“Hot honey is going to be big in snacks over the next six to twelve months. You should start working on it now.”
I remember having that conversation with a client at AI Palette.
They eventually launched. But it took them more than a year.
When we started AI Palette, I believed that helping companies identify consumer trends earlier would help them launch products faster. If we could compress months of research into a much shorter period, surely we could shorten the innovation cycle.
We could give them an earlier start. But there was a limit to how much that changed the finish date.
There were still concepts to validate, formulations to develop, ingredients to source, production trials to run and approvals to secure. An earlier consumer signal entered a process that largely worked as it had before.
That experience changed how I thought about the problem. Improving one step could only achieve so much when the outcome depended on how all the steps worked together.
It also pointed towards a broader change I believe FMCG companies need to make: from a process organised around departmental handoffs to one in which teams continually inform and adjust one another’s decisions.
The limits of the handoff
The traditional stage-gate process should look familiar.
Consumer research informs a brief. The brief moves into product development. Formulations are tested. Commercial assumptions are reviewed. Sourcing, manufacturing and regulatory requirements are resolved. The product moves towards launch.
In practice, these activities overlap. Experienced teams involve different functions early, and stage gates serve an important purpose. Companies need checkpoints to assess evidence, allocate capital and protect quality and safety.
But the work between those checkpoints can still behave like a relay race.
Each function advances the project before passing it on. Information travels with the project, and assumptions made in one department may only be tested properly when another department picks them up.
A promising concept encounters an ingredient cost problem. A formulation requires a production process that is difficult to scale. A launch plan assumes capacity that is already committed elsewhere.
The brief goes back. The formulation changes. Another review is scheduled.
Every team may be working hard. The product still moves slowly.
I remember a beverage company developing a ready-to-drink product for homemakers in India. The concept stood out, the consumer need was real, and the product performed well in consumer testing.
But producing it would require investment in new machinery. At the intended selling price, the economics did not leave enough margin to justify proceeding.
After two years, the company shelved the project.
Stopping may have been the right decision. The question is whether the consumer opportunity and the manufacturing economics could have shaped each other much earlier.
Could a different format have worked on existing equipment? Could the concept have supported a different price? Or could the team have reached the decision to stop sooner?
These are questions about how the business makes decisions together.
What a coordinated model looks like
In a more dynamic model, the brief becomes a shared set of assumptions that teams develop and test together.
Return to the hot honey opportunity.
Insights examines which consumers are interested, whether that interest is broadening and where the flavour might fit. R&D explores formulations. Procurement assesses ingredient options and costs. Manufacturing checks available formats and capacity. Regulatory reviews ingredients, labelling and proposed claims. Commercial evaluates price, channels and likely demand.
Early involvement matters. But the bigger change is what happens as those teams learn.
Imagine procurement discovers that the preferred ingredient would put the product above its target cost. That finding reaches R&D and commercial while formulation, pack size and price are still open decisions.
Manufacturing identifies an existing snack format that could accommodate the flavour without new equipment. The team revises the concept around that option. Insights then assesses whether the revised proposition still meets the consumer need that made the opportunity attractive.
Each function’s work changes what the others do next.
The same coordination continues after launch. Initial sales inform production volumes. Repeat purchase helps the team distinguish trial from sustained demand. Consumer feedback informs formulation changes. Production waste feeds back into the margin calculation.
A feedback loop is complete when new information influences a decision and the result of that decision generates further learning.
The operating model becomes dynamic because the teams remain connected as the evidence changes.
Stage gates can still provide formal checkpoints. Between those gates, teams need ways to identify consequential changes and respond while there is still room to act.
Why this matters beyond launch speed
Speed is one benefit of this model. Another is avoiding decisions that make sense within a function but fail when combined.
A consumer opportunity can look attractive until sourcing costs are understood. A cheaper ingredient can improve the purchase price while reducing consumer preference. A new variant can generate sales while adding enough production complexity to undermine its margin.
These tradeoffs exist throughout a product’s life.
Ingredient availability changes. Costs move. Consumer response differs from the original forecast. Production runs reveal issues that trials did not.
A plan approved several months earlier cannot anticipate every development. The company needs to recognise when its assumptions no longer hold and decide what to do about it.
For the beverage project, a coordinated approach might have helped the team find a viable production route. It might also have led to an earlier decision to stop.
Both would have been useful outcomes. The value lies in bringing consumer appeal, technical feasibility and commercial reality into the same decision.
What makes this more practical now
Cross-functional working has been an ambition for decades. Keeping it effective as information changes is much harder than arranging an early meeting.
Three capabilities can help: shared data, earlier visibility and AI.
Shared data gives teams a common basis for decisions.
Consumer insights, formulation records, supplier costs and production constraints need to be usable together. Teams need common definitions and a clear understanding of which information is current.
If commercial is working from one cost assumption and R&D from another, coordination breaks down even when both teams have access to a dashboard.
Earlier visibility allows teams to shape choices before they become commitments.
Procurement needs to see likely ingredient requirements before a formulation is nearly finished. Manufacturing needs to understand proposed formats and volumes before a launch date is promised.
That visibility should flow both ways. Available production capacity or a readily sourced ingredient can help innovation teams identify a more practical route to market. Operational knowledge can shape the opportunity from the beginning.
AI can help teams understand how a change affects the wider business.
The relevant information often sits across research reports, formulation records, supplier documents and production systems. Bringing it together takes time, and someone has to know what to look for.
Suppose an ingredient becomes more expensive. Which products are affected? What alternatives have already been tested? Would a substitution change taste, labelling, margin or launch timing?
AI can help retrieve the relevant knowledge, identify dependencies and assemble a view of the options. With reliable underlying information, it can reduce the effort required to keep different functions informed.
That makes ongoing coordination more practical. People still need to evaluate the evidence, resolve tradeoffs and own the decision.
The organisation has to change too
Technology can expose a problem quickly. The organisation still has to be able to respond.
If innovation is measured on launches, procurement on purchase price and manufacturing on long production runs, each function can meet its targets while making the overall opportunity harder to pursue.
Shared information needs to be accompanied by shared outcomes. Someone must be accountable for the combined result: consumer response, margin, speed and operational performance.
Decision authority matters just as much.
Who can reopen the brief when new evidence arrives? Who can approve a smaller production run? Who decides that a promising opportunity no longer justifies the investment?
Without clear answers, a connected system can produce more information for people waiting to make a decision.
Teams also need judgement about what deserves a response. A few social media posts should not overturn an approved plan. Losing a critical supplier or discovering that the product cannot meet its target margin may require immediate reassessment.
A dynamic operating model needs enough stability for people to execute and enough flexibility to respond to material changes.
For a leader, a useful starting point is one recent project. Trace the assumptions that each team worked with, when those assumptions changed and how long it took for the implications to reach everyone affected.
That will often reveal where better coordination could have changed the outcome.
When I started AI Palette, I focused on helping companies understand consumers earlier.
I still believe that matters. But seeing a change is only the beginning. The advantage depends on how quickly that understanding reaches the people who can act on it—and how their experience informs the next decision.
Scale has long been a major advantage in FMCG: bigger factories, wider distribution and larger budgets. Those advantages remain. But learning speed deserves a place alongside them.
A company learns when consumer feedback changes a formulation, when production experience reshapes a concept, or when new sourcing economics lead the team to reconsider a launch.
That requires information to move across departments and people to have the authority to respond. AI can help connect the evidence and assess the options. The organisation has to turn that learning into action.
The next advantage in FMCG will come from combining scale with the ability to adapt. That starts with teams learning together while there is still time to change the outcome.
Loading comments…