Search

Avoid These 5 Major AI Investment Pitfalls







So maybe you’ve decided to “get on the bandwagon” and do this AI thing. You’ve read up on why you need to implement an artificial intelligence effort and you’re ready to initiate a machine learning movement in your company and start reaping the benefits like a pro. Like, yesterday.

After all, you don’t want your competitors to leave you behind in the dust, right?


Well, before you get started you should know that rolling out an AI initiative – at least a successful one – is not an easy row to hoe. In fact, it’s estimated that 75% of organizations fail at their initial attempt to implement AI in their businesses.


Understanding up front what could derail your AI efforts can help you be part of the other 25% that enjoys all the benefits that come with a well-aligned, educated, and purposefully thought-out AI implementation.


Here are five common pitfalls that you can avoid with careful planning and thoughtful attention when implementing an AI initiative in your company.


Pitfall #1: Leadership isn’t really behind the effort.

Many times, AI has the full-fledged support of one or two champions within an organization, but everyone else who has the pull to actually make it happen are just… meh. The AI champions are left swimming upstream, fighting the currents of change-resistance in tepid waters. Even if the C-levels in your company are uber-excited about launching a sexy new AI initiative at your company, they may not fully understand how to execute it effectively.


Successful AI implementations require 100% commitment from the top down – and this doesn’t stop at simple dedication to and belief in the effort. It also means that those at the top are fully educated in the intricacies of data science: what it is, what it does, and how to bend it to your will.

Reluctance to spend the time educating top leadership in data science and how to effectively implement it (or simply ignorance of this necessity) is a sure-fire way to sabotage even the best of intentions.


Pitfall #2: The company’s culture isn’t aligned.

One of the top killers of any data science-related initiative is resistance to change. If everyone in the company – from the CEO down to the most seemingly inconsequential (to the project) front-line clerk – is not immersed in the value of your AI mission, it is destined to fail. And if leadership fails to educate the entire company on the value of supporting the company’s AI initiative and the importance of becoming a data-centric organization, employees will fail to capture accurate data and may end up unwittingly (or even wittingly) sabotaging the project.


As we mentioned above, a successful AI initiative requires in-depth knowledge, participation, and commitment from the top down. Without senior-level guidance and support, this sense of value in the project cannot permeate your company’s culture, and valuable investments in both time and money are washed down the drain.


Pitfall #3: Focusing on the wrong opportunities – no clear strategy.

Many companies approach a data analysis initiative with nothing but exploration in mind. They strike out blindly, throwing all the data they can find into their shiny new machine learning tool, then sit back and wait for the data gods to shine a heavenly light on some hitherto unseen insight that will transform their business and propel them to the top of the food chain.

It simply doesn’t work this way.


Other companies may start with a general goal in mind but haven’t clearly defined a specific question to answer, so they end up chasing proverbial rabbits down interesting, meandering data caverns, never quite arriving at the insights needed to enact real change.

Still other organizations jump in with a specific question to answer without fully analyzing the value of investing in that particular question. Lower hanging, ROI-heavy fruit are left dying on the vine, and the effort fizzles out when juicy benefits aren’t realized quickly enough to satisfy investors and key stakeholders.


Investments in AI must be predicated on specific, clearly defined questions which are carefully selected to align with existing business objectives and are likely to produce the measurable, actionable insights you need to make game-changing decisions.


Pitfall #4: Not involving the right people.

When it comes to assigning responsibility for an AI initiative within the organization, most fingers point to the IT staff. After all, they’re the techies, right? They manage the data in the first place…

However, while the IT folks at your organization may be experts at the technical aspects of managing data, they most likely lack the business knowledge inherent in your other company resources who live the business day in and day out. The ones who know your customers, your products or services, and your organizational processes like the backs of their hands. The ones who live and breathe your business operations on the daily.


Even companies who invest in adding a Chief Data Officer to their C-Suite (Gartner predicts that 90% of large organizations will do this by the end of 2019) sometimes fail to provide that person with access to the people within the company who understand the intricacies of the business. They unfairly expect this unfortunate person to waltz in with their data wizardry, snap their little magical digits, and deliver value from within a vacuum.


Furthermore, many companies fail to employ resources with a deep understanding of quantitative research, data models, and statistics – the backbone of true data science initiatives.


The truth is, data science projects require the diverse perspectives and expertise of a variety of roles within your company who can catch mistakes or misconceptions before they derail a project and are positioned to help ensure that the project remains aligned with the goals of the business.

When your data team is comprised of the right people with the right knowledge sitting at the table, working on the right opportunities… that’s when the magic starts to happen.


Pitfall #5: Data is not optimized for maximum value.

Guess what? If data is one of the most important building blocks for your data initiative, your data infrastructure is the cornerstone. Elementary, my dear Watson. Right?

Sure, you’ve got tons of data you can export from your CRM, your sales channel can produce years of sales history, and your financial system hosts a multitude of bits and bytes just waiting to be “scienced”.


But how clean is that data? How complete is it? You know what they say… garbage in, garbage out! And how easily can your data from disparate origins and business-driven silos be integrated and unified?


Cleansing and integrating data from multiple sources can easily consume up to 80% of a data scientist’s time, and cost between 60 to 80% of the total cost of your project.

Most companies fail to recognize the enormity of effort required to source, cleanse, and integrate data before the sexy analysis part can even begin. In fact, what is required is a total and fundamental cultural change toward weaving data into the very operational fabric of your business. Are you prepared to meet this requirement?


Conclusion

While the road to success is often paved with the best of intentions, the hard truth is intentions and enthusiasm alone won’t cut it. Success in implementing AI in any business begins with dedicated and educated leadership, cultural alignment at all levels with business goals, and many precisely fitted paving stones.

It isn’t easy, but the rewards are well worth it – if done right.


63 views
Get in touch