Generative AI In Business: Why Accenture Is Investing $3 Billion In AI

International technology consulting group Accenture has certainly been bitten hard by the generative AI bug. In coming years, it plans to double its headcount of data and AI experts from 40,000 to 80,000 as it invests $3 billion into its capabilities.

All of this is in order to improve its ability to leverage AI in three ways: To do better work for clients, to improve its own operations, and to transform the way it works for clients.

Recently, I caught up once again with Accenture’s Paul Daugherty. Five years after the publication of his book Human + Machine: Reimagining Work in the Age of AI, I took the chance to ask him about the developments we’ve seen since its publication. We also talked about how the arrival of generative tools like ChatGPT will affect the ongoing AI-driven transformation of enterprises.

Generative AI – the game changer

“Over the course of my career,” says Daugherty, “I’ve been really amazed by technology four times.”

The first was the Lisa – the early Apple computer that popularized the idea of a personal computer with a graphical user interface. The second was browsing the internet for the first time. The third was the arrival of the smartphone.

“The fourth was about 18 months ago when I started to see the progress in …specifically large language models (LLMs).

“The pace of advancement was becoming breathtaking, and some of the capabilities … to understand language and the generative capabilities to create content. Not just diagnose and evaluate or predict …really is game-changing in terms of the impact on business."

Amazing though it undoubtedly can be, Paul still sees businesses struggling to work out how and where to apply it. Often it isn’t that they have problems identifying use cases – there are plenty of them. Rather, stumbling blocks are around understanding which ones are likely to help them achieve business goals and where investment should be prioritized.

This is backed up by recent stats collated by Accenture that shows that 98 percent of executives believe generative AI will be essential to their strategy going forward.

“So that shows you the level of interest and the impact that executives see this having on their business," says Daugherty.

How Accenture is Using AI

Potential applications fit into one of the three categories mentioned at the start of the article. Further to that, they are broken down into five types of applications: creating, automating, advising, protecting and coding.

Accenture’s digital media business, Song, is a flagship for its use of creative, generative AI. Song is involved with innovating in content creation by understanding the plethora of new tools that are available. Client use cases it is involved with range from marketing and outreach to drafting regulatory documents for life sciences companies.

It has also assisted banking customers with automating back-office processes using GPT-3.5 and implemented its advisory capabilities in customer services operations.

The “protect” implementations are showcased by a worker safety solution created for energy companies. This involves pulling together real-time information on what is happening in a facility in order to improve safety outcomes.

And coding refers to using the super-human ability of LLMs to generate computer code, which enables software developers to become more efficient and productive.

Altogether, Daugherty tells me, Accenture is involved with helping over 100 of its clients implement generative AI projects.

Investing in Capabilities

Not all of the $3 billion in investment that Accenture has just announced it will make in its AI and data capabilities will go on generative AI.

“Generative AI is clearly game-changing for all the reasons I've talked about – but data and AI broadly have tremendous opportunities for companies," says Daugherty.

“It’s not even just about AI – it’s about data and how you pull together the data foundations, the modern data platforms and the cloud-based platform to drive the opportunity.”

As well as doubling the size of the data and AI team, Accenture has developed a tool called AI Navigator, designed to help its clients find their way through the baffling and ever-expanding range of enterprise AI opportunities open to them.

What models does a business need to work with, and how can they be used? Is it best to just use an existing API? Does data need fine-tuning or customizing? How can prompt engineering be used? "That's what the AI navigator for Enterprise will help our clients with," says Daugherty.

Responsible AI

We understand that algorithmic bias can create societal inequalities and decision-making done by opaque black box machinery, incomprehensible to most humans, has the potential to cause harm.

Then there’s the potential for a breach of privacy related to computer vision applications such as facial recognition.

These issues (and many more) have made us aware of the importance of pursuing "responsible AI."

Daugherty tells me that – as far as he is aware, at least - Accenture was the first organization to establish a formal compliance program around responsible AI.

“It’s even more critical with generative AI … we’re working with a lot of stakeholders, regulators and governments, as well as our clients, to establish … responsible AI capabilities.”

This “responsible AI framework” ensures that questions can be asked and answered around matters of bias, fairness, privacy and security relating to any AI projects.

On top of that is a “diagnostic tool” used to evaluate the risk level of any AI-related work.

Daugherty says, "There are certain types of AI categorized as high risk, versus … different levels of risk … and do you know, in your organization, where are you using what might be considered high-risk AI?

“I would submit most organizations really wouldn’t have an idea right now.”

Part of the answer lies in having a solid and reliable inventory of every application within an organization that involves AI. This means that each one can be individually risk assessed and monitored for outcomes as stakeholders build their own understanding of what has to be done to ensure AI is used responsibly.

The Future?

Where do Daugherty and Accenture see this all leading? And what other technologies will they be putting to work alongside AI to accelerate this wave of digital transformation?

Accenture’s 2023 technology vision document – titled When Atoms Meet Bits – highlights digital identity, trust and transparency, next-generation AI and what it refers to as “science technology” as the biggest trends to watch.

And while some believe the concept has gone off the boil, Daugherty is happy to report that he believes “The metaverse is alive and well.”

“The idea of making the two-dimensional digital world three-dimensional is real. There's a lot of activity happening around the metaverse; witness Apple’s new headset launch.”

As well as generative AI, next-generation AI covers a range of frameworks and concepts, including "common sense AI" and explainable AI.

Then there are next-generation computing architectures, including quantum computing and biocomputing. The first of which is already being implemented in Accenture client deliveries.

And after that? Only the final frontier: “There are lots of interesting advances happening in space technology, both what happens in space … satellite communications … experiments in space. And also how you can use information from space [like] low-Earth orbit satellites for transformative applications back on Earth.”

You can click here to watch my full interview with Paul Daugherty, Group Chief Executive, Technology and CTO for Accenture.