Academic Shield - AI-Generated Paper Detection for Conferences

If an entrepreneur "adopted" this product today, here's the realistic math.

Fermi summary

If you land 75 conferences at $600/year that's $45k ARR - but GPTZero is free and program chairs have no budget, so expect negative take-home in year one while you're building the customer base.

Market size (TAM)

$4.0M

~4,000 active academic conferences with meaningful paper submission volumes × ~$1,000/year average subscription - constrained by tight academic budgets and many volunteer-run orgs with no software budget at all

Year-1 ARR range

$10k - $130k

midpoint $45k

Gross margin

82%

Investment to production

$28k

Dev: $12k for PDF ingestion, multi-model detection pipeline, and reporting dashboard. Integrations: $6k for EasyChair/HotCRP/OpenReview API

Probability of success

11%

P(reaching mid case in 12 months)

Expected take-home Y1

$-23930

probability-weighted, after investment

Go-to-market motion

Cold email to program chairs of top-tier CS/engineering conferences (NeurIPS, CVPR, AAAI tier-2 and below) offering a free trial on their next CFP cycle, then convert to annual subscription before the next call for papers opens.

Key risks

AI detection false-positive rate (~10-20%) means wrongly accusing a human author - a career-damaging accusation - exposing the conference AND your tool to serious blowback or legal action
GPTZero, Copyleaks, and Turnitin already offer academic/nonprofit pricing at $0-$150/year, making it nearly impossible to charge meaningful rates without a demonstrably superior accuracy story
Buying authority doesn't exist at the program-chair level - volunteer chairs must escalate to IEEE/ACM society boards or CS department procurement, stretching sales cycles to 12-18 months
OpenAI, Google, or a major academic publisher (Elsevier, Springer) could ship a native submission-level detector and make this category free overnight

Generated by the Wishdeal Factory financial-analysis agent. Numbers are honest Fermi estimates, not guarantees. Real outcomes depend on the operator. The studio is bullish on the engineering quality, agnostic on the business outcome.