We’re all feeling it right now.
When’s the last time you went a week at work without some conversation about AI?
To some degree, AI use is becoming part of everyone’s job. The technology is transformative. Everyone can see it, as we use tools like ChatGPT, Copilot, and Claude for work or personal use. We know this big wave of change is coming.
And yet to some degree…the change isn’t coming to fruition yet.
McKinsey’s 2025 State of AI reported that 66% of companies implementing AI solutions were stuck in pilot of experimentation phases.
Here’s a survey of technology leaders in the financial industry who are feeling the weight of supporting legacy systems while struggling to get returns on new technology innovations.
What does it actually take to get AI working for your business?
Everyone is still building the plane as they’re flying. But there seem to be three principles worth following and focusing on now, that can make a difference this year and into the future.
Get your data in order.
In one way, AI tools are the same as any software you’ve ever used.
Garbage in, garbage out is a principle as old as computing. AI just makes the consequences more visible, more quickly.
AI tools run on context — your data, your history, your specifics. A large language model doesn't know your work, your products, or your processes unless you give it that information in a structured form it can actually use.
There are two main ways you can provide context to an AI tool. If you’re just using a basic chat interface like ChatGPT or Claude or Copilot, you can upload documents or share examples and build a base of reference context. Or you can directly connect these tools to the ones you use for your work everyday.
Bottom line: the quality of your data directly correlates to how well AI can support your work.
Are your business files neatly organized on OneDrive? Or does your business have core information stuck in spreadsheets saved locally to someone’s desktop? Anyone inspired to do some spring cleaning?
Document your processes before you automate them.
AI adoption will force you to find a new level of clarity regarding how your business actually works.
Most teams have processes. Few have them written down. Even fewer have them detailed to the level needed to identify where AI could support the process, or to have AI totally take over a process by running an automation or workflow.
Doing the work of documentation reveals something valuable on its own: the best candidates for automation. Sometimes the right move is to automate a step. Sometimes it's to build an agent that manages an entire process. Other times, it’s actually best to keep a step or a whole process human-led and managed, because that human touch is the meaningful value derived from that work.
For us at Lighthouse, we spent the last two years implementing Entrepreneur Operating System (EOS). Now, across all our different business functions, we have a foundation to draw upon and see where AI can support us, speed up routine work, all that.
If you haven't documented your core processes yet, that's a great place to start — regardless of what you decide to do with AI.
Decide what a human still has to own.
AI makes confident mistakes. Have you ever asked ChatGPT to answer a question, it responds, you tell it the answer is wrong, and it gleefully tells you you’re right before producing a new answer?
That’s because these tools are prediction engines, not truth-seekers. They can produce polished, persuasive, flat-out wrong outputs — especially without the right context.
As with any new initiative, it’s important to continually keep humans in the loop when building new AI tools and automations into your work.
Who is responsible for testing these tools, collecting feedback and implementing improvements? Who’s keeping an eye on security and access and training team members on how to use these tools safely at work?
And critically: what parts of your work will you intentionally not hand off, because the human touch is part of what makes it valuable?
So what do we do with all this?
Regardless of where you’re at with bringing AI into your business, there are things you can do today to ensure those solutions make a measurable difference and change how you work tomorrow.
The businesses that do AI well are clear about what stays human, and they protect those things on purpose.
Speaking of keeping the human in the loop…
Humans on your team will always be in the loop when it comes to managing their email boxes. Humans are susceptible to certain communication styles. And hackers take advantage of that in phishing campaigns.
Technology and AI play a great role in supporting email security, and our team uses tools like Proofpoint to secure our own email domains and our clients’.
Want a free way to train your team how to avoid getting phished? Learn the SLAM method. I learned this in a security training years ago and never forgot it. Hopefully it can help your team too.
What We’re Hiring For This Week
Our Technology Talent Managers are always tuned in to the market, creating opportunities for technology professionals and innovative businesses to connect.
Here’s a few roles we’re recruiting for this week:
AWS Technical Lead: This role leads the design and implementation of cloud-native solutions on AWS, with a strong focus on Master Data Management (MDM). 6+ month remote contract-to-hire opportunity, $65-70/hour.
ServiceNow Technical Lead: This role leads the transition of ServiceNow applications from implementation to production, ensuring stability, scalability, and long-term sustainability. 6+ month contract-to-hire opportunity, remote but must work Eastern time hours, $65-70/hour.
DevOps & Application Systems Analyst: This role serves as a principal application designer for major GitLab and Sonar modifications, leveraging analytical and technical skills to evaluate client requirements and processes. 12+ month contract, four days onsite in Buffalo, $55-60/hour.