What If Alexander the Great Taught Your Kids?

01 /The Concept

Humy helps teachers run more engaging lessons by letting them summon AI-reconstructed historical figures that students can actually talk to.

The platform already hosts over 2,000 AI personas of historical figures – from Socrates and Alexander the Great to Winston Churchill and Nelson Mandela.

To bring a historical figure into the classroom, a teacher first selects a purpose: either a) the AI character engages students in conversation about the lesson topic as a mentor, or b) it presents a specific challenge and guides students toward a solution. Students then receive individual links to chat with the character directly.

Beyond historical figures, Humy helps teachers build lesson plans, generate learning materials, and create assignments for in-class and homework.

The platform also automates feedback on a wide range of student assignments – multiple choice, essays, research projects, and more.

Pricing for individual teachers is $14.99 per month. Separate tiers exist for schools and districts that want to give access to all teachers, negotiated separately.

Over 50,000 teachers at schools and universities are already using Humy. The team published news of the latest platform update on Product Hunt a week ago. In May, the startup raised its first $150K seed round.

02 /Why It Matters

There's a teaching method named after the ancient Greek philosopher Socrates – the Socratic method – based on how Socrates himself worked with students.

The core idea: the teacher's job is to teach students how to think, not to hand them "correct" answers. Especially since in most real-world situations, there are no correct answers – only better and worse framings depending on the angle and the context.

This applies even to the hard sciences. Classical Euclidean geometry held for centuries that parallel lines never meet – then non-Euclidean geometries came along and upended that assumption, and they turned out to have important practical applications.

In social sciences and humanities – the domains Humy is built for – this ambiguity is even more pervasive. Life itself could be described as a social science: there are no right answers.

That's why the Socratic method's goal is to teach students to think – so they can decide for themselves what "right" looks like in a given situation, from their own perspective and from others'.

The mechanics: instead of delivering material from a single authoritative source, the teacher conducts a dialogue – asking provocative questions that push students to think, to take positions, and to imagine the situation from multiple perspectives.

The goal isn't for the student to arrive at the right answer. It's for them to recognize that many answers can be legitimate.

This is precisely what Humy enables. Bringing different historical figures into a lesson lets students see the same issue through radically different lenses – what would Churchill think was the right move here? What would Alexander the Great say? What would Nelson Mandela argue?

Creating AI personas of historical figures for conversation isn't a new idea. Dopple ([related review](/review/individualnost-cepljaet-i-prinosit)), for example, built a platform for creating AI versions of film and animation characters, game characters, and historical figures, raising $1.88M in its first round. But Dopple remains primarily a technology platform – it's not entirely clear why someone would return to it regularly.

Humy found a concrete and compelling use case – one where it's obvious who pays and why. Teachers who want to make their lessons more engaging and more thought-provoking are a clear, motivated customer base.

It's worth noting that pinning a specific historical figure's name to an AI persona is partly a compelling packaging decision. The AI has a distinct point of view and personality – the historical framing makes that more vivid and memorable.

SocialAI ([related review](/review/ne-igrushka-a-polza-i-dengi)), which raised $3M, decided to skip the historical framing entirely.

The startup built an app that functions as a "social network of one" – just the user and a cast of AI characters who comment on their posts. The user can choose their own AI subscribers with different personalities: troll, skeptic, optimist, fan, hater, advisor, contrarian, and so on.

The result is a private, interactive journal. The user writes thoughts, observations, and experiences – and a chorus of differently-minded characters responds, giving multiple perspectives on the same idea. The user can then engage those characters in dialogue, working to explain themselves more clearly and understand the other views – ultimately helping them form and articulate their own position.

03 /Opportunities

The broad opportunity is building platforms that give AI personas with distinct personalities and specific viewpoints a useful, recurring job.

Humy found one answer: using those personas to run lessons.

SocialAI found another: embedding group therapy or group mentoring dynamics inside an app – where different voices push the user to clarify and own their thinking.

Hyperbound ([related review](/review/za-takoe-obuchenie-kompanii-tochno-zaplatjat)) found a third: a sales training simulator, post-YC, that raised $1.5M. Salespeople practice conversations with AI personas of different roles and temperaments – "abrasive and will shut you down," "polite but immovable," and so on.

The pattern scales: any domain where expertise is scarce, expensive, or intimidating – financial coaching, legal guidance, language practice, negotiation training – is a candidate for AI personas with distinct specializations. The differentiator isn't the persona itself, it's the domain-specific knowledge base and the quality of the conversation engine beneath it. Which underserved domain would you tackle first?