← All posts

AI Is Already Working in Your Company: A Practical Overview for Busy Executives

AI is already working in your company.

This is for the non-technical executive that is trying to better understand AI in the context of their business. Maybe you build HVAC solutions. Maybe you work on Energy solutions. You are too busy to come up for air long enough to wrap your head around this fast-moving advancement in technology. If this sounds like you, then I hope this helps.

The good news is you are already benefiting from AI even if you don’t explicitly know it. There is already an “AI workforce” inside your company. Distributed across your employees and software systems, continuously expanding.

Let’s peel back a few layers and see exactly where it’s showing up.

AI in Chat Tools

This comes in the form of an employee using one or more AI chat tools like ChatGPT, Claude, Gemini or an equivalent in a web browser or a desktop app to help them do their job. These tools are inexpensive (typically $20/month per person) and can do A LOT including connecting to things like email, calendars, documents and many SaaS products. The costs are likely buried in the P&Ls across your departments.

The bad news is these inexpensive AI tools have a hidden cost. The data your employees send the AI tool can and very likely will be used to train future models. You are trading off data privacy for affordability and convenience. You are trusting the AI tool provider (OpenAI, Google, Anthropic, etc) to be good stewards of your company data that is inadvertently shared.

You can address this by providing your employees with the “For Business” version and putting controls around “connectors”. Most major providers (OpenAI, Anthropic, Google) now offer Enterprise plans that explicitly do not train on your data and include zero-retention options. The extra cost is real, but so is the peace of mind. This also requires oversight for what is allowed for access. Very doable, likely important and a wise investment for your company.

AI in SaaS

AI as a feature inside SaaS products is also becoming very common. Your teams are likely using AI inside of a SaaS product such as Zoom. The costs are more visible as this would typically be additional expenses on your monthly statements. The AI in SaaS tend to be limited to the vendor’s ecosystem. In other words AI in Zoom will help you summarize a Zoom call but it won’t help you check your inventory that lives in another system.

In an HVAC business this might mean the AI inside your service-management platform is already predicting equipment failures from sensor data. In energy, the AI in your SCADA system is optimizing load balancing in real time. If you are on older systems that have not modernized with AI then think of this more in terms of what is possible.

AI in IT/Development

If you have technology teams or technology contractors they are using AI to build, test, configure, deploy and monitor. These AI tools are often from the same providers of the “Chat Tool” but with more capabilities geared towards technical roles. They are more expensive ranging from $100 to $1,000s per month per person. The costs are likely buried in the P&L alongside other costs such as “development tools” or “internal software expenses”. Contractors are not charging you for this as it is reducing their labor costs.

Same bad news here as the Chat Tools i.e. your data is not private and you are trusting the AI provider to use your data for model training carefully. For technical roles the system access is a bigger consideration. The person in marketing by default already has limited access to things. The person in IT by default has much more access to things. Fortunately IT has to be highly aware of security as a fundamental skill of the job so they tend to be cautious by default.

You can address this in the same way by paying for the “For Business” or Enterprise version and ensuring there are controls around IT specific access. It will cost more and it will require more oversight. Still very doable, likely important and a wise investment for your company.

Agentic AI

There is a lot of buzz around AI Agents. Think of an agent as a virtual team member that can help operate or manage elements of your business automatically or at least semi-automatically. You can have agents doing work for you in just about every business unit you can think. Agents that run marketing functions. Agents that handle customer service. Agents that monitor systems for you. It is a wide ocean of possibilities (often referred to as a “fleet of agents”).

Your technical teams are likely building their own agents to help automate routine tasks. Non-technical teams are likely using agents from their SaaS platform that runs their particular slice of the business. Costs are likely in the P&L for technical teams under similar labels such as “development tools” or “internal software expenses” and can be wildly variable (could be in the tens of thousands of dollars). Non-technical team costs will be bundled with the SaaS costs.

Personal Agents

A personal agent is similar to a company agent but only interacts with owner. It is like having a team of highly skilled personal assistants that can help you with personal or professional endeavors in just about any domain. The personal agents retain preferences, memories and build a working knowledge base of the owner. This is where tools like OpenClaw sit. These are often referred to as “Orchestrators”. An orchestrator sits on top of your agents and allow you to communicate and collaborate across your fleet just like you would employees or your friends. You can text them, use Teams, Slack, etc. I know, it sounds a little sci-fi. But it’s real, and it’s already happening.

It is possible that your employees are using personal agent fleets to do parts of their job while also helping them manage their personal lives. This type of agent can feel the most human-like and can be highly adaptable to bridge many different systems, personal projects and work projects.

The agentic type of AI is going to continue to become more and more prevalent in both business and personal with the lines continuing to blur as they become easier to set up and manage. Many SaaS products are building “agent friendly” api access to embrace this movement versus fighting it. Many of the AI providers now offer hosted virtual agents and many companies are launching agent hosting solutions to make it easier and easier for people to spin up these personal agent fleets.

Conclusion

I hope this was somewhat helpful in illuminating the reality that your company is using AI even if you don’t have an explicit plan, strategy or budget for it. It is more beneficial than harmful but you should seek to understand if you are inadvertently sharing company data and allowing potentially harmful levels of access to your systems. You should lean in to this. Work to embrace AI and provide safe and powerful access to your teams. Resistance is futile.

Agents are very powerful. It is likely that everybody will have a fleet of agents that work for them and for their company. It is analogous to when the first iPhone came out and now we all have smartphones. They are extremely useful and will continue to be more and more accessible as agent solutions get more polished and easy to set up and configure.

To shed light on your company’s current situation you can start with a few simple questions:

  • Which AI tools are already on company credit cards or buried in P&Ls?
  • Are we on the free/consumer tiers or the Enterprise versions?
  • Is there an AI usage policy and who owns it?