporting evidence.
Among the statistics that KPMG does share is the finding that enterprise investment in the IA market – which includes artificial intelligence, machine learning and RPA – is expected to hit $232 billion by 2025, compared to just $12.4 billion today.
KPMG also carried out a worldwide executive study on the challenges of deploying intelligent automation at scale. The consultancy found that, while most companies recognise the potential value of RPA and AI, they remain unsure about the technologies and unprepared for how they will impact their employees – a situation not helped by press coverage of the technologies’ potential to sweep aside millions of jobs.
However, jobs will be lost to the technologies, says KMPG.
But some of the other challenges are more practical in nature and centred within the enterprise, found the study: two-third of respondents cited a lack of in-house talent to deal with the technologies, while half of organisations are struggling to define clear goals and objectives for deployments and accountability – when it comes to artificial intelligence, for example.
The findings imply that many organisations feel that AI is something they should have, but are unsure exactly why or what they should do with it. Despite this, over the next three years, 40 percent of respondents will increase their AI investment by 20 percent or more, says the report, while nearly half (49 percent) of organisations expect to use the technology.
As for RPA, 74 percent of companies say they will use it to some extent over the next three years (up from 16 percent currently), says KPMG. Nearly two-thirds of respondents plan to fully implement RPA within that timescale.
However, slow rates of adoption and implementation internally, coupled with looming organisational challenges, could hinder business leaders from reaping the full benefits of these technologies, it adds.
In particular, nearly one-third of leaders (31 percent) are uncertain of how RPA will impact their employees.
More, the technology mix will begin to erode the traditional organisational boundaries separating human resources, finance, procurement, and other functions, says the report, resulting in fewer isolated or vertical functions.
In total, less than one-quarter of respondents (24 percent) have intelligent automation pilot projects or proofs of concept in place, suggesting that many organisations may be planning to use the technologies without first testing the waters or carrying out due diligence.
According to KPMG, the results “underscore the need to not only act quickly but to plan deployments strategically with scale in mind.
“Most companies’ executives acknowledged they are still experimenting only with RPA, applied to legacy applications and processes,” says the company. “With such a narrow focus and a bottom-up approach, they have not positioned themselves to transform their business and operating models so they can become and remain competitive with digital-first companies.”
Not all organisations can emulate Amazon’s one-click experience, acknowledges the report. However, they can close the gap if they act quickly, understand the urgency, and define and execute a comprehensive intelligent automation strategy – one that looks not just at technology, but also at business and operating model opportunities and constraints.”
Companies should consider alternative investment strategies, such as divestitures and alliances, to “disrupt against themselves, isolating innovation from day-to-day running of the business,” it concludes.
Internet of Business says
It stands to reason that AI, RPA, and related technologies, need to be implemented strategically, and not in a tactical, cost-cutting arms race.
This is a point made by a number of recent reports (see Internet of Business, passim), which identify a mismatch between what vendors believe they’re providing – technologies to augment human skills – and what many buyers believe they’re getting – a shortcut to easy cost-savings and job-cuts.
That said, it’s sometimes tempting to wonder if companies such as KPMG have automated the production of these reports, particularly when they ignore what vendors such as Microsoft, IBM, and Google are saying: that these technologies exist to complement human skills, not replace them.
Either way, business leaders should be wary of any report that makes airy, unsubstantiated, marketing-like claims such as, “Organisations that can power up their intelligent automation efforts can radically improve operations, transform their business models, and become long-term winners,” as the KPMG one does. Or, “A boundary-less enterprise will ultimately produce a more customer focused business model”. Or, “Internet-native and highly digital-focused companies are outpacing others in reaping the benefits of such efforts right now.”
And they should be especially wary of suggestions that organisations should become focused on technology “supported by people” rather than the other way around. The days when companies re-engineered themselves to support vast, inflexible enterprise systems had just two major beneficiaries: management consultants and professional services companies.
KPMG also appears to have missed the obvious tension in its report. On the one hand it is saying that organisations should simply go hell for leather and automate en masse now, becoming customer-serving machines at the behest of their new enterprise systems.
But on the other, it acknowledges that many organisations don’t understand AI and RPA, are unsure of their implications, and are pursuing failing, piecemeal programmes on the fringes in order to learn about the technologies and experiment – a pragmatic approach.
The implication is that they should grab a management consultant’s hand and take a massive leap into the dark, in order to emulate companies such as Amazon. That implicit sales message runs through this report.
But while companies such as Amazon and Facebook have, of course, redefined the business world, they are not without risks of their own.
For example, one reading of Facebook’s latest financial results – published last week – is that it is a highly automated organisation that is now unravelling in slow motion, taking on more and more people to fix the systemic problems caused by its own automated processes. For more on this, read our detailed analysis.
On Friday, Facebook CEO Mark Zuckerberg and other executives were sued by investors for disappointing financials that wiped $120 billion from the company’s share value.
Amazon too, could be poised to take on massive risk. While its own financials last week showed booming retail sales in the US, and a cloud business, AWS, that sits at the centre of another highly automated, diversified organisation, it is not too big to begin unravelling.
For one thing, it is becoming so diversified and automated that it is unclear who its core customers are anymore: a danger for any business that is overreaching itself. And second, it could be best placed out of the big three cloud infrastructure providers to win the Pentagon’s $10 billion cloud services contract, confirmed over the weekend.
Good news for its investors, you might think. But what would its retail customers think in a socially connected world?
In these politically charged times – in which Facebook has been brought low by Cambridge Analytica and other scandals, and Google has been forced to reverse out of the Project Maven deal, and Microsoft and Salesforce.com have been slammed for their relationships with globally reviled US immigration policies – imagine the potential damage to a popular consumer retail business from being handed $10 billion by the US Department of Defense.
This is the problem facing an automated business that no longer has a definable customer: it started as a bookseller and grew into an online retailer. Now it’s selling facial recognition systems to police forces and could run the US’ defence infrastructure in the cloud – while flogging you insurance and healthcare, making TV shows, and putting a robot in your hallway and a camera in your wardrobe.
You could see that as a measure of extraordinary success. But equally, it could be seen as a looming existential crisis on a global scale. What happens if the retail customers that still sustain its US business begin saying ‘No. We know no longer no who or what you are, or what you stand for”?
Perhaps this is the logical output of a system that puts technology, not people, first.