Now availableVAL 2.0 — generally available
Rebuilt to be run by AI, not just used by you.
An AI agent inside VAL reads your documents, builds a dashboard or a small app on request, and can change your data directly — the things a person used to do, now done by AI.
Generally available now.
Not starting from zero — built and kept up to date for you, before you even start.
Canvas
Ask, and it's built.
Describe the app or dashboard you need. Canvas builds it live, drills as deep as you need, and every number traces back to real SQL — never a guess.
Canvas, at a glance
A live app, not a static report
Ask a question. Get an app.
An answer, a report, or a tool your team keeps using
Summary to source, in clicks
Click any order, see exactly why — real screen, demo data
An exact, traceable answer
What the platforms really charge, outlet by outlet
A few of the apps Canvas has built
Yesterday's sales, told like a morning paper
An FP&A canvas
Every Grab sale, matched to the bank
An AR canvas
Why invoices don't match — and what posts
An AP canvas
Tomorrow's kitchen order, worked out
A procurement canvas
And just as easily: screeners, benchmarks, break queues, brand monitors, quote pages, heatmaps — whatever shape the question needs.
The VAL agent
This is just the beginning.
What the VAL agent can do — right inside VAL, or from any agent you already use. It does the work; you stay in charge.
It builds the page
Ask for a page, not just an answer.
Ask for an outlet screener and VAL builds it as a page on your real numbers — saved, and ready to use again.
It reads your files
Drop in a file. It checks it against your data.
Drop in a supplier's price list. VAL reads every line and checks it against your contract prices, flagging what's above.
You approve
It proposes. Nothing changes until you say yes.
VAL lays out the change it wants to make — here, three items into Supplier issues — and waits for you to approve or reject.
It uses tools
Excel, code, a chart — when it needs one.
When the job needs a tool, VAL uses it: exports to Excel, runs code, draws a chart, then posts the write-up to your team.
Bring your own agent
Talk to VAL from your own AI harness.
A second way to reach the VAL agent — through Claude Code, or whichever AI harness your team already runs in. It builds, configures, and answers there too, drawing on the same library of skills for context and guidance, so power users get VAL supercharged into the tools they already live in.