We use artificial intelligence in many facets of our work.
According to a survey by RLEX, the most common use cases of AI are geared towards increasing efficiencies or worker productivity (51%), informing future business decisions (41%), and streamlining processes (39%).
Clearly, AI can be helpful, but it also comes with its costs. To help you decide whether to implement AI at your organization, we’ll review common benefits and costs associated with it.
The benefits of AI
Professor Agarwal, of University of Toronto’s Rotman School of Management, made the case that AI reduces the cost of prediction. He essentially explains that humans spend a lot of time and work trying to predict outcomes, when AI can just be trained to do so (without all of the cognitive biases). When AI has made its prediction, people (like your employees) can then use their judgement and make informed decisions quickly.
Agarwal used the example of self-driving cars. Like humans, the car needs to predict which way to go to avoid crashing. Although the self-driving car hasn’t been perfected, when you take a look at your business processes, like supply chain management (a highly predictive process), it’s clear that AI can be helpful.
AI also allows organizations to up-skill their team and foster stronger employee loyalty. According to ZDNet, “If companies invest in their employees as part of their AI initiatives, they have a better chance of retaining employees and of building the skills and capabilities of their human workforces.”
Now, what are the costs of AI?
We’ll break down its costs from a monetary and time-to-value perspective.
The monetary costs
Custom solutions: According to FX, custom AI solutions cost anywhere from $6,000 to over $300,000 (that number also includes development and rollout). Additionally, for AI consultants, you’re looking at around $200 to $350 per hour.
Pre-built solutions: If you’re looking for something a little lighter, like a prebuilt chatbot, you should expect to pay around $40,000 per year.
Considerations: According to Sandra Carrico, Vice President Engineering and Chief Data Scientist at GLYNT.AI, her organization’s AI project costed 15 times more than they expected. This included implementing their AI infrastructure (their biggest cost), customer data protection, and more.
How long until I see value?
The length of your AI project will be based on your company, what the project is, and what resources you need, to name a few.
Prasad Vuyyuru, of Infosys Consulting, says that your first AI win should be within 8-12 weeks. He argues that seeing value quickly can be effective in keeping your senior stakeholders engaged.
Beyond a quick first win, you should expect your AI project to be a long-term initiative, as it’s in a constant circle of learning.
Do your benefits outweigh the costs?
If you’ve determined that your use case for AI outweighs the costs and time-to-value, keep these tips in mind:
- Spend time taking care of and cleaning your data. It is a key component of a successful AI project. As ZDNet explains, “AI emulates the human mind, but at warp speed. It is only as good as the algorithms and data that are fed to it.”
- Make sure to be very specific on what you want your AI system to help you predict. What is your AI goal? Agarwal explains that “Due to the methods used to train AIs, AI effectiveness is directly tied to goal-specification clarity.”
- Don’t stop teaching your AI machine. As it gets more information, it will be in a better place to predict and help your project.
Wondering what CIOs think about AI? We plan to cover this in our Modern CIO newsletter, edited by 7X CIO Mark Settle. Sign up for the newsletter.