The Right Answer Is the Least Useful Feedback You Can Give

01 /The Concept

Feedback that tells a student whether they got the right answer is the least useful kind. EdLight is built around a more radical premise: structured feedback that teaches students how to think – and uses a simple physical medium to make that thinking visible.

The platform's core insight is that all student work is submitted as digital images. The student doesn't need to use any software to prepare their submission – they can write or draw on paper and hold it up to a webcam or phone camera.

On that paper, a student might draw an annotated graph, a table of conclusions, or a formula with worked-out reasoning. Any format works – what matters is that the process of thinking is visible, not just the final answer.

Feedback on EdLight isn't a grade. It's commentary on how the student reached their answer and advice on how that thinking process could improve. Teaching someone to think is the goal; assigning a mark is a byproduct.

Teachers give feedback by adding annotations directly to the image the student submitted – sticky notes with drawings, text, or even voice messages, placed wherever relevant in the platform's editor.

All submitted work with teacher annotations accumulates on the platform as a learning portfolio – a visual, chronological record of how a student has progressed through each subject.

Reviewing that portfolio, a teacher can see the full arc of a student's thinking over time – identifying systemic gaps that need focused attention. Students can do the same for themselves. Reflection and self-analysis are core parts of learning, and the platform includes explicit guidance for teachers on how to help students use their portfolios for that purpose.

Portfolios can be shared with parents, tutors, and other teachers – making a child's education a collaborative effort grounded in a shared record. And crucially, that record isn't a grade sheet. It's a direct visualization of how the child thinks.

Teachers can use a limited version of the platform for free. The standard plan is $15/month or $144/year. Whole-school and district licensing is available under separately negotiated terms.

EdLight currently serves 20 US schools plus a number of individual teachers. Revenue tripled in 2023 compared to 2022, reaching $655K.

The startup had previously received only grants – $150K from Google for Startups, and $100K plus $3M from the Bill & Melinda Gates Foundation. It has now raised its first outside equity investment of $4M.

02 /Why It Matters

The number of startups helping teachers give better feedback is growing – driven by two converging forces: feedback has always been the most important part of learning, and AI now makes it feasible to deliver at a scale and frequency that were previously impossible.

Several startups in this space are worth noting.

Graide ([related review](/review/pomogi-im-masshtabirovatsja)) helps teachers automatically generate feedback for student work that resembles work they've already commented on. The insight: in any subject, students make the same mistakes repeatedly. Once a teacher has given feedback on a particular error pattern, the AI can replicate that feedback for every future instance of the same pattern. Graide has raised £2.5M.

Sizzle ([related review](/review/ii-vzorvjot-obrazovanie-sovsem-s-drugoj-storony)) built a homework help app that gives students step-by-step hints – nudging them toward the correct approach without handing them the answer. It raised $7.5M.

Packback ([related review](/review/jeto-mozhno-sdelat-ne-tupo-a-horosho)) gives students guidance on how to improve their writing and gives teachers AI-generated suggestions for evaluating student work using the same criteria used to coach the student while writing it. Its AI also moderates student discussion threads and provides automated feedback on messages. Packback has raised $12.2M.

Articles about EdLight describe it as an "AI platform," but the startup's site doesn't make explicit AI claims. It seems likely they'll add an AI assistant purpose-built for analyzing visual student responses and generating feedback in visual and audio formats – that's the natural next step.

The visual format of student work deserves its own attention as a concept. Adults routinely explain complex ideas on a napkin or whiteboard – sketching shapes, arrows, numbers, and rough diagrams to make ideas visible and communicable. It works because thinking is often spatial, not just verbal.

Yet we rarely teach children to think visually, or teach them to externalize their own reasoning that way. "Teaching on napkins" – making visible thinking the medium of instruction – is a compelling concept. EdLight has taken a decisive step in that direction.

Two more startups sit in the same current. Piggy ([covered here](/review/nezametnoe-izmenenie-na-mnogie-milliardy)) raised $7.7M for a mobile "stories" platform – vertical-scroll documents mixing images, video, and text, inspired by a student who submitted a TikTok in response to an essay prompt because that was how she actually thought. Antimatter took the same impulse further, building a platform where students write memes about what they're studying: you can only write a good joke about something you actually understand. It raised $2M.

03 /Opportunities

The thread running through this space is consistent: education's core goal is teaching people to think, not to memorize and repeat. That requires feedback on how a student arrived at an answer, not just whether it was right. AI can now support that feedback loop at a scale and frequency that would exhaust any individual teacher. And the formats students actually use to communicate outside the classroom – visual submissions, video, memes – belong inside it too.

The combined opportunity: educational platforms and tools that teach thinking, use modern communication formats to do it, and integrate AI for feedback at every stage of the process.

This space is large, and the startups mentioned here have only sketched a few of the available approaches. Somewhere at the intersection of these threads are opportunities worth billions. The task is to find where they converge.

The entry angle most underexplored: visual and audio feedback tools at the K-12 level, where AI adoption still lags far behind higher education.