The first version of NoRedInk was not born in a pitch deck. It was hiding in a stack of high school essays. Jeff Scheur had spent eight years teaching English in Illinois, including at Chicago's Whitney Young Magnet High School, and the arithmetic of conscientious grading had become absurd. In one year he taught close to 170 students. At fifteen minutes a paper, a single assignment produced more than 42 hours of work. The red pen kept moving long after the weekend had given up.
The deeper frustration arrived when the papers went back. Students saw the grade, stared at the marks, and often had no clear next move. Some misplaced the paper. Some threw it away. Scheur had supplied a great quantity of correction but not necessarily a usable route to improvement. The distinction bothered him enough to start making copies of graded essays and cataloguing the misconceptions that appeared again and again.
His classroom identity extended beyond the papers. Scheur directed debate at Whitney Young and led the team to a national title in 2010. He earned National Board Certification in 2011. Both details complicate the familiar picture of a founder fleeing an old profession for a shinier one. He had become accomplished inside education before leaving the classroom, and he continued to define the new job in educational terms: broaden the reach of useful practice while respecting the people who knew the work firsthand.
His taxonomy became impressively particular. One code explained how to introduce a quotation with a colon rather than a comma. Another covered the present tense in literary analysis. He assembled thousands of rules into a writing manual. Still, he found himself circling the same comma errors and reading rewrites at two in the morning. The manual could describe the skill; it could not make timely practice appear beside every student.
The taxonomy before the technology
During his seventh year in the classroom, Scheur posted an advertisement on Craigslist. He hired an engineer for roughly ten to fifteen hours a week and paid out of his own pocket, eventually committing $20,000 to the early work. He had always been interested in computers and could mock up the central mechanics: generate lively sentences from a student's interests, notice which concepts caused trouble, and deliver practice that changed in response.
The order is worth noticing. He did not begin with adaptive learning as a fashionable technology and go searching for somewhere to put it. He began with a repeated classroom failure, labelled its component parts, and then asked software to perform the repetitive portion at a useful speed. Eight years inside the problem had given him a severe filter for nice-sounding ideas.
That filter also produced mundane features with unusual value. A teacher could preview an exercise exactly as a student would see it. For Scheur, the feature was obvious because a teacher must know what is about to land on a student's screen. In software, obviousness is often the reward for standing in the user's shoes long enough. No grand theory was required, only the memory of planning a lesson, anticipating confusion, and wanting to inspect the material before thirty teenagers did.
A student attempts the skill in context.
The misconception becomes specific.
Feedback arrives with another try.
Progress becomes visible to both people.
His students supplied the product's name. NoRedInk won a classroom vote over two alternatives that deserved, if nothing else, a dignified retirement: Grammar Ninja and Writing Beast. The choice did more than produce a memorable brand. It placed the students inside the founding story as collaborators, which is where Scheur had placed them all along.
Fifty teachers, then a small stampede
In February 2012, Scheur shared a finished early version with 50 teachers at an Illinois conference. Four weeks later, it had 1,500 users. By nine weeks, it had 15,000. Hundreds of schools asked for more material. By November, the number was approaching 100,000. Teachers were presenting it to their departments and recommending it to colleagues. The adoption loop looked uncannily like the learning loop: try something useful, see the result, pass it along.
NoRedInk's hook was simple and a little mischievous. Students supplied favorite celebrities, hobbies, fictional characters, and the names of friends. Exercises then used those interests in personalized sentences. Apostrophes and subject-verb agreement arrived wearing pop-culture costumes. One teacher reported overhearing a seventh grader announce, with the gravity of a rare scientific observation, that apostrophes were fun.
Scheur called this the “boring test.” A classroom activity could be accurate, aligned, and pedagogically respectable, but it still needed to hold a teenager's attention. The personalization was not confetti thrown over grammar. Motivation was part of the learning mechanism.
That September, NoRedInk won NBC's $75,000 Innovation Challenge. Scheur moved to California to join ImagineK12, the education technology accelerator later absorbed into Y Combinator. In 2013 the company raised $2 million from investors that included Google Ventures. The founder who had been paying a contractor from a teacher's salary could begin assembling a full team.
A teacher's argument against disruption
Scheur's classroom years did more than reveal a business opportunity. They also defined what the business should refuse to become. He has objected to education executives describing public education as completely broken. His preferred job for technology is less theatrical: help teachers and administrators do their work better.
That position draws a clean line around automation. Software can generate another targeted exercise without asking a teacher to circle the thousandth apostrophe. It can organize patterns across a class and let a student know immediately where an attempt went astray. It cannot provide the mentorship, inspiration, or relationship that Scheur considers essential to teaching. The computer handles repetition. The teacher keeps the human stakes.
The distinction helped NoRedInk expand from grammar practice into a broader writing curriculum: planning arguments, structuring essays, citing evidence, revising drafts, and cutting fluff. The premium product connected student work to teacher dashboards and district-level views. By 2021, students had completed more than 10 billion exercises. That August, NoRedInk raised a $50 million Series B led by Susquehanna Growth Equity, with True Ventures participating.
Scheur described the challenge of teaching writing with care: make the process less mysterious without flattening the art of expression. Targeted practice can make a difficult skill legible, but students also need to discover that a sentence has more than one defensible shape. A platform built from thousands of rules must leave room for judgment, voice, and surprise.
Feedback is valuable only when a learner can convert it into the next attempt. Volume matters less than timing, specificity, and a clear way back into practice.
The red pen meets the AI era
The technology surrounding writing has changed sharply since Scheur hired that first engineer. NoRedInk now offers AI-assisted grading, reading supports, guided drafting, and tools that help teachers judge how student text was produced. The company frames AI as an aid to feedback and personalization, while keeping critical thinking and student agency in view. Scheur's 2026 ASU+GSV appearance placed him in a conversation about the next step, from AI copilots to agents for educators.
Writing makes the boundary especially delicate. A grammar rule can be practised repeatedly, but expression cannot be reduced to a single approved sentence. Scheur has described the task as demystifying how strong writing works without diminishing its art. Scaffolds should make a hard move approachable, then loosen their grip. The aim is not to make every paragraph sound alike. It is to give more students command of the choices available to them, including the confidence to choose differently.
The new tools make his original boundary more relevant, not less. A machine can comment instantly and at scale. The important question is what the comment invites the student to do. If it completes the thinking, convenience wins and practice disappears. If it identifies a gap, offers a scaffold, and returns the pen, the feedback loop survives.
By 2026, students had completed more than 20 billion writing exercises on NoRedInk. Scale can make a founder's story feel inevitable, polished backward into a straight line. This one remains agreeably untidy. There was a manual too large to use, a Craigslist post, a $20,000 personal bet, an Illinois conference, and a naming contest won by the least embarrassing candidate.
Most of all, there was a teacher paying close attention to what happened after he handed back the paper. Scheur did not merely notice that grading consumed his nights. He noticed that all those hours could still fail the student at the precise moment feedback was supposed to help. NoRedInk grew from the decision to redesign that moment.
It is a founder's story with a teacher's moral: the work is not finished when the correction is delivered. The work is finished when the learner knows what to try next.