🔒 Password Required

This presentation is password protected.

VeriMath Cover
VeriMath Workflow
ESL × Wolfram

VeriMath

Natural Language → Verified Mathematics

A New Revenue Channel for Wolfram Research

ESL — Engineering Software Lab Wolfram Technology Partner, Israel

Prepared for Abel Nazareth, International Business Development Manager

The Problem

Researchers are choosing ChatGPT over Mathematica for mathematical work.

🔴 ChatGPT

  • Type English → get answer instantly
  • Zero learning curve
  • $20/month subscription
  • 300M+ weekly active users

But: answers are often WRONG

◆ Mathematica

  • Learn Wolfram Language first
  • Write code → get answer
  • Steep learning curve
  • ~303 institutional customers

But: answers are ALWAYS correct

Result: Wolfram gets $0 from these users.

Real Customer, Real Problem

Prof. David Gilo — Tel Aviv University, Faculty of Law

"If there were software where you could simply type in English or upload a file and start asking mathematical questions, and it gives accurate answers because it uses Mathematica and translates to plain English — that would be a breakthrough." — Prof. David Gilo, after testing both ChatGPT and our prototype

ChatGPT (what he used first)

  • Wrong answer: γ_crit ≈ 5.9 (incorrect)
  • Found OLD paper online, used wrong formulas
  • 5+ attempts, all with errors
  • Hours of manual correction

Our Prototype (what we built)

  • Correct result in single prompt
  • Interactive .nb with 3 Manipulate panels
  • Self-consistency check caught the real issue
  • Minutes, not hours

The Solution: VeriMath

ChatGPT's natural language interface + Wolfram's computational engine

📄 Upload PDF
→
💬 Ask in English
→
◆ Wolfram Computes
→
✅ Verified Answer

📝 Input

"Verify Lemma 1 in this paper. Check if it applies to CRRA utility. Show interactive plots."

◆ Under the Hood

AI generates Wolfram Language code → Wolfram Engine executes FindRoot, Solve, NIntegrate → Results verified

📊 Output

Plain English explanation + interactive plots + downloadable .nb notebook + step-by-step proof

How VeriMath Grows Wolfram's Revenue

💰 Direct Revenue

Engine
Licenses
  • VeriMath runs on Wolfram Engine
  • ESL pays Wolfram per server
  • Scales with user base
  • Net-new revenue from $0 users

📈 Upsell Pipeline

Mathematica
Licenses
  • Every session → downloadable .nb
  • To modify .nb → need Mathematica
  • ESL sells as Wolfram partner
  • "Gateway" to full Mathematica

🌍 Market Expansion

New
Users
  • Law, economics, social sciences
  • Users who will never learn WL
  • Graduate students
  • Industry researchers

The Conversion Funnel

Every layer generates revenue for Wolfram

ChatGPT Users (300M+)
Pay Wolfram: $0
▼
VeriMath Users
Wolfram gets: Engine license revenue
$$
▼
Download .nb Files
Discover Wolfram Language
▼
Buy Mathematica License
ESL sells as Wolfram partner
$$$

Revenue Projections — Wolfram's Share

Year 1 Year 2 Year 3
VeriMath Users 50–100 500–1,000 2,000–5,000
Engine Licenses 1 3–5 5–10
Engine Revenue to Wolfram $5K–$15K $15K–$50K $25K–$75K
Mathematica Upsell (licenses) 5–10 20–50 50–100
University Site Licenses — 2–3 5–10
Total New Revenue to Wolfram $10K–$30K $50K–$150K $150K–$400K

All from users who currently pay Wolfram $0. Zero investment required from Wolfram — ESL builds and operates VeriMath.

Go-to-Market Strategy

Phase 1
Months 1–6
Israeli Universities
TAU (reference customer)
Hebrew University
Technion
Free pilot programs
Phase 2
Months 6–12
International Academic
US universities (175 customers)
UK (17), Germany (11)
India (Abel's network)
Existing Wolfram licensees
Phase 3
Months 12–18
Industry R&D
Pharma (clinical trials)
Finance (quant research)
Engineering (verification)
Government labs

Abel's international network is the distribution channel for Phase 2.

Competitive Defense

If Wolfram doesn't offer this, someone will build it on:

🐍 SymPy (Free)

Open-source Python CAS. Growing fast. If someone builds "English → SymPy" — Wolfram loses the entire non-programmer market.

📐 MATLAB / MathWorks

Already has large academic install base. MathWorks + AI = direct threat to Wolfram's education revenue.

🍁 Maple (Maplesoft)

Strong in symbolic computation. If Maple builds a natural language interface first, Wolfram loses mindshare.

✅ VeriMath = Wolfram's Answer

Ensures the "English → Math" bridge runs on Wolfram Engine. Built by your existing partner. No investment needed from Wolfram.

What ESL Needs from Wolfram

✅ Required

  • Wolfram Engine commercial license at partner pricing
  • Written approval to build product on Engine
  • "Powered by Wolfram" branding rights

🤝 Nice to Have

  • Co-branded case study (Prof. Gilo, TAU)
  • Referral pipeline ("want English math?" → VeriMath)
  • Mention in Wolfram partner directory
  • Conference co-presentation

What Wolfram Gets Without Doing Anything

Engine license revenue + Mathematica upsell pipeline + market expansion into non-programmers + competitive defense against SymPy/MATLAB/Maple.

Zero investment. Zero risk. ESL builds and operates everything.

The Ask

1
30-minute live demo
We'll analyze a research paper in real-time. You'll see Wolfram Engine computing under the hood.
2
Engine license agreement
At partner rates. Starting with 1 instance, scaling as needed.
3
Approval to launch
"Powered by Wolfram" branding. Standard technology partner terms.
4
Optional: Joint go-to-market
Co-branded case study. Referral pipeline. Conference co-presentation.

The Bottom Line

$0
What Wolfram gets from
ChatGPT math users today
→
$$$
What Wolfram gets when
VeriMath redirects them
Every VeriMath user is a user who would otherwise give their money to OpenAI. We redirect that spend through Wolfram's ecosystem — via your existing Israeli partner.

The prototype is built. The first customer is waiting.
Let's schedule the demo.

Daniel Liezrowice — ESL Wolfram Technology Partner, Israel
1 / 14