Skip to content

ML for formulation · built on your know-how

Faster R&D with ML
starts with your know-how.

Pick several competitor products, extract and validate their ingredients, and get an F0 starting formula shaped by your own formulation habits. Every sample and change feeds the next round.

Local training & inference · your data never leaves

Reuse know-how first, model by category next, validate in the lab.

Machine learning FAQ
Illustrative R&D scene linking formulation notes, cream trial samples and historical records Illustrative R&D scene
Your formulas · your materials · your labEvery change and result informs the next round
Benchmark many at onceIngredient extraction, cross-checked at scale
F0 that fits your labBuilt on your material, pairing and use-level habits
Know-how that compoundsSampling results return to your local know-how base

From know-how to formula iteration

Turn the formulas you already have
into reusable R&D know-how.

Connect formulas, raw materials, processes and lab results, and keep the reason behind every change. First make existing know-how searchable and reusable — then let validated machine learning narrow the next sampling round.

Stock photo of organizing existing R&D records
Start from real records and build up know-how.
01 / Organize records

Make know-how findable

Link formula versions, raw material grades, actual percentages, processes and results — and keep why something changed or was rejected.

Value now: less digging, more reuse
02 / Focus on a task

Teach models specific questions

Within one category and system, model use-level reference or a measured property, and validate on cases held out from training.

Prerequisite: comparable data, validated advice
03 / Validate & reuse

Give every iteration direction

Use validated references to screen candidates. Each new sampling round records changes and results, which join the know-how base after review.

Ongoing value: find the next direction faster
Local deployment decides where your data lives; what a model can do depends on the task, data quality and validation. Start by reusing know-how, then switch on validated predictions step by step.

Runs on your servers · R&D data protected

Machine learning stays local.
Your R&D data never leaves.

Real formulas, raw material compositions, batch ratios and lab records stay on your own servers. Models are trained and run locally, and your R&D know-how is managed by you.

Your on-premise environment
Local dataFormulas · lab
Local trainingYour patterns
Local inferenceSuggestions

AI organizes public information; your private formulas and lab data are never sent to external models.

01 / Why large language models

So much information —
the meaning hides in every sentence.

One review can praise "spreads easily at first" yet complain "still tacky two hours later." AI reads the context to separate experience, condition and wording — turning scattered comments into questions R&D can act on.

  • Less time reading and sortingKey points come first, so formulators focus on checking what matters.
  • Different wording, one theme"Sticky after applying" and "tacky feel" land in the same experience theme.
  • Every insight traces to a quoteConditions and sources are kept, so you can judge whether it applies.
From a paragraph to a clear question
Review quote · Example

"Spreads easily at first; still tacky two hours later, and pills under sunscreen."

↓ Read in context, keep the conditions
Spreadability"spreads easily" · at firstPositive
Tackiness"still tacky" · after 2 hoursTo improve
Pilling"pills" · under sunscreenTo improve

Extracted opinions need checking; consumer perceptions and lab measurements are stored separately.

Click to switch examples · Fictional examples, not customer data or live model output

Within your on-premise environment, formula and lab data are stored locally and machine learning trains and infers locally — no data leaves
On-premise deployment · organize data first, then build and validate models

02 / Local machine learning · formulation know-how

Machine learning stays local.
Your formulas never leave.

Formulas are your core asset. Model training and inference run entirely on your own servers, so your lab's know-how shapes every next R&D decision.

Never uploaded to external platforms or modelsFormula ratios, raw material compositions, processes and lab records all stay on your servers.
  • Learns your formulation habitsFinds the patterns in your past formulas — preferred materials, pairings and use levels.
  • Turns competitor references into F0Combines competitor ingredients and validated data into a starting point that fits your lab.
  • Every lab round adds upReviewed changes and results return locally for future search and model updates.

03 / From competitors to F0

From multi-competitor benchmarking
to your own F0 starting formula.

Pick several competitors, extract their ingredients, validate against large-scale data and learn how ingredients are combined. Local machine learning then adds your historical habits to propose an F0 starting formula that fits your lab.

Your formulation habits are learned locally · private formulas and lab data never leave

Click this step to see how the system builds an F0 starting formula from multiple competitors, or use the left and right arrow keys.

Competitor selection · multi-product benchmark

Pick several competitors
and compare them side by side.

Choose comparable creams as the target set, and set this round's direction by positioning and sensory experience.

Why it helpsSee shared traits and differences in one place — no collecting and switching one by one.
Select competitorsSeveral different competitor creams selected together in the system and grouped into one comparison
Multiple productsOne comparisonTarget direction
After F0, know-how keeps building
Sample & testLog changes & resultsReview & save locallyFeed the next F0

Click any step to view · Illustrative scenes, not the real interface or customer data; F0 suggested percentages still require lab validation.

04 / Machine learning for formulation · FAQ

Your formulation questions,
answered up front.

From data and security to F0 sampling — how your existing formulation know-how becomes the basis for the next round.

From your own know-how to faster iteration.Organize real records first, then learn and validate by category. Machine learning helps find the direction, and every lab run adds to your know-how.
Machine learning runs locally — R&D data never leavesFormulas, raw material compositions and lab data stay on your servers, and model training and inference run there too.

INCILABS product notes · Updated

What is machine learning–assisted formulation?

Machine learning–assisted formulation learns patterns from your historical formulas, actual batch percentages, raw material compositions, processes and lab results to support material selection, use-level estimation and formula iteration. INCILABS combines it with multi-competitor benchmarking to deliver an F0 starting formula for formulators to review and sample.

Does machine learning send our formulas or lab data outside?

No. INCILABS machine learning trains and infers entirely on your own servers. Private formulas, raw material compositions, batch ratios and lab data are never sent to external models or services. Local know-how is governed by your own permissions, and later samples and adjustments keep flowing back locally.

How do multiple competitor products become an F0 starting formula?

Select several comparable competitor products, extract their published ingredient lists and cross-check them against INCILABS ingredient, raw material and filing data. Local machine learning then applies your learned habits in material choice, pairing and use levels to propose a raw material list, suggested percentages and supporting evidence. After formulator review and sampling, changes and results are written back.

What formula data do we need? Can we use Excel records?

Yes — you can start from Excel formulas and lab records. Prioritize linking formula versions, raw material grades, INCI composition, actual percentages, process conditions and the matching test results. Keep the reasons behind both adopted and rejected options, flag missing fields clearly, and build up comparable R&D cases over time.

We don't have much formula data. Can we still start?

Yes. Start with one category you make often: organize its records, search similar formulas and reuse past know-how. Whether predictions can be enabled depends on the task, the number of independent cases, data quality and validation results — near-identical versions don't count as many independent samples. Validate one specific question first, then expand.

Does local machine learning mean training a large model from scratch?

No. You can build task-specific models for use-level reference, candidate ranking or a particular measured property, without pre-training a general-purpose large model. Local deployment defines where training and inference run; what a model can do still has to be validated on your own data.

How do ingredient order, regulatory limits and VCRP help estimate percentages?

Ingredient order, applicable regulatory limits, VCRP historical use data and other use-level surveys, similar historical formulas, and each raw material's INCI content and recommended use level together narrow the candidate range. Every source keeps its origin and conditions of use. VCRP is one reference among several — a published ingredient list cannot reconstruct a competitor's actual formula.

Can an F0 starting formula go straight to production?

No. F0 is a starting point for formulator review and sampling. Suggested percentages must go through actual sampling, performance and stability testing and regulatory checks, then be approved through your own process. Changes and test results feed the next round of local learning.

How do we know machine learning actually speeds up formulation?

Within the same category and task, compare against your existing manual process: time spent searching information, time reviewing candidates, sampling rounds needed to hit the target, and validation pass rates. Test on new cases not used in training and record where it applies; the actual speed-up depends on your own validation results.

Make every formula change drive the next iteration.

Start from the formulas and lab work you already have: capture the know-how first, then build validated local R&D capability step by step.