Kiddom: The Business of K-12 Curriculum & Learning Intelligence

Before you reason about Kiddom as a technology asset, understand what kind of company it actually is. Kiddom is not an "app" and not a content publisher in the traditional sense — it is a digital curriculum platform that fuses high-quality instructional materials with the software teachers use to plan, teach, assess, and analyze. The company's own framing has shifted over a decade from "teacher toolkit" to "digital curriculum platform" to, most recently, learning intelligence — and that arc is the single most important thing to internalize, because it defines the technical mandate you'd inherit.

The Founding Story

Kiddom was founded in 2015 in San Francisco by two college roommates from the University of Illinois who reconnected years after graduation. Ahsan Rizvi (CEO) had bootstrapped an iOS math game; Abbas Manjee (Chief Academic Officer) had been a math teacher for over-age, under-credited youth in New York City and had built a digital gradebook for his own students (after an earlier stint as an investment-banking analyst at Merrill Lynch). Those two artifacts — a learning app and a teacher's gradebook — are the genetic material of the company: a product instinct fused with genuine classroom credibility. The CAO being a former teacher (not a marketing title — an actual practitioner) is a cultural fact you'll feel in every product decision.

The stated mission is deliberately "human-first": technology that gives teachers time and judgment back, rather than replacing them. That framing matters strategically in the AI era — Kiddom's entire AI posture (Module 3) is "AI that supports the teacher's decision," not "AI that teaches the kid." Keep that north star; it's a positioning moat as much as a values statement.

The Strategic Arc: Gradebook → Curriculum → Learning Intelligence

Funding, Ownership & Valuation

Kiddom has raised roughly $56.5M across three disclosed rounds (some trackers cite higher cumulative figures; treat the exact total as an estimate). The defining round was the Series C of $35M in August 2021, led by Altos Ventures, with participation from edtech-focused investors including Owl Ventures and Khosla Ventures, plus Cambria Group, A-Street, and Edovate Capital across rounds. A frequently-cited post-Series-C valuation lands near ~$247M (2021 estimate — directional, not a fresh priced mark).

• Investor profile: Owl Ventures is the largest dedicated edtech fund in the world; their presence signals the company is being managed toward edtech-specific outcomes — district ARR, net revenue retention, and adoption wins — not generic SaaS vanity metrics.
• Capital posture: The Series C predates the 2022-2024 edtech funding winter and the ESSER funding cliff (the expiry of federal COVID relief money that inflated district budgets through ~2024). A company that raised in 2021 and is operating in 2026 has had to demonstrate capital efficiency and real revenue, not just growth — which shapes how you'll be asked to spend on engineering.
• No near-term IPO. Treat this as a private, milestone-driven company: patient-ish capital, but a board that wants to see the AI/learning-intelligence thesis convert into renewals and expansion.

Who Runs It

• Ahsan Rizvi — Co-Founder & CEO. Product/engineering origin; M.S. Public Policy, B.S. Industrial Engineering (Illinois).
• Abbas Manjee — Co-Founder & Chief Academic Officer. Former NYC math teacher; the academic-credibility anchor and the voice of "does this actually help teaching?" — your most important non-engineering partner.
• Kent Donges — Chief Revenue Officer; 30 years in education as teacher, principal, and sales leader. District go-to-market runs through here.
• Plus functional leadership across legal, sales, and customer success / professional learning (the implementation muscle that makes district adoptions stick).

Recurring public signal: Kiddom's co-founders have taken the stage at the ASU+GSV Summit (the industry's marquee edtech conference) on the theme of instructional coherence — the idea that fragmented tools and content hurt learning, and a single coherent curriculum-plus-platform is the fix. That word, "coherence," is the company's strategic vocabulary; you'll want it in yours.

You cannot architect for this business without understanding how K-12 actually buys — because the buyer, the budget cycle, and the adoption mechanism dictate the product's hard constraints (accessibility, interoperability, evidence of alignment) far more than any user-experience preference. K-12 is one of the most distinctive B2B markets in technology, and the people who run Kiddom live and breathe its rules.

The Buyer Is the District, Not the Teacher

The Kiddom of 2015 sold to teachers; the Kiddom that became a company sells to districts. That changes everything:

• Long, seasonal sales cycles. Decisions track the school-year and budget calendar. Pilots in spring, decisions in summer, deployment in fall. Revenue and load are both seasonal — and predictable.
• Procurement, not checkout. Purchases go through RFPs, committee review, board approval, and formal procurement. Your product must pass requirements (privacy, accessibility, interoperability) before it can be considered, regardless of how good it is.
• The decision-makers aren't the users. A superintendent, curriculum director, or school board chooses; teachers and students use. You build for two audiences with different success criteria.
• Budget cliffs are real. Federal ESSER relief funds (which inflated edtech spending 2021-2024) have expired. Districts in 2026 are scrutinizing spend and consolidating tools — a tailwind for "one coherent platform" and a headwind for point solutions.

The HQIM Movement — Kiddom's Tailwind

The single most important market force behind Kiddom is the rise of High-Quality Instructional Materials (HQIM). Over the last decade, a substantial body of research and advocacy (RAND, Chiefs for Change, and the nonprofit EdReports — effectively the "Consumer Reports of curriculum") established that the specific curriculum a school uses is a high-leverage, remarkably low-cost driver of student outcomes. Swapping a mediocre curriculum for a high-quality one is cheaper than almost any other intervention and can move achievement measurably.

This created a policy and purchasing shift that Kiddom is built to ride:

• States publish HQIM lists and run adoptions. State boards vet curricula against frameworks and publish approved lists; districts are steered (sometimes funded, sometimes required) to choose from them. Getting on the list is a distribution unlock.
• Decade-long lock-in. Adoptions are infrequent and durable. California's 2025 math adoption was its first since 2014 — the choices made now shape instruction for ~a decade. Winning an adoption is closer to winning a long-term contract than a sale.
• OER → supported digital HQIM. Much of the best curriculum (Illustrative Mathematics, EL Education, OpenSciEd) is openly licensed. The market problem was never "is good content available" — it's that raw open content is hard to implement, assess against, and keep coherent. Kiddom's wedge is turning open/HQIM content into a supported, digital, assessable, AI-augmented implementation.

Market Size & Shape

• The global EdTech market is frequently sized around $120-160B (2024) with aggressive multi-year growth projections — but these top-line numbers mix consumer, corporate, and K-12, so treat them as context, not Kiddom's addressable market.
• Kiddom's real TAM is narrower and more concrete: U.S. K-12 core instructional-materials + curriculum-platform spend — districts buying math, ELA, science, and social-studies programs plus the software to deliver them. This is a multi-billion-dollar, slow-moving, adoption-gated market dominated historically by a few large publishers (Module 5).
• The growth vector Kiddom is betting on is the convergence of two budgets that used to be separate: the curriculum/instructional-materials budget and the software/LMS budget. Kiddom sits at the seam and wants both.

"Kiddom" is really four products fused into one platform: a digital curriculum, a lightweight LMS / instruction layer, an assessment-and-analytics engine, and — increasingly — an AI layer. Districts buy a single coherent thing; engineering maintains four distinct subsystems with different SLAs. Treat this module as the functional spec of the business.

WHAT A DISTRICT BUYS WHEN IT BUYS "KIDDOM"

  ┌──────────────────────────────────────────────────────────────┐
  │  1. DIGITAL CURRICULUM  (the content)                         │
  │     IM v.360 math · EL Education ELA · OpenSciEd science      │
  │     Core Knowledge soc-studies · print + digital · EN / ES    │
  └───────────────────────────┬──────────────────────────────────┘
                              │ delivered & taught through...
  ┌───────────────────────────┴──────────────────────────────────┐
  │  2. INSTRUCTION / LMS LAYER                                   │
  │     lesson planning · assignments · gradebook · Spotlight     │
  │     (capture paper work → project to class) · rostering       │
  └───────────────────────────┬──────────────────────────────────┘
                              │ measured by...
  ┌───────────────────────────┴──────────────────────────────────┐
  │  3. ASSESSMENT & ANALYTICS                                    │
  │     embedded/formative assessment · standards-based grading   │
  │     teacher + admin dashboards · ANet formative integration   │
  └───────────────────────────┬──────────────────────────────────┘
                              │ augmented by...
  ┌───────────────────────────┴──────────────────────────────────┐
  │  4. AI LAYER  ("learning intelligence")                       │
  │     Kiddom Assistant · Auto-Feedback · Auto-Scoring ·         │
  │     Practice Generator · Lesson Clipper · ATLAS (differentiate)│
  └──────────────────────────────────────────────────────────────┘

1. The Curriculum

The content spine is a set of high-quality, mostly open or nonprofit-authored programs that Kiddom delivers digitally (and in print). The flagship is Kiddom IM® v.360 — Illustrative Mathematics' 360 curriculum, with state-specific variants (e.g., Kiddom IM v.360 California, NYC, Maryland, Virginia editions) aligned to each state's standards and framework. Content ships print + digital, in English and Spanish, with teacher guides, multilingual-learner supports, digital manipulatives, and project-based experiences. ELA comes via EL Education and Odell; science via OpenSciEd and OpenStax; social studies via Core Knowledge; plus phonics, formative assessment, and supplemental partners (Module 4).

2. The Instruction / LMS Layer

This is the teacher's daily surface: planning against the scope-and-sequence, assigning work, a standards-based gradebook, and classroom tools. A representative differentiator is Spotlight — capture paper-based student work from a device camera and project it to the class. That feature tells you something important about the product philosophy: Kiddom embraces the hybrid paper+digital reality of real classrooms rather than forcing everything online. Your architecture has to treat "a photo of a worksheet" as a first-class input, not an edge case.

3. Assessment & Analytics

Assessment is embedded in the curriculum, not bolted on — the point is to connect "what was taught," "how students did," and "what to do next." Standards-based grading rolls evidence up to mastery; teacher and admin dashboards turn it into instructional and reporting signal. Kiddom integrates formative assessment via ANet (Achievement Network). This subsystem is where Kiddom's "learning intelligence" claim is earned or lost — the data here is the fuel for the AI layer.

4. The AI Layer — "Learning Intelligence"

Since 2024 Kiddom has shipped a family of AI features, all sharing one design principle: grounded in the adopted high-quality curriculum, with the teacher in the loop. This is the "coherence" thesis applied to AI — the opposite of a generic chatbot that invents content from nowhere.

Atlas is the strategic centerpiece. Announced February 2026 (classroom availability Fall 2026, initially for Kiddom IM v.360 customers), Atlas runs a nightly batch over each class's formative assessment data and classwork, identifies misconceptions and gaps, and produces ready-to-teach differentiated materials and small-group instructions for the next day. CEO Ahsan Rizvi frames it as "a shift from reactive to proactive instruction," and Kiddom reports early in-school results of up to ~18% gains versus peers (treat as an early, company-reported figure, not an independent study). Atlas is built with assessment/PD partners (ANet, Teaching Lab).

This module covers the most distinctive — and most precarious — part of Kiddom's strategy: it builds a moat out of content it largely does not own. Getting this nuance exactly right will signal that you understand the business, not just the software.

The "Premium Digital Partner" Model

Kiddom partners with the nonprofit and open-curriculum authors whose programs the HQIM movement has anointed, and becomes their premium digital delivery + assessment + AI partner. The content's pedagogical authority comes from the partner (e.g., Illustrative Mathematics' brand is the credibility); Kiddom supplies the technology, the implementation support, and — increasingly — the AI and analytics layer on top.

State Adoptions Are the Distribution Engine

If partner curriculum is the content, state adoptions are the rights-equivalent distribution unlock — the closest thing in this business to Fanatics' exclusive league licenses. Getting a Kiddom program onto a state's approved HQIM list steers (and sometimes funds) district purchasing toward it for ~a decade. Recent wins:

• California — Math (Nov 2025): Kiddom IM® v.360 California approved by the State Board of Education for K-8 math + Algebra 1, in California's first math adoption cycle since 2014. Aligned to CA Common Core, the 2023 Math Framework, CA ELD standards, and Environmental Principles; English + Spanish; print + digital. This is a marquee, decade-shaping win.
• California — ELA/ELD (2026 cycle): Kiddom + EL Education submitted a comprehensive program for California's 2026 English-Language-Arts / English-Language-Development adoption (in process).
• Maryland — Math: Kiddom + Illustrative Mathematics launched a Maryland math curriculum for 2026-27 K-12 implementation.
• Virginia — Math: a Virginia math program aligned to the 2023 standards.

The Fault Line: A Moat Built on Borrowed Content

Kiddom sits at the collision of two large, previously-separate markets: the Learning Management System (LMS) market and the instructional-materials / curriculum-publisher market. Plus a fast-moving swarm of AI insurgents. Knowing who owns what — and what each can't do — is essential to any strategy conversation.

The Two Adjacent Markets Kiddom Bridges

The LMS axis (the "empty pipes"): Canvas (Instructure), Schoology (PowerSchool), and Google Classroom own the workflow layer — assignments, gradebook, distribution — but are content-agnostic. They're powerful pipes with nothing flowing through them; the school still has to source and assemble curriculum separately. Google Classroom in particular is near-ubiquitous (Chromebook + Google Workspace districts) and effectively free, which sets a brutal baseline — Kiddom even integrates with it rather than fighting it head-on.

The publisher axis (the "static content"): the legacy "big three" — HMH (Houghton Mifflin Harcourt), Savvas (ex-Pearson K-12), and McGraw Hill — plus modern curriculum specialists like Great Minds (Eureka Math), Amplify, and Imagine Learning. These own deep content and district relationships, but their platforms are often weaker, more closed, and less coherent than a software-native experience, and several are tied to a single curriculum they author.

Kiddom's wedge is the seam: curriculum + platform fused — a software-native experience wrapped around best-in-class HQIM. An LMS is pipes without content; a publisher is content without a great platform; Kiddom is trying to be both, with AI on top.

Competitor Map

Where the Battle Is Actually Fought

• Adoptions: getting on state HQIM lists (CA, TX, etc.) is the highest-leverage contest — a decade of distribution per win. Kiddom's CA math adoption is a real point on the board against far larger publishers.
• Coherence vs. fragmentation: Kiddom's pitch is "one integrated system" against both the LMS (no content) and the AI tools (no curriculum). This is its cleanest differentiation.
• The AI narrative: the insurgents (MagicSchool et al.) are winning mindshare on "AI for teachers." Kiddom's structural advantage is that its AI is grounded in adopted HQIM and real student data — but it has to make that advantage felt, or it loses the story to flashier ungrounded tools.
• Interoperability: because districts run Google Classroom/Canvas/PowerSchool, Kiddom frequently has to integrate with the incumbent rather than displace it — making LTI, rostering, and SSO competitive features, not plumbing.

This is the heart of the CTO conversation: what a curriculum-plus-platform-plus-AI business actually has to build and own. Below is a subsystem-by-subsystem reference architecture, each tied to a business force from earlier modules. These are general best-practice patterns for this domain (not Kiddom's confidential internals), framed so you can reason about trade-offs out loud.

KIDDOM — SYSTEM ARCHITECTURE (REFERENCE MODEL)

  CLIENTS:   Web App     iOS / Android     Print (PDF/offset)
               │             │                  ▲
               └──────┬──────┘                  │ single-source
                      │  (API gateway / BFF)    │ publishing
   ┌──────────────────┼─────────────────────────┴───────────────┐
   ▼                  ▼                                          ▼
 CURRICULUM       INSTRUCTION / LMS                      ASSESSMENT &
 CONTENT PLATFORM  • planning · assignments              ANALYTICS
 • canonical model • standards-based gradebook           • item bank (QTI)
   program→unit    • Spotlight (paper capture)           • auto + AI scoring
   →lesson→item    • submissions (digital + photo)       • mastery rollups
 • versioning      • classroom / real-time               • teacher + admin
   (CA/NYC/MD/VA)        │                                  dashboards
 • EN / ES i18n         │                                • district reporting
        │               │                                       │
        └──────┬────────┴──────────────┬────────────────────────┘
               ▼                        ▼
      STANDARDS ALIGNMENT GRAPH    AI / "LEARNING INTELLIGENCE"
      • CCSS · NGSS · state         • RAG grounded in HQIM corpus
        frameworks (CA/TX/VA)       • Auto-Feedback / Practice Gen
      • crosswalks · evidence       • ATLAS nightly differentiation
      • powers search, reporting,   • de-identify PII · human-in-loop
        adoption compliance         • evals / safety gates
               │                        │
               └───────────┬────────────┘
                           ▼
   INTEROPERABILITY  +  IDENTITY/ROSTERING  +  DATA WAREHOUSE
   OneRoster · LTI 1.3 · QTI · Caliper · Ed-Fi · Clever/ClassLink · SSO
                           │
                           ▼
   CROSS-CUTTING:  FERPA/COPPA/SOPIPA privacy · WCAG/508 a11y ·
                   SOC 2 · multi-tenancy · bell-schedule scale

1. Curriculum Content Platform (the foundation)

The hardest unglamorous problem — the analog of a marketplace's canonical catalog. A curriculum is a deep, structured hierarchy: subject → grade/course → unit → lesson → activity/task → item, each node carrying standards alignment, multiple renditions (print + digital), and language variants (EN/ES). You need a canonical content model and single-source publishing so that one authoritative source generates the web experience, the mobile experience, and the printed workbook/teacher guide — and so a state edition (CA/NYC/MD/VA) is a composable layer over the canonical core, not a fork. Get this wrong and every new state adoption multiplies your maintenance surface; get it right and adoptions become configuration. This is the single most important architectural decision in the company.

2. Standards Alignment Graph

Every piece of content aligns to one or more standards — and standards are a graph that differs by state and changes over time: Common Core, NGSS, plus state-specific frameworks (California's 2023 Math Framework + ELD + Environmental Principles, Virginia SOL, Texas TEKS). You need crosswalks between standard sets and machine-readable alignment evidence — because a state adoption literally requires proving alignment. This graph powers search, the standards-based gradebook, district reporting, and adoption compliance. It is also a genuine data moat: high-quality alignment data is expensive to produce and compounds.

3. Identity & Rostering

Who is in which class — sourced from each district's SIS (PowerSchool, Infinite Campus, Skyward) via Clever, ClassLink, or OneRoster syncs, with SSO (Google, Clever, Microsoft, SAML/OIDC). Multi-tenancy is hierarchical: district → school → teacher → section → student. This is unglamorous, breaks constantly, and is the number-one cause of failed September launches. Rostering reliability is a renewal issue disguised as plumbing.

4. Instruction / LMS Layer

Planning, assignments, the standards-based gradebook, and real-time classroom tools. Two domain-specific demands stand out: (a) paper-first inputs — Spotlight means "a photo of a worksheet" is a first-class object that must be captured, stored, and tied to a student/assignment; and (b) low-bandwidth / Chromebook tolerance — many classrooms have constrained connectivity and aging devices, so the client must degrade gracefully, not assume fat pipes.

5. Assessment & Analytics

An item bank (ideally QTI-modeled), scoring (rule-based and AI-assisted for open response), mastery/standards rollups, and dashboards for teachers and admins. This subsystem produces the longitudinal student-performance data that is the fuel for the AI layer and the basis of the "learning intelligence" claim — so its data model is strategic, not just operational. District-level reporting often needs to export via Ed-Fi or into the district's data warehouse.

6. The AI / "Learning Intelligence" Layer

The strategic core. The defining pattern is retrieval-augmented generation grounded in the HQIM corpus + standards graph + student data — the technical embodiment of "coherence." Auto-Feedback and Practice Generator are interactive; Atlas is a nightly batch pipeline that, per class, joins formative results to the curriculum and standards graph and generates differentiated next-day materials. The hard parts are not the model call — they are the substrate and the safety:

• Grounding & coherence: outputs must stay inside the adopted curriculum's scope and standards (retrieve, cite, constrain) — this is the differentiator vs. ungrounded AI tools.
• Student-data safety: minimize/de-identify PII before it reaches any model; never train third-party models on student data; keep a clear data-flow story for district privacy review.
• Human-in-the-loop: AI produces drafts (feedback, scores, warm-ups); the teacher approves. This is both pedagogy and risk control.
• Evals & safety gates: hallucination, bias, and age-appropriateness must be measured continuously — an AI-quality regression in a K-12 product is a trust and PR event, not just a bug. Build the measurement culture as deliberately as the features.
• Model strategy: foundation models are a buy (use frontier models via API — e.g., the latest Claude); the proprietary value is the grounding corpus, the prompts/pipelines, and the evals — not the base model.

7. Interoperability — the K-12 Integration Tax

Districts run a zoo of incumbent systems, so Kiddom must speak the ecosystem's standards fluently. These are revenue-blocking features, not back-office plumbing:

8. Compliance & Privacy (a P0 constraint)

Because users are children and customers are public agencies, compliance is a hard gate on revenue, akin to HIPAA in health systems:

• FERPA — protects education records; Kiddom operates as a "school official" handling them.
• COPPA — under-13 data; school-as-agent consent model and strict data-use limits.
• State student-privacy laws — e.g., California's SOPIPA, plus the national Student Data Privacy Agreement (NDPA) framework many districts require you to sign.
• Accessibility — WCAG 2.1 AA / Section 508 / ADA; adoptions frequently require a VPAT. Inaccessible content can disqualify you from a state list outright.
• SOC 2 Type II — table-stakes security attestation for district procurement.

9. Print / Publishing Pipeline

Unusually for a software company, Kiddom ships physical materials — workbooks and teacher guides, in English and Spanish — generated from the same single source as the digital experience. Print and digital must stay version-locked: a teacher's printed book and the student's screen must show the same lesson, same edition. This is a real pipeline (content → typeset → PDF → POD/offset partner) and a frequent source of drift if the content model isn't the single source of truth.

10. Scale & Reliability — the Shape of the Load

Edtech load is distinctively bursty and predictable, which is both a challenge and a gift:

• Bell-schedule concurrency: within a time zone, classes start on the hour/half-hour — a sawtooth of synchronized spikes as every period begins. Capacity must absorb the peak, not the average.
• The September spike: back-to-school is the single largest, most deterministic load event of the year — you know it's coming months out. There is no excuse for a September outage; load-test against it explicitly.
• School-year seasonality: summers are quiet (your maintenance/migration window); testing windows and term-ends create secondary peaks.
• Cost discipline: districts are price-sensitive and margins matter — over-provisioning for a peak you can predict is wasteful; autoscale to the known calendar.

A CTO's job is not to build everything — it's to know which capabilities are moat-building (build and own) and which are commodity (buy / partner and move on). Here is a defensible framework for Kiddom, followed by the highest-leverage strategic bets.

The Build / Buy / Partner Matrix

The Five Strategic Bets

1. Grounded AI as the moat ("learning intelligence"). The single biggest prize. Make the AI layer — grounded in adopted HQIM, the standards graph, and longitudinal student data, with rigorous evals and PII safety — so good that it converts "we use IM" into "we can't leave Kiddom." Atlas is the spearhead; the platform is the long game. This is the thing the ungrounded AI insurgents and the bare-content resellers structurally cannot copy.

2. The content + standards graph as a compounding data asset. Single-source publishing with state editions as composable layers turns each new adoption into configuration, not a fork — making the GTM engine (Module 4) scalable. The alignment graph + evidence is also a reusable asset across search, reporting, and compliance.

3. Interoperability as a competitive feature. Because districts keep their incumbent LMS/SIS, world-class rostering, LTI launch, and SSO let Kiddom live inside the district's stack instead of demanding rip-and-replace. Reliability here ("September just works") is a renewal driver.

4. Assessment-and-analytics depth. The data spine that makes "learning intelligence" real. Deeper formative assessment, mastery tracking, and admin reporting both serve districts directly and feed the AI flywheel.

5. Adoption-driven expansion + multilingual reach. Each state adoption (CA, MD, VA, and beyond) is a distribution unlock; English/Spanish and accessibility broaden every one. Decide deliberately how far to industrialize the "new state edition" machinery so the sales team can promise new states without an engineering scramble.

This module converts everything above into action: the metrics that matter, the risks to watch, the board/leadership narrative, a first-90-days checklist, and the questions you should be ready to answer (and ask) as you onboard.

The Metrics That Define Success

The Risk Register

• Partner-content dependence — the moat is the wrapper, not the content; other IM partners sell the same curriculum. (Module 4)
• AI safety & trust — a hallucinated/biased/age-inappropriate AI output in a K-12 product is a PR and trust event; evals + human-in-the-loop + PII de-identification are mandatory.
• Compliance/accessibility gate — a privacy or a11y failure can disqualify you from an adoption entirely.
• September reliability — the one deterministic mega-load event; an outage here is a renewal killer.
• Integration debt — content model + data spine + AI must be unified to make "learning intelligence" real.
• Budget cyclicality — post-ESSER scrutiny; the platform must justify spend via consolidation ("replace five tools") and outcomes.
• Mindshare loss to AI insurgents — MagicSchool/Brisk/etc. own the "AI for teachers" story bottom-up; grounded-AI advantage must be made visible.

The Leadership & Board Narrative

This board is edtech-native (Owl, Altos, Khosla) and the founders include a former teacher (CAO Abbas Manjee). They care about four things — translate every initiative into one of them:

• Growth / ARR & retention — adoptions won, district renewals, seat expansion, consolidation wins.
• Outcomes / efficacy — measured learning gains (Atlas's ~18% claim must become rigorous, defensible evidence) — this is the currency with educators and investors.
• Moat / defensibility — grounded AI on proprietary data, the content+standards graph, switching cost.
• Risk — privacy, accessibility, AI safety, September reliability.

Never present a roadmap in pure technology terms ("we're rebuilding the content service"). Present it as "this lets us ship a new state edition in weeks instead of months, which unlocks $X in adoptable pipeline, while making the AI grounding stronger." Speak outcomes + adoption + moat, not services.

Your First 90 Days

Days 1-30 — Listen & assess:

• Sit with a real district implementation end-to-end: rostering sync → teacher planning → assignment → assessment → gradebook → admin report. Document every manual step and every break point. Then watch a teacher use Spotlight and the AI features in a real (or recorded) class.
• Map the content architecture reality: how canonical is the content model? How are CA / NYC / MD / VA editions actually produced — composable layers or forks? How version-locked are print and digital?
• Audit the data spine: is assessment/performance data unified enough to fuel Atlas, or fragmented across subsystems? This determines how real "learning intelligence" can be.
• Review the AI stack: what's the grounding pipeline, what PII reaches models, what evals exist, what's the human-in-the-loop guarantee? Treat this as your top trust risk.
• Pull the compliance posture: SOC 2 status, VPAT/accessibility conformance, signed NDPAs, the privacy-review pipeline for new districts.
• Map the interoperability surface: rostering (Clever/ClassLink/OneRoster), LTI, SSO — and last September's rostering failure rate.

Days 31-60 — Diagnose & plan:

• Produce the integration-debt register and a sequenced unification plan (likely: content model → data spine → grounded AI). Align with the CAO (academic), CRO (adoption pipeline), and product.
• Stand up the metrics dashboard above; get a clean baseline before claiming any improvement.
• Pick the highest-leverage AI bet to harden (likely Atlas grounding + evals) and define what "measured efficacy" means rigorously, not anecdotally.
• Stress-test the September/bell-schedule load model and the rostering pipeline now — months ahead of back-to-school.

Days 61-90 — Commit & communicate:

• Present a phased roadmap framed in board terms (growth / outcomes / moat / risk), with the "ground the AI on a unified content + data substrate" thesis as the spine.
• Define the build/buy/partner framework (Module 7) explicitly, so every future decision references it.
• Ship one visible proof point — e.g., a new state edition produced as configuration rather than a fork, or an Atlas-grounding/eval improvement with a credible efficacy read.
• Establish rituals: a recurring AI-safety/eval review, a content-model + data-quality review, and a hard pre-September readiness gate (load + rostering + compliance).

Questions To Be Ready For — and To Ask

They may ask you:

• "How do you make our AI better than generic AI tools for teachers?" → grounding in adopted HQIM + standards graph + proprietary student data; evals; human-in-the-loop; coherence over freelance generation.
• "How do we add a new state adoption without an engineering fire drill?" → canonical content model with standards/localization as composable layers; single-source publishing; alignment evidence as data.
• "What's our biggest strategic technology risk?" → partner-content dependence + integration debt; the answer to both is the unified platform + grounded-AI moat.
• "How do you guarantee September works?" → deterministic load-test against the known peak, hardened rostering, a pre-September readiness gate.
• "How do you keep student data safe in an AI product?" → de-identify PII before models, no third-party training on student data, clear data-flow story for district privacy review, continuous safety evals.

You should ask them:

• How unified is the content model and the assessment/performance data today — and is "ground the AI on one substrate" actually the mandate?
• What's the real state of September reliability and rostering success across districts?
• How does Kiddom intend to defend against other IM partners and the AI insurgents — and what's engineering's role in that?
• What's the appetite (and budget) for the unification/consolidation work versus net-new feature velocity?
• How is success measured for this role in year one — adoptions enabled, AI/efficacy milestones, platform reliability, or all three?