Five questions to drive AI success

Erik Bakstad at Ardoq proposes five questions that organisations must answer to improve the success of AI adoption projects

With an estimated $500bn set to be spent on AI in the next three years, it’s clear we are in the middle of a digital gold rush. AI has the potential to transform entire industries and those organisations that invest wisely stand to reap phenomenal rewards.

However, given the absence of a playbook to follow when adopting such new capabilities, it can be difficult for organisations to forecast the impact, and balance the risks against the rewards. It’s unsurprising that a recent report revealed barely half (53%) of emerging technology adoption projects deliver measurable business benefits or impacts.

While AI will create some big winners, failed adoption projects will almost certainly result in billions of wasted dollars.

When it comes to emerging technologies – especially ones as powerful and fast evolving as AI – it can be difficult to forecast the benefit and avoid sunken costs of investments that are quickly obsolete. IT leaders face an equal risk from inactivity, as they could find themselves with a significant capability gap. They therefore have little choice about taking risks if they want to remain at the forefront of their market.

An overwhelming majority of CIOs (79%) say if they don’t take risks on emerging technologies, they will go the way of the dinosaurs, while three-fifths (61%) feel compelled to invest due to fear of missing out. As such, businesses are faced with a tricky dilemma. To avoid being left behind by competitors, they must take a leap of faith into uncharted waters, knowing that they may not see a clear return on investment.

The five golden questions

To improve the ratio of success with emerging technology adoption, experimentation should always be tempered with pragmatism. Already we’re seeing how the rush to invest in AI is leading many to throw caution to the wind, while ignoring critical risks such as potential regulatory issues, ballooning costs, or outright project failure.

It’s essential that IT leaders have a modern enterprise architecture (EA) function that minimizes these risks and prevents wasted investments, by enabling them to analyze, design, plan, and implement a successful digital strategy. This will provide the insights to answer five key questions before launching into any research and innovation project: 

1. What proprietary data does the organisation have that is high quality and will remain relevant in the future?
2. What guardrails need to be in place to protect that proprietary data when experimenting with AI?
3. What principles and guidelines should be established to enable the organisation to make conscious decisions about what to build and what to buy?
4. Are there any emerging architectural patterns that could enable AI adoption, and does the organisation need to change its approach to leverage them?
5. How can the organisation architect AI solutions in a way that gives it the flexibility to change foundation models when newer and better ones emerge? 

Answering these questions will help IT leaders to identify potential gaps in the organisational structure that could lead to problems as they continue to invest in and implement emerging technologies. For example, a good place to start with AI is to build capabilities that leverage the organisation’s proprietary data, but IT leaders shouldn’t waste time developing solutions that already exist elsewhere.

They should therefore track open-source AI projects alongside emerging technologies, and focus on building data pipes that enable them to be integrated into existing workflows more readily. From an architectural standpoint, organisations should consider where there is an opportunity to embrace a composable mindset, by plugging into third-party capabilities through APIs rather than building all functionality internally.  

Making adoption projects easier

Unfortunately, most organisations often find that it is difficult to answer the five key questions effectively. A third (33%) of CIOs admit that their EA functions are outdated and riddled with time-consuming processes.

Organisations therefore often fall back on slow and error-prone impact analyses that fail to provide a detailed understanding of the risks, benefits, processes, and resources involved. This results in projects being delayed, IT teams left blind to the impact of the changes they are trying to make, and vast amounts of resources being wasted. 

To address these issues, organisations need to shift their mindset away from seeing EA as an IT function to conceive of it as a critical business discipline that drives alignment between people, processes, information, technology, and their overall goals.

EA has evolved far beyond the traditional once-per-year exercise conducted from an ivory tower to document the organisation’s systems and processes for an audit. It is now about providing continuous transparency and enabling all teams – both within and outside the IT department – to make more informed, strategic decisions.

IT leaders therefore need to create an EA that removes the reliance on manual processes and is driven by high quality, up-to-date data.

Evolve or die

By their very nature, emerging technologies are unpredictable and risky. Across all industries, the way in which AI is being adopted is once again demonstrating that there is no single approach that can ensure a seamless implementation process.

To have the best chance of success, organisations need to remain in control, by maintaining the insights needed to understand and mitigate risks, and enable a smooth transition throughout the research and adoption process.   

To achieve this, organisations should establish a modern EA function that spans the entire enterprise, so they can automatically gather the information that provides transparency and enables smarter, faster decisions. This will create a solid digital foundation that acts as both a springboard and a safety net for rapid adoption of new technologies.

We have seen time and again that organisations that are swift to adapt to new technological advancements and fuse them into the fabric of their business are the ones who conquer markets and dominate the cultural conversation. For all the ways in which AI is a game-changer, that is one aspect that will surely remain constant.

Erik Bakstad is CEO and co-founder of Ardoq

Main image courtesy of iStockPhoto.com and anyaberkut