My company is mandating AI usage, what do I do now?

In conversations recently, both a client and a friend shared their experiences with rolling out AI in their companies. In both cases, the companies had issued mandates to use AI without a real strategy for how and where to implement it. While the mandate may drive adoption of AI, a strategy is needed to ensure that it is a net benefit for the company and the employees using it. What can you do when a strategy isn’t articulated?

In one case, the top-down push focused on custom GPTs. People and teams inside the company can create specific GPTs for whatever task they think can be helped. This proliferation of custom GPTs has led to some improvements in efficiency by reducing the time spent on specific tasks. However, employees report confusion due to the sheer number of custom GPTs, a lack of clarity regarding who can create new ones, how they differ, and which one should be used for each task. One individual remarked, “I just don’t know which GPT to use. There are so many of them. How do I figure it out?”

In the other instance, my friend explained that their manager had mandated the use of AI and specified which task should be addressed based on where the organization was spending money on outsourcing. However, it became apparent that using AI for this task would not drive a net benefit for the team despite it being a time-consuming, rote task.  In fact, using AI would require additional oversight from the team as all the output would need to be reviewed. A new AI-driven process would more labor-intensive than the current outsourced workflow.

For companies with mandates resulting in hundreds of custom GPTs, it is crucial to pause and consider the purpose behind each GPT and establish an organized process. This organized approach might involve forming a small committee to identify and prioritize the repetitive tasks such as researching companies, preparing client presentation decks, and drafting proposals. These are ideal opportunities for AI, provided there is a plan and oversight. Either the committee or a designated AI coordinator needs to monitor and consolidate efforts across teams ensuring that similar GPTs are combined to create best-in-class solutions rather than ending up with multiple tools performing the same function.

In the case of my friend, it was important to step back and assess which tasks AI should be used for, and how to establish an effective process. My advice was to identify the top ten manual tasks, then determine which three could most easily be handled by AI (tasks that are time-intensive but not highly creative or requiring significant oversight). For example, document review and summarization is a practical use case depending on your company’s policies for uploading company information and documents. Start with this single workflow and use the win to build momentum for automating the next one.

 

TL:DR

• Appoint a point person or small committee to gather ideas on where AI can best help your department.

• Work with your company’s AI experts (or become one yourself!) to implement and learn what adds the most value.

• Take these learnings and lather, rinse, repeat.

Of course, the real power of AI comes when you move beyond the chat interface and build agents linked directly to underlying systems and data. But that is the topic for another post.

If you are the leader responsible for your company’s IA strategy, check out this blog by my NMP partner, Gerard, on how NMP approaches setting AI Strategy for an organization.