The Biggest Mistake CIOs Make in Adopting Artificial Intelligence
Many enterprises have taken to their digital transformation journeys with an increased focus on Artificial Intelligence. CIOs have started incorporating Artificial Intelligence into their strategies as they seem appropriate for their enterprises.
While the CIOs look to ensure their technology strategies are aligned with their enterprise’s business goals and at the same time balance their budgets between the mandatory programs and building new technology capabilities, it’s often hard for them to escape the “silo mindset”.
What is a “Silo mindset”?
Let’s say you are the CIO of a large financial services organization. Among the many technology goals you are trying to achieve, you are also contending on the problems your business partners came to you with. The business may be asking you to find how you can help them cut down the increasing volume of time being spent reviewing and categorizing certain sensitive business documents.
It now becomes fairly obvious to you that it’s a classic case of using NLP and Machine Learning and building a text classification model that can classify and tag these documents. So, all that’s required to be done is to fund a project that can accomplish this. Wouldn’t that be just as simple?
There is just one problem. You have no existing investment in NLP or Machine Learning technologies. You have no hardware to support anything Artificial Intelligence. You probably would be funded by your business partner, but then you also have those regulatory projects that you have to complete this year at any cost, lest you be fined by the regulators for non-compliance. The question now becomes — what is the quickest way to achieve the business goals? After all, as a CIO, isn’t it your job to align your strategies with the business goals?
This is where you enter into a “Silo mindset”. You decide that all you need is this one-off project and you figure out a way to leverage any part of your existing hardware to scale into what’s required for an MVP of the new AI solution.
Thinking outside of the “Silo”
Solving your first Artificial Intelligence problem by executing a one-off project would mean, you are equating an investment in Artificial Intelligence as an investment in any other traditional technology development programs. You would have failed to consider the potential re-usability of the models into other scenarios and blocked any further scaling of the solution you built to solve the problem.
Artificial Intelligence problems never come single. The moment you’ve forayed into one successful implementation, you are soon going to be faced with the next problem to be solved by AI and then another after that. But you now face a bigger set of problems, for which you don’t necessarily have any answer.
- How do you standardize the technology stack and hardware stack across your Artificial Intelligence programs?
2. With so many new problems, how could you build smaller experiments with different AI technologies to assess what works best for different types of problems?
3. How can you build an AI skill-set within your teams?
4. How can you balance between researching new ideas versus implementing AI solutions on identified problems?
5. Should you consider adopting Cloud based Artificial Intelligence deployment to optimize on cost and maintenance?
6. How can you reduce the time to market for solving new Artificial Intelligence problems?
As a CIO, unless you think outside of the “Silo” of an individual problem — the one-off project that solved your first AI problem, your organization enters the crossroads of ambiguity in attempting to solve any future AI problems.
Think of Artificial Intelligence as a “Practice”, not as a one-off “Project”
Adopting Artificial Intelligence in an enterprise is not a run-of-the-mill affair. Planning and execution of AI projects requires adopting differentiated metrics than that is done for traditional projects. AI also brings with it the need to think outside of the “Silo” project, and to envision solving a problem in a way that’s repeatable and predictable. Your organization should begin to cultivate a culture of understanding the need for the 6 questions outlined above.
Consider Artificial Intelligence as a Practice. Consider investing in a small Center of Excellence within the practice, that can perform relevant research, build small experiments, assess hardware needs etc and roll those findings into upcoming AI problems you may need to solve. Build in-house tools and frameworks that can aid in solving not only the first project, but which can also potentially be reused or re-adapted for future projects.
A “Practice” is harder work. It means you invest more than you do on a one-off project. It means that, within your organization you are building another miniature organization that would have a mind and mood of its own. It’s a holistic re-think of how your hitherto non-AI organization has functioned and now gets ready to be transformed into an AI-capable organization.
But to derive the most technological benefits and to have a justifiable ROI, consider Artificial Intelligence as a “Practice”. Not as a one-off project.
