August 28, 2024
Throughout history, we often see that good ideas and innovation often take time to gain a foothold. However, this recent wave of artificial intelligence (AI) advancements has been breaking the adoption curve, and we have seen some of the fastest rates of mass usage to date.
ChatGPT was launched in late November 30, 2022. As of early 2024, 22% of firms are aggressively integrating AI into their tech stack, and 33% are using it on a limited basis. Together, they outnumber the remaining 45% that are still only exploring its use. That’s fast — lightning fast. I wish I saw those adoption rates for cybersecurity tools. However, in this case, a little strategic caution may serve some companies well.
In this blog, I’ll spell out what businesses should know before they rush in.
The Challenges of Implementing AI in the Workplace
AI isn’t perfect, but what it can already do is nothing short of incredible. Most business leaders are excited about how it can increase productivity throughout their organization. But there are a few potential drawbacks to its widespread adoption that should at least be considered.
Here are the most common:
- AI is only as good as the information it can draw from
- It doesn’t integrate well with outdated tech
- AI-enhanced features are new, which means they can be expensive and may require new skills and training
- There may be complex ethical and legal considerations depending on which tools are used and how
- Some staff may associate AI implementation with downsizing
- It’s unclear how some AI models may be making their decisions
- Staff may become too reliant on AI technology and neglect critical thinking
- Competitors using the same tools are likely to gain similar insights
9 Steps for Rolling Out AI in Business
When it seems like every business is doing it, and almost every business publication is talking about it, it’s tempting to want to jump on board with AI. But here are the steps I’d recommend to make sure you are getting solid ROI on your company’s investment of time and money, and that your data is protected as well.
1. Build Your AI Council and Define Your Goals
AI-powered tools can require an investment of time and money. So just like any other business investment, before making it, you should understand exactly what you’re hoping to achieve and what success will look like. Only after you understand that can you start strategizing about which AI-powered tools would make the most sense.
One effective way to begin understanding how these tools can be used in the larger organization is by pulling together an AI council for your company. This oversight group will ensure that stakeholders from different departments and executive leadership are able to govern the rollout, adoption, and maturation of AI in your company. They will be responsible for ensuring that the benefits, goals, and KPIs are properly considered, along with the risks and ethical considerations. In addition, this group can provide the oversight necessary to ensure that proper controls are put in place, including compliance considerations, data security, and change management.
2. Complete an AI Readiness Assessment for Your Company
You may need to update your infrastructure — including your network — before you invest in any new AI tools. And you also may need to clean up your data. As I mentioned previously, AI is only as good as the information that it has at its disposal. If you can’t have a reasonable amount of trust in what it’s telling you, there’s almost no point in using it.
Bottom line — if you know you have data silos, gaps, or poor data quality, address those first. You’ll also want to ensure that accounts are employing proper access control, and that employees do not have access to large amounts of data that are necessary for their roles. This is known as the principle of least privilege and role-based access.
3. Think Through Ethical Considerations in Business AI
There are possible ethical considerations about using these tools beyond respecting copyright and intellectual property rights. Some examples that often come up include:
- Because human beings created AI tools, they can also contain many common human biases. If AI is used to automate decision-making and tasks of any kind, it is imperative that companies review the ethical guardrails first.
- Improper access control and labeling may result in the unintended exposure of company and customer information. Depending on your industry and the type of data you handle, this can become a significant liability. Even when data is kept secure, the company will need to consider if it has the legal and ethical right to use it for certain purposes.
- It needs to be clear at all times who is ultimately responsible for a decision. AI tools can seem certain about their responses when they are, in reality, incorrect. This situation is known as AI hallucinations.
- AI tools can reshape what roles are considered relevant. So it’s important to understand possible role and staffing changes — real and perceived — before you adopt AI tools and start crafting messaging.
At Marco, we’ve used robotics process automation not to eliminate staff, but to make our teams more efficient, improve customer service, and reduce errors. But if your staff reacts to any automation initiatives with a great deal of fear, they have their reasons. A significant number of recent layoffs have been blamed on AI. Be aware of these different perspectives as you decide how your business will adopt AI.
4. Understand the Additional Challenges of Data Privacy in AI Implementation
Most security experts advise their businesses not to put sensitive business information into public, less reputable, or newer AI tools. Even in the short time these AI tools have been available, this has proven to be a best practice for a number of reasons.
Depending on the tool, the following scenarios might be possible:
- The platform may use customer data to train models for other clients or share it with third parties
- An unethical user might try to extract private information through carefully crafted prompts
- Sensitive customer data might be exposed, potentially violating privacy laws
Before moving ahead with any tool, review the provider’s terms of service carefully. Some providers have put an enormous amount of time and resources into developing AI tools that are just as secure as anything else you’re already using. For example, the data security boundaries that are built into Microsoft Copilot for businesses are an example of what happens when this technology is developed taking these risks seriously.
5. Evaluate Microsoft Copilot and Other Business AI Tools
These tools may be based on similar technology, but they’re very different. They’re also evolving extremely quickly. When you evaluate an AI tool, you should evaluate all of the normal things, (e.g. user-friendliness, cost-effectiveness, business alignment, and technical compatibility), but pay additional attention to data privacy and security concerns — especially if you must maintain regulatory compliance.
6. Put Together a Realistic Budget
Depending on your business’s AI readiness, you may need to update your infrastructure significantly before you make the leap. You may also need to devote some serious time to cleaning up your data. Then there’s the ongoing costs of the tools themselves and any training required for your staff. These could potentially be big-ticket expenses and should be factored in.
7. Pilot Test Your AI tools
Start small. You might test a tool by choosing a suitable project in a non-critical part of your business to see if it can meet your goals in this area. Then, if it does well, you can try gradually scaling up its use.
8. Provide Effective Communication and AI Training for Employees
Once you’re ready, provide transparent and robust communication around what’s changing and what isn’t. Why are you making the changes? Why is it important to the company’s future? How will it help the company grow and get ahead of the competition? These are the questions that employees will be asking, so I recommend being proactive in your communication.
You may even want leadership or department managers to talk with their teams about a few potential situations that could cause problems. If there is any fear of job replacement in these conversations, it also gives the manager and leadership a chance to clarify the messaging that the company is putting out.
Of course, we must also remember data security. Through communication and training, ensure that your employees have a good understanding of how to appropriately use company data within these tools, including how not to use it.
9. Measure the Continued ROI of Your Business’s AI tools
One of the greatest things about cloud computing is that once you find that a solution isn’t working as promised, you probably haven’t already had to make a significant five-year investment. If your contract is month-to-month, you can simply cancel the subscription.
AI tools are changing quickly, and so are the needs of your business. So the same tool that brought you an enormous amount of ROI a couple of years ago may no longer be the best choice. Continually assess the value you’re receiving; these tools need to keep earning their keep.
Additional Resources for Securely Deploying AI at Your Business
At Marco, it’s our business to get to know new cutting-edge tools so we can advise our clients on which cloud investments are actually worth their while. Needless to say, our teams were very eager to play around with Microsoft Copilot and find out how it pairs with other Microsoft tools to simplify someone’s week.
Our recent webinar explored AI security and compliance concerns in more detail, including how to evaluate an individual organization’s readiness for Copilot. Click the link below to watch it at your own convenience!
