Google just threw a curveball at the ed-tech market. The search giant is embedding its Gemini AI models into Khan Academy's learning platform, starting with a new Writing Coach tool that launches today. Rather than replacing teachers, these tools are designed to guide students through the writing and reading process while giving educators better insights into which learners need help most.
Google and Khan Academy are reframing what AI can do in classrooms. At this week's Bett conference in London, the two organizations revealed they're combining forces to put Gemini's most capable models directly into Khan Academy's learning platform, starting today with the Writing Coach tool.
Here's what makes this different from the typical AI education hype cycle: instead of just generating answers, the Writing Coach actually guides students through the messy process of learning to write. The tool works in two modes - full interactive experience or feedback-only - and helps students outline, draft, and refine essays across persuasive, expository, and literary analysis formats. Right now it's available for grades 7-12, with beta access opening for grades 5-6.
"School district leaders are telling us that one of the biggest challenges they face right now is helping middle and high school students who are behind academically, especially in reading and language arts," said Sal Khan, founder and CEO of Khan Academy, in a statement. "We're proud to partner with Google to provide AI tools designed to improve reading and writing, enabling teachers to spend more time directly supporting the students who need their help the most."
This moves beyond the typical generative AI deployment in education. Ben Gomes, Google's Chief Technologist for Learning & Sustainability, grounded the partnership in learning science rather than just raw model capability. The Writing Coach adapts its feedback based on where individual students get stuck, offering specific examples to help them unstick themselves. Teachers get visibility into which students are struggling, but the system doesn't hand over finished work - it scaffolds the learning process.
The timing is significant. Education has been one of the slower sectors to adopt AI at scale, partly because educators remain skeptical about tools that simply automate thinking rather than enhance it. Recent research on AI in classrooms has shown that students learn better when AI acts as a tutor rather than a shortcut. Google and Khan Academy are betting that positioning AI as a learning coach rather than a homework machine will unlock deeper adoption.












