AI Adoption Done Right: The 5-Step Readiness Framework – Facility Executive Magazine

17 December 2024
colind88
News Feed

By Matt Deehan
From the December 2024 Issue

Facility management teams are eager to take advantage of the game-changing services Artificial Intelligence (AI) can provide, like predictive maintenance, resource allocation, and so much more. However, many FMs face potential risks in the speed of their adaption, facing obstacles in going too fast or slow. With risks like security exposure, unclear objectives, or data discrepancies, moving too fast toward AI can be detrimental to an FM team. Alternatively, failing equipment, labor shortages, and the threat of falling behind pose risks in moving too slowly toward AI adaption.

So how can facility executives effectively approach integrating AI into their daily operations without getting ahead of their skis? By conducting necessary, methodical research and planning before beginning any implementation.

This 5-step AI readiness framework can help FM teams prepare for successful AI adoption.

1. Stakeholder Task Force Led By FM

It’s important to think of AI as a business initiative rather than a technology initiative. Since facility managers are focused on improving and optimizing facility service delivery, the AI initiative should be led by the FM team. However, because AI is both complex and nebulous—and because it impacts other critical functions of business—FM executives can’t operate in a vacuum.

Establishing a cross-functional task force is critical. The team should include FM representatives who understand the day-to-day operations that AI can improve; IT personnel for addressing technical, information security, and data integrity issues; legal for compliance and governance; and external FM service provider partners who are integral to service delivery.

In addition to these roles, employing an executive sponsor who is aligned with company strategy and innovation can help drive priorities, champion the AI initiative across the organization, and secure required resources.

Above all else, remember that AI is at an inflection point when it comes to Generative AI technologies such as Large Language & Foundation Models (LLMs). Although they are in their infancy, they are advancing at lightning speed. This means everyone will be learning as they go. It also makes regularly scheduled task force meetings a must. Use these meetings to share and discuss new learning and developments and, when needed, readdress objectives, solutions, and phased implementation plans.

2. Establishing AI Objectives

There are many potential ways AI can improve and enhance facility services, both internal to the organization, such as employee productivity and operations enhancement, and external, like enhanced customer experience. Facility executives should identify and rank specific and, when possible, measurable goals that align with their corporation’s strategic priorities. Here are some examples:

Cost reduction: An FM executive might target a 15% decrease in energy costs. This could be accomplished by integrating AI with energy management systems to optimize HVAC loads and performance based on real-time occupancy and weather patterns. AI could also analyze historical data, forecast energy usage, and recommend proactive adjustments to meet this goal.

Equipment uptime: Facilities that want to save money and reduce unplanned downtime could use preventative maintenance tools that leverage machine learning and AI to analyze equipment performance data and recommend maintenance procedures for extending lifecycles by a desired percentage.

Skills improvement: FM managers can have their engineering team take a skills assessment and load the results into an AI tool that determines each employee’s relative strengths and weaknesses and suggests appropriate training. Taking this to the next level, FMs can leverage this data to support a real-time, enterprise-grade domain knowledge base for their employees, delivering essential information with governed permissions across the organization.

Enhancing tenant relations: If this is a strategic priority, improving tenant satisfaction might involve:

Deploying AI chatbots for immediate response to tenant inquiries.
Using AI to derive insights from work order data, such as which tenants are requesting support the most, what’s the most critical issue, and ideas on how to prevent the problem.
Using an AI tool to improve FM team responses to tenant inquiries. For instance, instead of an email saying, “OK, got it,” AI might recommend writing, “Thank you for your request. I’m investigating your issue, and it is now being worked on. I’ll be sure to keep you updated on the status.”

3. Data Availability And Integrity

Most organizations are increasingly data-driven. For example, some buildings use sensors to automatically gather data about occupancy patterns, weather, and HVAC performance, or barcodes and cell phones to scan equipment info during maintenance rounds. Others manually enter data. Too often, however, this internal source data isn’t leveraged to its fullest potential.

Facility managers can embrace automation by leveraging various AI tools and capabilities that accelerate the delivery of real-time, insightful data.

Keep in mind that generative AI is only as good as the data feeding it. If data is incomplete, inaccurate, or subjective rather than objective, any insights gained and actions taken may not be appropriate. They could lead to poor decisions and damaging outcomes.

That’s why, before implementing AI, organizations must ensure they have a strong data-strategy foundation in place. This includes examining the availability and integrity of their internal data systems and source data.

For instance, does the organization have enough sensors to gather all the information required for trustworthy AI analysis? If not, is there room in the budget to purchase more sensors? Do all pieces of equipment have barcodes and do all personnel scan them during their rounds? If not, rectify this before feeding data into AI solutions.

Some organizations may be tempted to use outputs from third-party LLMs such as OpenAI ChatGPT or Google Gemini. However, outputs from LLMs that are hosted outside the organization are not validated standard operating procedures. Tools like these provide results based on information that exists online, in the wild. As everyone knows, a lot of information on the Internet is questionable or so specific to certain situations that it shouldn’t be generalized or considered for standard operating procedures.

Using public-facing models for any confidential, proprietary, or enterprise function is strongly discouraged. It puts companies at tremendous risk of having sensitive information about their operations, finances, employees, and more in the wild.

4. Security First: Data Protection, Privacy Policies, and Governance

Organizations must take a security-first approach to AI technologies, usage, and solutions. This includes specifying that AI can be leveraged when it’s with a company-approved, owned, and licensed solution that is sanctioned.

It also includes establishing a clear Third-Party Generative AI & Interaction Policy that ensures no company confidential information is entered as inputs to these LLMs, to avoid security and data privacy risks.

Organizations must also make sure sensitive data about operations, tenants, employees, partners, and vendors is protected and complies with appropriate regulations, such as HIPAA, California Consumer Privacy Act (CCPA), and OSHA.

Training sessions are also vital for educating staff about the importance of data privacy and protection and informing them about the policies they need to follow.

5. AI Solution Evaluation

This requires identifying and understanding how standalone and integrated AI technologies work and their potential impacts. Because solutions are evolving rapidly, be diligent about staying on top of them. Remember there is no one-size-fits-all solution because every facility has different needs. Many times, an organization’s FM service delivery partner can help or even lead evaluations and keep the task force informed about new developments.

Facility management systems and applications such as project management, computer maintenance management, building automation, and work order management are starting to incorporate generative AI enhancements. Businesses can also fine-tune generative AI models to create their own custom systems.

The AI task force should look into their vendors’ existing and planned offerings—as well as systems their service delivery partners use—and evaluate them based on security, usability, interoperability, and functional capabilities, and adherence to policies. If needed, and the budget exists, consider replacement systems.

The cost implications of AI implementations also need to be considered. While AI offers long-term savings, initial setup and integration costs can be high, so organizations should weigh this against future benefits.

After evaluating systems, applications, and platforms, best practice is to begin by identifying a high-impact use case. Then work with the appropriate technology providers on delivering a Proof of Concept against the business requirements, and measure this against success metrics.

Artificial Intelligence is an incredibly exciting development in business. It promises enormous benefits to the Facility Management industry if properly understood and implemented. All FM teams should be in discovery and planning mode right now. Those who aren’t are at risk of falling behind.

Deehan is an industry veteran with twenty-five years of experience in the facility services business. He manages UG2’s facility management solutions team. Deehan has a deep knowledge and understanding of technical FM solution development, process improvement, transition project management, and strategic facility operations.

Do you have a comment? Share your thoughts in the Comments section below, or send an e-mail to the Editor at .