Quote requests: specs, past prices and current material costs pulled into a first-pass estimate.
How it works
Four steps.
01
Map the business
A short discovery: how work, materials, money and information move through the business, and which systems hold the records.
02
Find and automate
The analyst reads the records, finds missed money and wasted hours, and automates the routine office work behind them.
03
Test bigger changes
Each proposed change to how you work is run against a fixed set of your past orders first. Only what comes out ahead reaches you.
04
Grow
Later, it looks for sales in your order history: reorders that slipped, products customers buy elsewhere, open capacity.
The innovation system
A swarm of AI agents that explores ideas for your business systematically.
Systematic inventive thinking gives people a small set of patterns for finding new solutions inside what a business already has. We adapted those patterns for many AI agents working in parallel. Together they map the space of possible changes, explore it far wider than a team could, and test each idea before anyone sees it.
Explore
Wide, by design
Agents apply structured idea patterns to every part of the business, not only the known problems.
Challenge
Argue against each idea
Other agents look for the reasons an idea would fail, and check it against your data.
Keep
Only what holds up
What survives reaches you with the evidence attached. The rest is dropped, and the system learns from it.
Two ways in
It looks for improvements on its own, and builds the automations you ask for.
On its own
It doesn't wait to be asked. The innovation system runs regularly across every part of the business and brings forward the changes that hold up.
Track record: the same system has won prizes on InnoCentive, where manufacturers and foundations pay only for solutions that pass expert review, including a Gates Foundation-backed agriculture challenge.
On request
Tell us the routine that eats your team's week. We write down what starts it, what it needs, what should come back and where you approve it. Then we build it on your data and add it to the analyst's skills.
The result: the analyst keeps every automation it learns, so it gets more useful each month.
Industries
Running now in converting plants.
Pilot open
Labels and packaging converters
Unbilled changes, repeats at stale prices, customer stock past its term, quote prep, tooling orders, press scheduling.
See the converting-plant page →
Next
Other industries
The approach works wherever a business runs on an ERP or job system, email and spreadsheets. If that describes yours, we'd like to hear about it.
Early pilot
We work closely with a small group of pilot partners.
This is an early pilot. We start with a short discovery and standard exports under a mutual NDA, and report what the analyst finds, what it automated and what it was worth, separately.
01
Starts from exports. Nothing to connect at first. Later, a read-only account.
02
You approve every action. Customer emails, commitments and operating changes.
03
Your data stays yours. Confidentiality terms agreed in the first call.
Team
Built by experienced entrepreneurs, industry veterans and leading scientists.
Yaron Oliker · Co-founder & CEO
Second-time founder with more than a decade building AI companies. Previously co-founded and led Unbotify, an AI fraud-detection company acquired by Adjust in 2019.