Hey {{first_name}}, if you joined us from the webinar last week—welcome!

In last week’s webinar, we talked about how AI is helping us launch products more quickly by shifting from “what consumers say” to “what consumers actually do.”

Next week, I’m attending the Food Processing Expo in Sacramento. I’ve been attending more of these in-person conferences to get boots-on-the-ground understanding of what’s changing in the F&B space.

Planning to attend? Reply to this email and let’s meet up!

We're continuing our series on AI tools reshaping food and beverage product development.

Last week we covered Tastewise, which helps brands understand what consumers want before they build it. If you missed that one, catch up here.

Today we're looking at the other side of the equation: not what to build, but how to build it faster without burning through your R&D budget.

Most product failures start six months prior to production, when someone commits to an idea without knowing if it can survive cost pressures, sensory testing, and shelf stability all at once.

I've watched teams burn through trials for months, chasing a formula that sounded good on paper but fell apart at scale.

Here's what usually happens:

→ Formulate based on intuition or competitive reverse-engineering
→ Run sensory panels
→ Hit a cost ceiling or texture issue
→ Reformulate
→ Repeat until time or budget runs out

That cycle works when you have infinite resources

(Most of us don't)

This is when I discovered Proxy Foods.

Proxy Foods’ AI platform models the outcome space before you commit to iterations.

Instead of guessing which ingredient swaps might work, their AI platform models the outcome before you commit to iterations in the lab.

Think of it like this:

You're not replacing R&D judgment. You're giving it a preview of what's possible when you adjust sodium, increase protein, swap binders, or cut cost by 12%.

What Proxy actually does:

The platform unifies ingredient data, internal R&D knowledge, and predictive models into one workflow.

→ You set constraints, cost thresholds, sensory targets, nutrition specs, and it shows you which formulation paths are feasible before you waste weeks testing impossible combinations.

Where I see this mattering most:

→ Clean-label reformulations where you need to remove an ingredient without destroying texture
→ Alternative protein development where flavor and mouthfeel tradeoffs are brutal
→ Marginal-improvement projects
→ Regulatory-driven changes

The part that matters to me:

Technology like this doesn't replace the consumer’s palate.

It allows for faster iterations. It lets teams focus on refining what works instead of beating a dead horse (aka—a bad recipe).

What this looks like in practice:

A brand needs to reduce sugar by 30% without making the product taste like a science experiment. Instead of running 15 lab trials to find out which combination of sweeteners and bulking agents holds up, they model sensory similarity and cost impact in the platform first.

They run 3 focused trials instead of 15.
They protect margin.
They stay on timeline.

The truth is…

Food has to delight people. It has to feel familiar and trustworthy. No model replaces that human connection to flavor and experience.

But when you can compress months of trial-and-error into weeks of targeted refinement, you give better ideas a real chance to make it through the gauntlet of commercialization.

Where I think this goes:

Tools like Proxy pair well with market intelligence. When you know what consumers want and what's technically feasible at the same time, you stop building products that only work in theory.

You build products that survive the factory floor.
(And taste like the promise on the label)

If you're working on reformulation mandates, cost engineering, or innovation sprints that need to compress timelines without losing quality, this is worth understanding.

Not because it's flashy. Because it works.

Inspired to develop a new product? Let's chat!

BTW—did you miss the webinar?

Reply back to this email and I’ll send you the recording + slides.