James Henderson

On AI, ‘you have to show the magic’

“Do you know Arthur C. Clarke?” asked Nick Eshkenazi, triggering a wave of intense note-taking and Googling.

“He’s one of my heroes and if you haven’t read any of his work, then I encourage you to do so. He coined three laws and after publishing the third he stopped because Isaac Newton had three and who can do better than Newton?”

Clarke was a British science fiction writer who co-wrote the screenplay for 2001: A Space Odyssey. His third law is best known and most widely cited – any sufficiently advanced technology is indistinguishable from magic.

What started as an observation on the nature of technology and discovery now resonates most profoundly in the unfolding field of artificial intelligence (AI).

“You have to show the magic,” advised Eshkenazi, speaking as Chief Digital and Transformation Officer of Astellas Pharma.

“We have an accountability to go back to this law because the technology doesn’t matter if you cannot show the magic. If you cannot show the magic of your solution or your project then you’re wasting your time.”

Nick Eshkenazi (Astellas Pharma) speaks on AI: Enthusiasm or Execution? at Moxie Authority 2025 (Photo: Christian Gilles)

With over 30 years of experience in technology, data and operations across global organisations – such as Woolworths Group, Costco Wholesale, PACCAR and USAA – Eshkenazi has a reputation for pragmatic innovation, exponential thinking and strategic leadership.

“There are magical things happening as we speak, all around us,” Eshkenazi expanded.

“We have a responsibility within our company to identify the problem and turn that problem into a magical solution. But it has to make people look at it and say, ‘wow’, because only when you hear that ‘wow’ do you know that the magic has arrived.

“We all crave it. We all look for it. But it’s so rare and that’s why I wake up every day looking for that. And when I find it, I go deep and then I go deeper and I try to deliver it as fast as I can.”

At Astellas Pharma, Eshkenazi leads the DigitalX and Transformation Office functions, focusing on growing the company’s digital and transformation capabilities while executing strategic agenda ambitions.

“We all have to try and make a difference every day for the mission that we have subscribed to,” he added.

“In our case, that is turning innovative science into value for patients. That’s what we exist for – changing people’s lives and creating life extension solutions. Very powerful. It wakes me up in the morning and keeps me up at night.”

Impassioned and impactful, Eshkenazi held the attention of 400 executives during the opening keynote of Moxie Authority 2025AI: Enthusiasm or Execution?

His innate curiosity and commitment to lifelong learning inspired a new way of thinking at this inaugural conference in Sydney, housing the most influential minds in business and technology across Australia.

Show me the AI magic

If reaching a state of alakazam with AI is the end game, then Eshkenazi directed focus to the revolutionary world of voice – “we can show the magic with speech”.

Namely through Sesame, an interdisciplinary product and research team focused on making voice companions useful for daily life.

Founded in 2023, the start-up is based in San Francisco and specialises in building voice companions capable of holding natural conversations and devices that make them more effective.

According to the company’s website, Maya and Miles are Sesame AI companions with natural technology so “expressive and conversational that you’ll forget you’re talking to an AI system”.

“By the way, I don’t own any stock in Sesame nor am I the CMO but in my personal opinion, it’s fantastic technology,” Eshkenazi shared.

Moxie Research shared at Moxie Authority 2025

“Go check it out and play with their two personas. This is the first voice AI that I have wanted to come back to and talk to again – they have maintained context very nicely by remembering previous conversations.”

The goal of Sesame is to achieve “voice presence” – widely considered as the magical quality that makes spoken interactions feel “real, understood and valued”.

The key components of that approach are:

  • Emotional Intelligence: Reading and responding to emotional contexts
  • Conversational Dynamics: Natural timing, pauses, interruptions and emphasis
  • Contextual Awareness: Adjusting tone and style to match the situation
  • Consistent Personality: Maintaining a coherent, reliable and appropriate presence

During his time at Woolworths, Eshkenazi was part of the core team tasked with building Olive, the conversational AI platform which was rolled out in 2019.

“We were using powerful technology at the time with Dialogflow and TensorFlow from Google but it was still a struggle,” he recalled.

“The process was painful and difficult at times but that’s no longer the case – we were never able to deliver the experiences delivered today by something like Maya and Miles.

“Play with Maya and Miles and you’ll see the magic.”

A state of AI enthusiasm or execution?

Aligned to the theme of Inspired Knowledge, Moxie Authority 2025 housed the most influential figures setting the market agenda in business and technology across Australia.

This inaugural and invite-only conference in Sydney hosted more than 400 industry front-runners spanning all ends of the ecosystem, from CIOs, CTOs and CISOs to CDOs, CEOs and Founders.

As part of the pre-registration process, all attendees were asked to submit the one question they most wanted to be answered during the event. Out of 400 questions, more than 140 questions centred on AI.

Such strong market intrigue was mirrored by new findings from Moxie Research – 87% of Australian businesses will allocate additional budget to AI projects in 2025. In addition, 78% will build a cross-functional team to set AI strategy while 69% will define clear AI business objectives.

“Can you have execution without enthusiasm?” Eshkenazi asked. “But you probably can have enthusiasm without execution.”

Labelling AI as simply an inflection point is not enough however. This is an inclusive technology capable of changing conversations across all industries and geographies.

Nick Eshkenazi (Astellas Pharma) speaks on AI: Enthusiasm or Execution? at Moxie Authority 2025 (Photo: Christian Gilles)

“What’s different is that it’s moving so fast,” Eshkenazi observed. “This could be the first time that by the time you complete a proof of concept [POC] – which probably took a sprint or two lasting 2-4 weeks – you’re already behind.

“It’s already too late because upon completion, the technology has changed under your feet. That’s how fast it’s moving.”

For example, as organisations assemble AI solutions in sandbox environments, new large language models (LLMs) are being released in parallel and at pace. Or a new feature under construction is already being released.

“That’s the speed,” Eshkenazi expanded.

Adding a different dimension to a timeworn topic is no easy task. Every press release, analyst report and earnings call is littered with mentions and intentions related to AI.

Consequently, Australian businesses are struggling to differentiate between market hype and market reality.

“We’re at a very early stage of acceleration but more will come because the competition is just starting now,” Eshkenazi shared. “The news on AI is now hourly, not even daily anymore.”

According to Moxie Research, senior leadership teams are “championing AI initiatives and actively driving AI strategy” at 72% of Australian organisations surveyed.

Support for AI adoption is shared company-wide with each department driving positive change and new initiatives, spanning both technology and business:

  • IT Department: 58%
  • Senior Management: 46%
  • CEO: 31%
  • Business Operations: 27%
  • Customer Service: 27%
  • Product Development: 25%
  • Sales and Marketing: 21%

As outlined by Eshkenazi in his own words, three AI archetypes have emerged:

  1. Buy and Deploy: “Use an off-the-shelf solution to solve the problem in more of a lift-and-shift scenario. Replace a process, deploy an AI agent and move on. This is absolutely available now.”
  2. Optimise: “Take an existing business process and start to challenge it. Explore if there is another way and understand if an existing new phase can be entered.”
  3. Reinvent: “This is the most exciting – entirely reinvent a business model. Mostly this is a case of, ‘hey, we have a business problem so let’s go and fix it’. But partnerships must be in place internally for this to succeed, especially in the digital space.”

In this context, Eshkenazi explained that archetypes can be blended and definitions can vary based on the business brief. But remember, the market remains immature and organisations continue to learn as they go.

“You can buy a platform and build on top of it,” Eshkenazi shared. “But I would encourage staying away from customisations because when things are moving fast you become stuck.”

Based on Moxie Research, 84% of Australian businesses are currently running internal AI working groups to shape strategy and drive adoption.

Moxie Research shared at Moxie Authority 2025

While progress varies based on maturity, the key benefits of cross-collaboration between business units is to primarily drive efficiency and process optimisation (59%).

This is alongside creating innovation and a competitive edge (51%), plus enhancing customer experience (50%) and kick-starting change management and integration processes (49%).

“Are you missing out? I would say, no, I don’t think you are,” Eshkenazi clarified. “And I’ll tell you why – because things are moving so fast, it’s good to watch and learn.

“Is there such a thing as first mover advantage in AI? Let’s see. Some companies are rehiring for roles that were previously discontinued because the changes they made have not worked as expected.”

In other words, understand the pearls and pitfalls of AI in equal entirety.

Define critical capabilities, build teams and go deep

Even though the majority of organisations are favouring a “watch and wait” position on AI, this strategic pause has not stopped technology leaders from taking a lead role in laying the technical groundwork in the background.

Should a POC need to be spun-up at speed and potentially at scale, IT environments must be primed and ready to deploy.

Based on Moxie Research, that translates into a commitment from CIOs and CTOs during the next 6-12 months to:

  • Select the right technology stack / infrastructure: 68%
  • Evaluate data readiness to maximise AI: 65%
  • Strengthen governance and ethics framework: 57%

Within this context, the key pillars underpinning AI strategies in Australia are:

  • Data and Infrastructure: 70%
  • Strategic Business Alignment: 57%
  • Operational Integration: 55%
  • Governance and Ethical AI: 53%
  • Performance Measurement / Optimisation: 49%

Despite the direction of travel, more than half of businesses across the country (52%) currently lack “modern, scalable data infrastructure optimised for AI”. Limitations are commonplace and capabilities in some instances can be described as “basic” at best.

“Articulate the problem well, start with that,” Eshkenazi outlined. “Most organisations have a tendency to favour a tool without having any clarity on what that tool is supposed to fix.”

Instead, ask those clichéd yet critical questions. What’s the problem we’re trying to solve? What does success look like? What will be different upon completion?

When trying to create something that is yet to exist, Eshkenazi advocated the need for businesses to take a deep dive on AI. This is beyond a simple licence for Microsoft Copilot – more exists to explore below the surface.

Nick Eshkenazi (Astellas Pharma) speaks on AI: Enthusiasm or Execution? at Moxie Authority 2025 (Photo: Christian Gilles)

“You can tell by my accent that English is not my first language, it’s actually my third,” he stated.

“So yes, emails have significantly improved across global companies working in different companies through AI writing tools. But that’s only 1% of what it can do.”

Play with the tools absolutely but demonstrate progression by identifying a “small and very effective” set of use cases. Then go deep, define the success criteria in advance and garner support from leadership.

“Because once you meet that success criteria through a POC or pilot, you have to go live otherwise you’re wasting your time,” Eshkenazi claimed.

According to Moxie Research – and by order of severity – the top AI project deployment challenges currently facing organisations across Australia are:

  1. Data Management and Transfer
  2. Cost Management
  3. Limited Availability of GPU Resources
  4. Infrastructure Integration
  5. Scalability Issues
  6. Security Concerns
  7. Performance Issues
  8. Cooling and Power Consumption
  9. Technical Expertise
  10. Software Compatibility

“There is a saying that everyone is an expert with hindsight,” Eshkenazi acknowledged.

“The challenge with AI is that we’re all trying to be an expert with foresight and that’s impossible. It’s absolutely impossible so let’s not try. But we can try the tools and we can try to solve some problems because when we succeed, success breeds success.”

At the crux of it, Eshkenazi cautioned against companies creating many moving AI targets. This isn’t a throw spaghetti at the wall and see what sticks type of scenario.

“You create mediocracy,” he warned. “But if you’re very purposeful and focus on one, two or even three well-articulated problems, then deploy the right teams to solve them you will have a better opportunity to achieve success.”

As a leader in business and technology, Eshkenazi is well-versed is bridging the gap between both domains through a pragmatic and common-sense approach to transformation, innovation and execution.

“I strongly believe that an agile mindset combined with effective ways of working and operating models are significant enablers when deploying something new,” he said. “That’s already been proven in AI projects.”

Translated into tangible takeaways, Eshkenazi recommended the creation of “small, cross-functional teams” consisting of five to seven people – a maximum of 12 – with business knowledge of the problem that needs to be solved.

Pair this side-by-side with engineers, analysts and product managers – “but only the right people with the right skills” – and watch as they collectively deploy, evaluate and reassess until the success criteria is met.

“And they run fast,” Eshkenazi added. “I mean that literally in the sense that they define the problem on Monday and have a functional prototype ready on Friday.”

Moxie Research shared at Moxie Authority 2025

Achieving this level of speed can only be achieved through an insourced approach to AI deployments, Eshkenazi explained.

“If there is one learning that I draw on over the years is that you have to internally define your critical capabilities as an organisation,” he said. “These are the areas that you want to invest in and nurture and for everything else, then you can use partners.”

By applying this philosophy to Astellas Pharma, Eshkenazi is “hiring all the time” with a global outlook.

Any place where the right skills exist, then leveraging an organisation with a worldwide footprint will come into play. Through the “privilege” of hybrid ways of working and a follow-the-sun mentality, this is a company that works 24/7 around the clock with people located in all corners of the planet.

“If it’s a strategic capability then I insource it,” Eshkenazi continued. “If it’s strategic advice then I partner to bring in an external thought process to kick-start something new.

“But what if I need to own it for the future? Then I will own it, buy it, build it, end-to-end across the entire lifecycle through a small team of specialised experts.”

According to Moxie Research, the Australian market is divided on the insource vs. outsource question specific to AI. Based on the data, organisations seeking to maximise the benefits of AI will:

  • Increase partnerships with specialist third-party IT vendors / partners: 46%
  • Build internal skills and increase in-house capabilities: 45%
  • Acquire companies with proven AI expertise: 9%

“Over time, some organisations become heavily dependent on third-parties and then everything has to be done through a third-party,” Eshkenazi qualified. “In those instances, a project doesn’t start unless a third-party is involved.

“In the space of AI, partners are learning at the same speed as we are because the technology is so new. So you need to decide, do you pay the partners for them to learn with your money? Or do you pay your own team to learn with your money? You have to decide and make that call.”

Moxie Authority 2025 housed the most influential figures redefining business and technology across Australia. This inaugural and invite-only conference in Sydney hosted more than 400 industry front-runners spanning all ends of the ecosystem, from CEOs, CIOs and CTOs to CDOs, CISOs and Founders.

SIGN UP FOR INSIGHTS VIA MOXIE MAIL

Inform your opinion with executive guidance, in-depth analysis and business commentary.