How AI Tools Change Your Team’s Work (And What to Do About It)

---

Your new AI tool is fully integrated into your team’s workflows. You’ve made sure everyone has the training they need. Now all you have to do is sit back and watch the productivity gains roll in, right?

Not quite.

Workplaces are complex organisms. You can’t change one thing (like automating certain tasks) without ripple effects. The seldom-discussed Phase 2 of AI transformation is figuring out how automation tools change employees’ work and adjusting as needed to account for those changes.

In this piece, we’ll take a look at why AI tools can have ripple effects within an organization, how to identify these effects, and how to adapt to make sure you’re still enjoying the benefits of increased automation.

The Ripple Effects of AI Tools

Picture a call center for a telecom company. The company introduces an AI agent that can handle the simplest 15 percent of customer calls. After a period of adjustment, the tool is working great – but call center agents are facing new challenges.

Now, because all the simplest calls are handled by a bot, human agents are facing more complex customer scenarios in every conversation. What’s more, customers with complex issues still have to go through the AI agent to get to a human – and they’re often frustrated by that. In some cases, the AI tool isn’t providing helpful summaries of a customer’s needs, so that agents are forced to ask people questions they’ve already answered.

The result for call center employees is that the nature of their work has changed. Now, instead of dealing with some combination of simple and complex calls, they’re dealing exclusively with complex calls and with a higher portion of agitated callers.

To do this changed work, they may require a different skill set or different tools – stronger deescalation skills, for example. This might require additional training for some employees.

And what about the systems call center agents use to look up customer accounts? Are those efficient? Do they load quickly and let agents view a customer’s entire history with the company from one screen? If not, the wait times while agents load information could further exacerbate caller agitation.

Things continue to ripple from here. Will agents need additional breaks to recover from stressful interactions? New mental health benefits? Will you need to have additional managers available so agents can escalate calls more easily? Will you need to give agents greater latitude in granting refunds or applying promotions to appease angry callers?

I could go on, but you get the idea. Automate work, and the remaining work changes. Now let’s take a look at how you can identify and address those changes before they cause  new problems.

How to Identify What AI Tools Change

While we don’t know in any given AI transformation what exactly a new AI tool will change, we do know that it will cause changes. This is why it’s so important to have a way to track those changes and measure their impact on employees’ ability to do work.

FOUNT’s work friction framework is designed for exactly this. It involves conducting targeted surveys of impacted workers. Unlike traditional employee experience surveys, which typically ask how employees feel about their work, work friction surveys ask about the work itself: how did the AI tool impact your ability to do xyz. What is your satisfaction with doing abc. Etc.

From these surveys, we gather data on two things: moments (aka tasks workers complete throughout the workday) and touchpoints (aka people, processes, and things workers interact with to do their work).

We then assess two components of the moments and touchpoints we measure:

1. Impact, meaning to what extent it impacts the work a person does; and
2. Satisfaction, meaning how well that moment or touchpoint is currently working.

With this data, it’s easy to measure where the AI tool is having the greatest positive and negative impacts – and where unexpected consequences may be playing out.

Now let’s take a look at how to react to that data to make sure your organization continues to benefit from your AI investment.

How to Adapt After Implementing an AI Tool to Maintain ROI

Let’s return to our hypothetical call center. The AI bot is handling 15 percent of call volume, but the remaining callers are both more complex and more demanding than they were pre-AI tool. As a result, the productivity gain you’re seeing is only about half of what you budgeted for.

(Related: How to Assess the ROI of Current AI Initiatives & Prioritize Future Investments) 

You need to change something to get those productivity numbers up, but what?

The answer lies in your work friction data. Look for moments and touchpoints that have a high impact score and a low satisfaction score.

These are the components of work that make a big difference to an employee’s ability to get their job done and that are not functioning well.

Maybe you discover two areas with high impact and low satisfaction:

3. The AI-to-agent handoff
4. Looking up customer information in the company database.

From text responses to the survey, you learn that agents don’t have enough time to read the AI’s summary before being connected to callers. This is stressful, in part because callers tend to get frustrated when they have to repeat their situation to the human agent.

You also learn that looking up customer information is frustrating because agents have to access multiple databases that aren’t always synced. Load times can lead to delays, which can add to callers’ frustration.

From here, you have two obvious levers to pull, one of which is easy and low risk. You increase the delay between handing a caller from the AI to a human agent; when you survey workers again a week later, they’re much more satisfied with that part of their work. What’s more, text comments note that having more time makes it easier to diffuse customer frustration.

The internal systems are still a pain point, but now you’ve bought yourself some time to figure out this bigger, longer-term issue.

In AI Transformations, Launch Is Only the Start

Launching an AI tool may feel like a culmination: you’ve done the research, done the training, perfected the tech setup, and then you’re live! And while go-live may be the end of the first phase of your AI journey, it’s only the start of the rest of it.

Introducing automation to a system as complex as an employee’s work inevitably changes it, often in unpredictable ways. To make sure you’re providing employees with the tools and resources they need to do their jobs and staying on track to enjoy the benefits you planned for from your AI investment, plan to measure work friction after launch.

Identifying places of high friction will help you know what the next best steps are as you proceed down the road to becoming an organization powered by AI.