Seven keys to building enterprise automation, innovation and AI capabilities that deliver strategic impact

Over the last 10 years, I’ve helped many organisations build automation and innovation capabilities that harness the ever-increasing power of AI to deliver tangible value to customers, employees and the bottom line of their organisation.?One thing has struck me though: for every organisation that succeeds in this endeavour, I see three organisations fail to scale. Whether they lose interest or focus, build excessive ‘shelf ware’ or just don’t deliver any tangible benefits at all.?These relative failures mean years of effort and investment are wasted, and organisations are unable to rapidly pivot to the emergent possibilities of generative AI (GenAI), which is, for many, a challenging market for investment.

In this article, I've summarised seven important lessons from my experience with effective innovation, automation, and artificial intelligence (AI) teams in various sectors, including financial services, the public sector, retail and higher education. These lessons are crucial for organisations trying to effectively use GenAI and move from initial experiments to fully implemented solutions that provide strategic benefits.

Lesson 1: Have confidence – the business case for a single use case isn’t the be-all and end all!

?I’ve seen many organisations over-analyse business cases for AI models or automation use cases in the early stages of setting up their enterprise capability.?Whilst fiscal fortitude is admirable, creating economic hurdles for innovation in the early stages can be counterproductive and stifle the very thing that you are trying to achieve.?

?The companies that quickly expanded their automation and AI used a strategic approach. For example, a financial services institution tapped into the CEO's contingency fund to drive GenAI innovation and find the best uses for it. Similarly, a major Australian university boosted its AI and automation efforts when the CFO allocated a significant budget for developing new use cases, proof of concepts and pilots across five functional teams.

?

Lesson 2: Be business-led, but technology and data-enabled.

The term ‘business led’ can be cliché, but in my view, it is critical. Technology and automation teams exist to help the broader business solve business problems and provide capabilities to drive competitive advantage.?The successful automation and AI factories we’ve supported really analyse the business challenge and gauge whether technology has a role to play in the solution, before considering which individual technologies may be appropriate.?

?With a leading Australian public sector organisation, we've successfully combined their process excellence capability with their automation and digital solutions team. This approach ensures that automation, workflow and AI are used only on processes that truly benefit from them, rather than on a wide range of inefficient ones.

?

Lesson 3: Incentivise experimentation and minimise the reasons not to adopt.

From my experience, successful AI and automation initiatives often make it easy for business users to engage and adopt their services. Some organisations do this by centrally funding the early stages, allowing different departments to benefit from AI and automation improvements without incurring initial costs. This is crucial because creating secure enterprise platforms from the start can be expensive and complex.

?Two of our clients have mostly financed their automation, innovation, and AI projects with central funds – one covers all expenses through the CFO, while the other covers the majority of costs centrally, but requires contributions from each department for key projects to ensure commitment. In addition all successful organisations I've worked with have invested considerable effort in teaching their employees about the art of the possible and encouraged them to share creative ideas.

?

Lesson 4: Perseverance is a virtue – when applied wisely.

Being persistent is important, but it can also lead to problems. Teams that excel in automation and AI know how to find the right mix between trying out different applications and knowing when to quit on projects that are unlikely to succeed, lack business support, or would use too many resources. Essentially, these teams regularly reassess their priorities and are not scared to 'fail' and put an end to unsuccessful projects.

At the same time, some organisations succeed with new technologies after multiple attempts. A major client of ours first tested process mining software but didn't see great results and shelved the idea. Later, they revisited it with different processes, a new partner, and software, which led to much better outcomes creating the ability to optimise their entire value chain using their extensive data.

?

Lesson 5: Leverage partners and your enterprise ecosystem.

At the risk of coming across as self-serving – my clients’ experience has shown that organisations which leverage their trusted partners in the data, AI and technology space have a better chance of success.?With AI and associated fields moving forward rapidly, insights from ecosystem partners are invaluable to upskill and hone your automation, analytics, data and AI strategies.?As one CFO said to their peers recently “you have to distil through the sales spiel from partners to find the gold – but it’s worth doing!”.?

Partners play a key role in offering validation, sharing experiences, providing specialised skills that may be hard for your organisation to maintain, and giving insights into worldwide trends. It's also crucial to exchange knowledge and learn from peer organisations, ideally on a global scale, to see how you measure up.

?

Lesson 6: Focus on risk and security by design (and do it early!)

A common trait I’ve noticed amongst clients who’ve successfully built automation and AI capabilities that have scaled to have enterprise impact, is that they’ve created guardrails early on that enable adoption, experimentation in relative safety. For one of our clients, this meant creating the risk framework for automation and AI early – aligning all activities to existing risk and governance policies and adjusting these when it made sense.?

?We've observed that some organisations require the use of a GenAI platform, which is safeguarded within their broader cloud data platform, complete with stringent security measures and usage protocols. Another client uses GenAI to better handle hundreds of thousands of customer interactions annually, reaping substantial rewards while intentionally involving a human reviewer to check and finalise all GenAI-generated initial responses before they reach the customer.

?

Lesson 7: Sculpt your AI or automation hub to suit your operating model. ?

Over the last few years, many organisations have asked “where should I place our AI capability?” or “where should I build my automation hub?” They wonder whether it should it be centralised, hub-and-spoke, fully federated etc.?For me, it’s horses for courses, but I believe that a centralised component is critical to get scale, velocity and make sure that people with some of the key skills are working with ‘likeminded folk’ to learn from each other and constantly innovate.?

In all the successful, large-scale operations, a central team has been essential. It brings the 'factory' model to life, enabling cost savings, pattern reuse, and the development of learning and career opportunities. An AI leader highlighted the benefit of having data experts collaborate directly with peers, rather than being isolated in support roles or individual business units. Additionally, I've noticed that effective innovation, automation, and AI centres often foster unique subcultures where teams excel at overcoming business challenges, quickly solving problems, creating value, and nurturing talent within the organisation.

?

I hope the above helps you on your journey to scale your innovation, AI and automation capability.?To find out more about the EY.ai holistic approach to AI, visit ey.ai/au or reach out to me directly if you’d like to discuss the above.

The views reflected in this article are the views of the author and do not necessarily reflect the views of the global EY organisation or its member firms.