What Is the Cost of Implementing AI Today?

What Is the Cost of Implementing AI Today?

You’ve surely heard exciting news about artificial intelligence (AI), but is it right for your company? After all, you’re seeing headlines about investors pouring funds into AI: SAS is investing $1 billion in AI in 2019.

Fortunately, there’s some good news. The cost of implementing AI in the enterprise is steadily falling. Before you start to build a business case in Excel, you need to answer a key question.

Before You Count the Cost of AI Implementation

Before you dive into the details of the costs of AI implementation, you need to look at the bigger picture for a moment. Specifically, look at the opportunity cost of ignoring AI. If you weigh the costs and decide AI is still too expensive, what will happen to your organization?

Take a look at the short- and long-term implications of ignoring AI implementation. In the short term, you’ll start to see your growth plateau as other competitors boost their growth rate. In the long term, your organization’s productivity levels are going to decline. You’re going to start bleeding IT staff to competitors who are investing more heavily into innovative technology.

Given those realities, reframe the question to, “How much should we invest in AI?” rather than, “Should we invest in AI at all?” Now, look into the different cost elements to bringing AI to life at your company.

Identifying the Cost of Implementing AI: Direct Costs

While AI may feel new and innovative, implementing this technology is like any other project. You need to build a reasonable business case and win support.

Develop Use Cases

You’ll need to use your own time at work or somebody else’s time to analyze the ways AI could be used. While there’s no cash cost to this activity, you may have to negotiate internally to delay other work to make time available. In one or two days of analysis and research, you should be able to come up with a few use cases. If you’re using an employee with a $75,000 annual salary to do this task, a two-day work effort will cost the organization around $500-$600.

Purchase (or Build) the AI Platform

The next cost is to buy or build an AI platform. Some companies, such as Microsoft and IBM, are investing heavily to build their own AI teams and applications. The other 99% of organizations don’t have that skill set internally. Thus, you’ll have to look at purchasing an AI solution. The cost of an AI solution is all over the place. The recommendation: Look for an AI application focused on solving a clearly defined problem, such as handling routine IT security tickets.

Pilot Project Fees

In some companies, you’re expected to pay for a small pilot project first. By experimenting at a small scale, the organization sees some results and reduces the risk of a large-scale project failure. To estimate this cost, check with other project managers to see if this cost applies in your situation.

Indirect Costs for AI Implementation

With the direct or “hard” costs estimated, it’s time to look at the indirect costs to bring AI into production.

Project Management

Depending upon the size of your project, you may need to involve a project manager, project assistants, and business analysts. This cost may be charged internally to your department through a project cost. Alternatively, you may have to contact a recruiter or consulting firm to bring in talent. This cost component will ensure that your AI implementation project is well managed and is implemented on time.

Resource: Are you a novice in project planning? Use our article on planning a multi-factor authentication implementation project to guide your approach.

Training and Change Management

If you implement an AI-powered virtual agent to handle routine tasks, that’ll have an impact on your company. Specifically, your help desk staff may wonder what these changes mean for their daily work. To address this concern, set aside some time and budget for change management. For example, you may increase budget to help affected staff earn new certifications (e.g., check out TechRepublic’s article titled 13 In-demand Tech Certifications That Will Help You Get a Job for ideas), so they can contribute more value to the organization.

Monitoring Cybersecurity
Left to its own devices, an AI implementation poses increased cybersecurity risks. That’s why you need to set a budget and time aside to develop a cyber-monitoring process. For example, you may ask your team to conduct simulated attacks against the platform to ensure it’s resistant to attack. You can minimize this cost if you choose Avatier, because we have significant expertise in managing cybersecurity risk.

Written by Nelson Cicchitto