First steps towards success

What is the AI adoption success formula?

Yaroslav Kuflinski
© Shutterstock / ronstik

We are in the middle of yet another wave of digital transformation with AI at its core. Business leaders’ skills are put to the test here, as transformation is never an easy task. Let’s look at the common hurdles of AI implementation.

Artificial intelligence is no longer a mere tech buzzword but a driving force for many businesses. There are little to no doubts about AI’s unprecedented potential to provide value across verticals. Conversational systems, process automation tools, recognition and personalization systems, and predictive analytics are all the nascent use cases of AI actively probed right now.

Although AI is still considered to be a relatively immature technology, the business world is beginning to massively invest in custom software powered by AI. The recent study by Cognilytica called ‘Global AI Adoption Trends & Forecasts 2020’ indicates that 91% of the surveyed companies are planning to adopt AI in the next five years. According to the Global 2019 AI Survey by McKinsey, 63% of respondents report revenue increases due to AI adoption, while 44% point out cost savings driven by AI.

SEE ALSO: Dealing with past developer decisions – keeping up or staying still

The overall vector is clear – AI’s global mass adoption has already begun. Let’s look at the common hurdles of AI implementation and the steps that companies should take to mitigate the risks and successfully integrate AI in their products or operations.

AI implementation challenges and pitfalls

As with any other emerging technology, roadblocks on the way to its adoption are inevitable. Novelty often creates uncertainty, and hype around the technology may lead to impulsive, poorly strategized decisions. Based on the stories of other companies that had the bravery to implement AI in its early stages, we have the luxury to study their experience and avoid their mistakes. Although every organization has its own very specific characteristics that impact the implementation strategy, there are certain recurring issues that have been observed across many companies in a wide range of industries.

What’s the matter?

Somewhat surprising, if you ask some of the adopters about their reasons to implement AI, their answer may easily fall into the ‘just because’ category. I mean, how can an organization be on the frontline of technological advancement without adopting the most talked-about tech, right?
It’s hard to ignore the hype, we get it. Companies keep falling in the same trap of coming up with a legit reason for AI adoption after making the initial decision to adopt it anyway. Regardless of what change you want to introduce to your enterprise, it all starts with the estimates of added value and ROI.

Data quantity and quality

Data, the gold of the 21st century, is a fundamental building block of any AI initiative. Again according to Cognilytica, an astounding 39% of companies report data quantity as their main challenge to AI adoption. While there are seemingly infinite amounts of data, its accessibility is very limited. For example, when it comes to healthcare, sensitive patient data is extremely hard to obtain due to regulatory issues.

Moreover, even with non-confidential information, AI algorithms are very picky about the quality of data they run on. More often than not, companies are left perplexed over having lavish amounts of data but failing to integrate it for AI to digest.

Lack of talent

According to Cognilytica, 27% of companies report that limited accessibility to AI talent is the main reason for the technology’s slow adoption on their side. Undoubtedly, in the context of emerging technologies, talent is one of the key ingredients of a successful implementation strategy.

Despite AI being a key enabler of digital transformation, organizations may be struggling to find relevant talent in their region or simply lack the budget to attract experienced AI specialists.

usiness leaders may think that this problem can be solved by employing ready-made AI solutions. Unfortunately, organizations still need people with sufficient expertise to deploy and manage these AI-powered systems.

What’s the plan?

An AI implementation strategy definitely deserves more than a few paragraphs. However, given that businesses keep falling into the same implementation traps, it’s worth outlining the key factors that underpin successful AI adoption.


The first step of an effective AI implementation strategy has nothing to do with the technology itself. The key question here is ‘what are our business objectives?’ There is no doubt that AI can greatly enhance production processes or automate manual tasks. However, can it sufficiently meet the end goals of the organization? And most importantly, is AI objectively the best way to fulfill the identified business needs?

Once stakeholders, including managers, executives, and data-scientists, agree that the adoption of an AI-powered tool is economically feasible and aligns with the current business goals, it’s time to define the organization’s AI vision. All in all, you need to have the answers to at least these few questions:

  • How are our competitors using AI and what are their plans for further integration of intelligent technologies into their pipelines?
  • How will AI change our corporate culture, business model and internal processes?
  • Lastly, how drastically are we ready to change? In other words, how far are we willing to go?

Finding the exact use case

Now it’s time to find the most relevant use cases. Try to answer these questions first:

  • Where exactly are we going to implement AI? Are we looking to make operations more efficient or improve our product?
  • Which use cases align with our established vision for AI implementation?
  • Do we have enough high-quality data?

You will most likely end up with a list of use cases. It’s important to prioritize them based on their implementation complexity and potential value.

SEE ALSO: Women in Tech: “Don’t let irrational advice keep you from tech!”

Preparation and execution

AI adoption is never a mild change. It will most likely result in organizational governance changes, cultural shift, and most obviously, new talent hired.

As it was mentioned above, hiring AI specialists is a major roadblock to the technology adoption right now. However, McKinsey suggests that workforce retraining is probably the best solution to talent scarcity. According to their Global AI Survey cited above, 83% of surveyed companies plan to retrain at least some of their employees in the next three years in the wake of AI adoption.

Moreover, other employees that don’t work with AI per se still need to be prepared for the drastic change that AI brings. Everyone within the company needs to have a clear understanding of how AI will change their working routine.

A cultural shift is inevitable and transparency is a key enabler of smooth AI integration. In fact, Jess Shutt, Lead User Researcher on Salesforce’s Einstein and AI team, puts the application of effective change management tools at the forefront of successful AI adoption and stresses the importance of prioritizing people over the technology.

Companies also need to figure out their data strategies. All things considered, a dedicated team of data scientists is often a necessity. Data has to be cleansed and prepared, data sources have to be identified and constantly updated, and data integration processes have to be closely monitored. Moreover, AI does not completely evolve on its own, and algorithms need to be continuously adjusted and optimized. In addition, legacy IT infrastructures often need to be revamped to make room for new data architectures.

We are in the middle of yet another wave of digital transformation with AI at its core. Business leaders’ skills are put to the test here, as transformation is never an easy task. While it’s crucial to remain calm and resist the urge to jump on a hype train, those who fail to capitalize on the present opportunities run a major risk of falling behind the competition.

Yaroslav Kuflinski

Yaroslav Kuflinski

All Posts by Yaroslav Kuflinski

Yaroslav Kuflinski is AI/ML Observer at Iflexion. He has profound experience in IT and keeps up to date on the latest AI/ML research. Yaroslav focuses on AI and ML as tools to solve complex business problems and maximize operations.

Inline Feedbacks
View all comments