Not many people have an assistant anymore. That’s why I’m so pumped to have a new helper — her name is Amy, and she is awesome. Amy only has a couple of years of work experience, but she is super friendly and responsive.

The main thing that Amy helps me with is scheduling meetings. Conference call ping-pong drives me crazy, so I just copy Amy when I need to set up a meeting, and she takes care of the rest. She knows my scheduling preferences better than anyone, and now that she has learned how I like to work, I trust her to communicate directly with both coworkers and clients. Amy is a valued coworker, just like the rest of my other 1,300 ISG colleagues.

Oh, and by the way, Amy is not human. She’s a bot from a company called X.ai. And I can’t imagine my day without her. I mention Amy here because she is a great example of how intelligent automation is working in companies today. An increasing number of people depend on bots to augment their tasks and help them make decisions.

“Augmentation” is the key word. Amy doesn’t replace the executive assistant role — she simply augments and, in some cases, replaces the tasks that executive assistants perform. By taking over the mundane work of setting up meetings, Amy frees up time for an administrative assistant (or me, in this case) to focus on more value-added work, which makes me more productive.

Our research reinforces this concept. Earlier this year, when ISG surveyed business and IT leaders, 68 percent of respondents either agreed or strongly agreed with the statement: “Our automation and AI initiatives are focused on automating tasks, not roles.” And nearly three-quarters agreed or strongly agreed with the statement: “Automation and AI will free up staff to do more value-added work.”

Task-based automation is happening across all enterprise support functions. In IT, virtual engineers are restarting web servers for systems engineers. In finance, RPA bots are setting up customer contracts for AP clerks, and in customer care centers, chatbots are partnering with agents to reduce social engineering.

Enterprises are focusing on task automation over role automation for two reasons:

First, it’s about the technology. True artificial general intelligence does not yet exist — and is not likely to for a long time. This means the intelligent automation technology we are applying today is narrow in its focus; it’s not (yet) a human. It’s really hard for intelligent automation tech to apply knowledge learned from one process to another process. This means each bot, virtual engineer, or algorithm focuses on accomplishing a set of narrow, specific tasks.

Second, it’s about specialization. Inside most companies, people perform several formal and informal roles. And those roles are made up of a lot of tasks. Large IT and BPO (business process outsourcing) providers do, however, have scale that allows them to specialize at the task level. This is one of the main reasons we’re starting to see layoff announcements. These layoffs are usually being positioned as performance-related, but in reality a lot of them are triggered by the introduction of intelligent automation. As intelligent automation takes over new tasks, people who are highly specialized need to elevate their skills. If they can’t, then they are let go for “performance” reasons.

So if you’re headed down a do-it-yourself path for your intelligent automation initiatives, my recommendation is to not create your business case around immediate cost savings because they likely won’t materialize in the near future (unlike savings from, say, outsourcing). Instead, focus on using your new digital workforce to increase productivity (by doing more with the same number of people), avoid costs (by not filling open job requisitions), improve speed (by having bots operate when humans typically don’t), and increase compliance (by logging everything the bots do).

If near-term cost savings is what you’re really looking for, a service provider can use their intelligent automation infrastructure to help you reduce cost, which, by the way, can be really significant, even for generation-three outsourcing buyers.

Stanton Jones is a Director and Principal Analyst at ISG, a global technology research and advisory firm.