For Nat Chidambaram · NETGEAR Enterprise
Every partner quote should leave with Insight and Exium attached.
NETGEAR has the hardware momentum. The growth is in the recurring line. We'd train a model on your winning partner quotes so every design is valid, all NETGEAR, and ships with the subscription.
- ✓Healthcare · 3 locations · staff, imaging and guest networks
- ✓WBE758 → PoE++ switch; PR30SG user and VLAN limits
- ✓Multi-site healthcare deals that closed carried a 3-year Insight term
Why now PR30SG started shipping with Exium in the box on 27 Oct, and Insight 10.0 is moving toward AI-defined networking. A model that knows your catalogue and your winning quotes is the piece that turns both into recurring revenue. NETGEAR Q3 2026 results (4 Nov) · Insight 10.0 (30 Jun)
See it in 60 seconds ↓01 · The moment
One partner brief. Two quotes.
A reseller asks for a NETGEAR design for three dental clinics. Play it with a general-purpose model, then with one trained on your catalogue and your winning quotes.
Synthetic customer and conversation, written from NETGEAR's public products. The pilot runs the same comparison on your real traffic.
02 · Tested on your workflow, not a benchmark
We grade every quote the way your sales engineers would.
The tests come from your catalogue, your compatibility rules and your won and lost quotes. A model only ships when it passes them.
Illustrative scores on synthetic quote requests written from NETGEAR's public catalogue, not measured on NETGEAR data. The pilot measures all four on 500 to 2,000 of your past quotes.
03 · The upside
What a higher attach rate is worth.
Move the sliders. The defaults are assumptions to replace with your numbers in week one.
Assumptions, not NETGEAR data. Default value per quote = 12 devices × $160, the Insight 3-year list price from 1 Aug 2026. Exium terms come on top.
04 · The 2-week pilot
Two weeks to prove attach rate on your own quotes.
We train on past partner quotes, test against your engineers' designs, and shadow live requests. No partner sees anything until you say so.
- Days 1–3Build the testPast quotes with won or lost outcomes, the catalogue and compatibility rules become an eval suite.
- Days 4–7TrainPost-train an open model on winning designs and score it against the best frontier model.
- Days 8–12Shadow live requestsThe model drafts quotes in the background; your sales engineers compare side by side.
- Days 13–14ReadoutAttach rate, design accuracy and deal size, plus a plan for a live test with APEX partners.
What we need from you
- 500–2,000 past partner quotes with won or lost outcome
- Product catalogue and compatibility rules
- 50–100 designs reviewed by your sales engineers
- Partner price book, under NDA
- One product owner, about 2 hours a week
What you get on day 14
- An eval suite for partner quoting that you keep
- The trained model weights, owned by NETGEAR
- A hosted endpoint ready for the partner portal or Insight
- Measured attach rate and design accuracy vs frontier models
- A plan for a live A/B test with APEX partners
Runs on past quotes and synthetic briefs only; no end-customer data is needed. Nothing reaches a partner until your team approves it.
Who we are
Dvar helps companies grow with AI models they own.
We build evals from your live workflows, improve an open model where it makes sense, and run it for you. Backed by General Catalyst, Picus Capital and Da Vinci. Team from IIT Delhi, IIT Bombay, Harvard and Cornell, with experience at Microsoft, McKinsey, BCG, Bain Capital, Nava and Google DeepMind.