This article was originally published in Fast Company.

It’s the best of times and the worst of times for people whose job it is to teach others how to do their own jobs better, adapt to change, and get the hang of new technologies. Organizations are seeing greater need than ever for learning and development (or “L&D” for short), with spending on training climbing nearly 33% to $90.6 billion in the U.S. last year.

At the same time, L&D professionals are rightly wondering whether they’ll soon be automated out of existence. After all, Google and YouTube are the de facto training departments for many employees: they’re ubiquitous, free, and packed with seemingly limitless content. But more sophisticated technology is on the rise, too. For example, AI can determine what someone needs to learn based on their performance data and career stage, then push content to them as they need it. This leaves many corporate trainers to stake their own value on curation, controlling the quality and consistency of training resources–but that’s already shrinking territory for L&D experts to stake their value on. Companies like Pathgather, Degreed, and Edcast are all developing tech-driven solutions to solve the curation piece of the puzzle.

So should we all reconcile ourselves to a future in which robots and algorithms help bring us up to speed as we advance our careers and adapt to change? Not necessarily. One area where the L&D professionals still have an advantage is in bridging the learning-doing gap. The science of learning suggests that the art of teaching is still very much a human endeavor. Here’s what L&D experts (and the organizations that employ them) will need to do not just to keep the robots at bay but to wind up with better-trained employees now and in the future.

Design and deliver brain-friendly content

The form content in which content is communicated is just as, if not more, important, than the information itself–and training experts can focus on determining form. For example, people pay special attention to stories; think of how easy it is to get lost in a movie or book. And stories are unusually powerful ways to create shared understanding–they get the brains of different people to align. Effective L&D professionals, in other words, are great storytellers. They’re also pros at tailoring instructional demos to different audiences, since we know that people also learn through observation.

Simple, clear goals also help the brain organize what it’s learning, so corporate trainers who can edit and condense instructional content may be more effective at delivering it than a YouTube video. The things people typically say they want in training resources–like high production value and gamification–don’t necessarily aid learning.

Connect content to personal goals and values

Want to motivate people to put what they’ve learned into practice? Then personalize the instruction. For example, a new manager might be studying how to have difficult conversations. Their personal goal is to get better at conveying information quickly without losing their team members’ trust. If the communications skills the manager is asked to learn relate directly to that goal, they’ll be more likely to absorb those techniques. One crucial new role for L&D professionals is to draw out those individualized connections, acting like coaches to guide people through training programs that keep them each personally motivated.

Create opportunities for deliberate practice

You can’t learn a new behavior without doing that behavior. The brain systems involved in learning new concepts, like the hippocampus, are great for recalling facts but totally distinct from those that learn actions, including the striatum–which learns through experience. In other words, you can’t reach every component of the brain with good arguments alone. Bridging the learning-doing gap requires flexing each of these brain systems simultaneously to make sure the desired habit actually sticks. A tech platform may be able to curate the right instructional information at the right time, but it still usually takes a human to devise real-world opportunities to put it into practice.

Reactivate the brain later

No matter how brain-friendly the content design, how motivated the person is, or how much a learner practices, the effort will likely fail if the learner engages with their new knowledge just once. The neural pathways representing new ideas and new behaviors have to be activated multiple times for long-term neuronal connections to grow and survive.

Research even suggests that our brains should retrieve newly learned information three times for the best results. What’s more, the learning process is enhanced by sleep–so those opportunities for memory recall should ideally happen on separate days. L&D experts can help craft these schedules, checking to make sure that the content of the training has a chance to lodge in employees’ minds for good.

The fact is that in order to adapt to the future of work, we’ll all have a lot to learn–not just once but continuously. And access to great content from robo-instructors won’t be enough. We’ll need to know how to close the gap between learning and doing. So far, the science of learning suggests that the ones best suited to help us solve that problem aren’t algorithms but other people.

This article was originally published in Fast Company.