What separates two organisations that have embarked on a journey to transform itself with the power of AI? This has been a common question in business circles. According to our Data Pulse survey, 89% business leaders are planning to adopt more AI solutions in the next 12 months, and this begs the question: how can they do it the right way?

As our experts have pointed out on this series, there are several pitfalls when working with AI. Working with AI requires a long hard look at whether the company’s IT infrastructure and data storage are able to manage the increased workload that comes with the new technology. It also opens the organisation up to new levels of security risks, and our experts discuss how to be ready for, and circumvent them. Finally, working with a complex technology requires special talent to navigate its complexities.

In our final post about the four pillars of AI success, our experts delve deeper into the investment and strategies crucial for an organisation to have in place before venturing into the world of AI. Implementation isn’t just a costly venture – as a rapidly-evolving technology, it is also a complex one. Here are some top tips from our panel of esteemed AI experts:

  1. My organisation has just started using AI. What must we do to implement AI successfully?

BS Teh, Seagate

In my recent experiences, data is an important foundation of a successful AI journey, but it is also the most overlooked part of it. Data is the lifeblood that powers AI applications. The algorithms it produces to derive better and more meaningful insights all run on data – copious amounts of it. The growing use of AI in an organisation will mean an exponential increase in data volumes, and by extension, will require a robust data infrastructure that is future-proof.

The problem? While 74% of our respondents in APAC have already implemented AI in one or more areas of their business, 20% believe they are still not ready to handle the increasing stream of data. A majority of them believe the solution lies in further investments in IT infrastructure and a state-of-the-art data storage solution.

Finally, success in AI requires business leaders to empower lines of business with a clear strategy and direction. As the adoption of technology moves at such a rapid pace, various areas of the business are yearning to understand the strategies behind technological shifts within the company, and their role in working with these technologies to power the business forward.

Akash Bhatia, Infinite Analytics

One of the most important areas in successful AI implementation is the underlying data infrastructure. AI thrives on good data. Even if the organisation runs on a Relational Database Management System – a structured language to perform tasks such as update or retrieve data – or No-SQL – which is unstructured – these systems and motivations for using them need to be reviewed regularly.

The frequency of data refresh – a process where data is added/removed from the dataset– also plays a considerable role in determining the infrastructure. If your application/product needs real-time responses, the infrastructure requirements will be quite different.

Johnny Chou, Viscovery

AI technologies are based on the premise of deep learning – they require large amount of data that are structured into algorithms and training models to continuously optimise the process.

To optimise the progress, it’s crucial to collect the real data in order to train the machine learning models.

Moreover, after the deployment of AI, organisations will need to collect, sort, store and even migrate more volumes of data. Therefore, they should plan for the expansion of their IT infrastructures, and invest in training more AI talents to manage the AI system effectively in order to achieve successful results.

Rohit Pandharkar, Mahindra Group

There should be clear communications from the top about the most pressing business challenges, and how they can be solved by machine learning.

Furthermore, organisations should start with the business problems where machine learning can have the biggest impact in the shortest amount of time to start showcasing the value that can be derived from this technology.

  1. What forms an ideal AI strategy? What should my organisation be considering/evaluating?

Rohit Pandharkar, Mahindra Group

Start by identifying the top three business challenge that can be solved using data. You need to figure out how these problems can be resolved mathematically, so it requires a deep understanding of how algorithms work. Finally, positive impact needs to be measured in terms of profitability, revenue, or increased savings.

  1. What are the top 3 challenges/barriers facing organisations looking to implement AI?

Rohit Pandharkar, Mahindra Group

As AI is a relatively new space, the number one challenge right now lies in talent acquisition. Every organisation looking to transform by implementing AI is asking the same question, “Where can I find ‘rock star’ data scientists to help me level up my business?”

The second challenge is really getting buy-in from the organisation’s leadership. This requires a shift in mindset starting from the boardroom level – they need to recognise the potential impact that predictive analytics can actually make.

Lastly, organisations need to clearly justify that the benefits of AI implementation are actually worth the effort and investment that has been put in.

  1. What are the first tasks in the organisation that I can automate with AI, in order to free up more resources, which can then be re-invested?

Rohit Pandharkar, Mahindra Group

The first areas to automate with AI is to identify customer segments and reach out with more personalised product offerings. Another area is to make business projections using science and dashboard automation to ensure planning can be done quicker – for example, an automated task list for sales people.

  1. How do I convince the boardroom of the value AI brings to the organisation (i.e. push AI implementation higher up the boardroom’s agenda), and secure the investment required for its successful implementation?

BS Teh, Seagate

Convincing the boardroom to take the leap and make large investments is always a challenge. The first question to consider is about the level of digital maturity in an organisation – has its IT moved from a cost-center to being a core driver of the business? This can help determine what the next steps are – for example, if the board has already seen the benefits from investing in digital transformation, and whether the company is ready to take the next leap and make a bold investment in automation.

From an innovation perspective, it’s important that leaders know the potential of AI in enhancing forecasting, based on the data an organisation has generated and stored over the years. For any business at all, being a few steps ahead always translates well for the bottom line.

The board also needs to understand that AI is so ubiquitous that its ability to maximise the value of data and transform it into strategic decisions and actions can benefit all areas of the organisation. We haven’t even begun to scratch the surface of what areas AI can help – cost savings and greater efficiency are often mentioned, but its use-cases continue to grow by the day.

Andrew Burgess, APD

The best way to convince the boardroom is to invest a small amount in a pilot, or a Proof of Concept that clearly demonstrates the value proposition of an AI investment. No board will ever decide against making an investment that adds to the bottom line, so this value proposition needs to reflect growth.

Johnny Chou, Viscovery

AI is a long-term and cumulative investment for all organisations. It takes time no matter on IT infrastructure build up, talent cultivation and enterprises implementation.

That said, the value of innovation and potential business model changes are hard to ignore for companies that look for a greater competitive advantage.

Therefore, companies must start as early as possible to formulate their strategy around AI investment, and commit to a long-term approach to continuously adjust and improve their AI roadmap in order to create robust value and stand out among their competitors.

Rohit Pandharkar, Mahindra Group

One can start by demonstrating successful examples and results of the algorithms deployed. These outcomes have to ladder up to profitability, revenue, or increased savings. Another good way to urge the board to invest in AI is to showcase examples where peers or competitors have been able to successfully implement the technology.


Leave A Comment