AI agents for consumer research, product claims, and regulatory filings
Kolena AI agents read your documents, apply your review rules, and write the results into your systems.
The Kolena Advantage in Consumer Packaged Goods
Seven functions, one platform. Insights, regulatory, legal, quality, manufacturing, safety, and product testing each run their own agents on their own documents.
Document Workflows Kolena AI Agents Automate
Each starts from an expert-built agent, then adapts to your claim library, specs, and output templates. We size these against your own volumes during a pilot.
Consumer Study and Verbatim Analysis
Pre-built AI agent: Consumer Interview & Qualitative Study Analysis
Kolena reads agency debriefs, U&A studies, sensory panels, focus group transcripts, and thousands of verbatims, and codes them to your own attribute framework.
- Codes verbatims to your taxonomy of benefits, attributes, and purchase drivers.
- Compares results across waves, markets, and brands, and flags where a finding contradicts a prior study.
- Makes a decade of archived study findings searchable in plain language, with the source page cited.
Regulatory Submission Assembly
Kolena AI agents assemble the submission from formula records, supplier documentation, test reports, and safety data, then reformats it for each market's requirements.
- Checks INCI listings, allergen declarations, and concentration limits against current restricted-substance lists.
- Generates market-specific submission sets for US, EU, UK, ASEAN, and China from one source of truth.
- Re-runs the whole portfolio when a regulation changes and returns only the affected products.
- Checks ingredient lists, Nutrition Facts panels, and allergen declarations against current labeling requirements.
- Checks that on-pack nutrient and health claims trace to the nutrition analysis behind them, and flags where that data has changed.
- Re-runs the portfolio when a formulation or supplier changes and returns only the affected SKUs and label proofs.
- Checks every on-label efficacy claim against the approved EPA master label — pathogens, contact time, and approved surfaces.
- Assembles the FIFRA registration package from efficacy test reports, toxicological data, and environmental assessments.
- Reconciles SDS, GHS labeling, and hazardous-substance requirements, and flags labels that have drifted from their registration.
Claim Substantiation and Artwork Review
Pre-built AI agent: Product Claims & Concept Development
Every claim has to trace to a test result. Kolena reads clinical, instrumental, and consumer-perception reports and builds the substantiation record behind each one.
- Checks artwork proofs and campaign copy against the approved claim library, market by market.
- Flags unsupported superlatives, missing asterisk footnotes, and claims whose study has expired.
- Produces the substantiation package for a retailer review, self-regulatory body, or competitor challenge.
Raw Material COA and Spec Verification
COAs arrive from hundreds of suppliers in as many formats. Kolena AI agents map every result to your internal specification and clear or hold the lot before it reaches the floor.
- Compares assay, microbial, heavy metal, and impurity results to spec limits and flags out-of-range lots.
- Reconciles supplier spec sheets, SDS, allergen and origin statements against your approved material master.
- Tracks certificate expiry across palm, organic, halal, vegan, and eco-label programs.
Batch Record and Deviation Review
Filling and co-packing partners send batch documentation in their own formats. Kolena reads each packet against your GMP requirements and returns a release recommendation with the open items listed.
- Verifies in-process checks, yields, line clearance, and signature completeness per ISO 22716.
- Classifies deviations and CAPA records, then surfaces repeat root causes by site and by line.
- Prepares the audit file for a retailer audit, ISO 22716 certification, or internal quality review.
- Verifies in-process checks, yields, CCP monitoring records, and signature completeness against your HACCP plan and GFSI scheme.
- Classifies deviations and CAPA records, then surfaces repeat root causes by site and by line.
- Prepares the audit file for a GFSI audit, customer audit, or regulatory inspection.
- Verifies in-process checks, yields, line clearance, and signature completeness against your quality system and establishment requirements.
- Classifies deviations and CAPA records, then surfaces repeat root causes by site and by line.
- Prepares the audit file for an EPA establishment inspection, customer audit, or internal review.
Safety Incident Intake and Reporting
Pre-built AI agent: Consumer Complaint Analysis
Complaints arrive as free text from care centers, retailers, and health professionals. Kolena AI agents classify severity and causality against your own procedure and assemble the report.
- Separates serious adverse events from quality complaints and routes each to the right owner.
- Screens each case for reportability under MoCRA and EU cosmetovigilance rules, and links it to batch and market.
- Assembles the periodic safety report with every case traced to its source correspondence.
- Separates potential serious adverse health events from ordinary quality complaints against your own criteria.
- Links each case to lot, production date, and distribution while the trace is still actionable.
- Assembles the record behind a reportable-food determination and the supporting file for a recall classification decision.
- Separates reportable incidents from ordinary complaints against your own criteria.
- Classifies misuse against product defect, linking each case to lot, label version, and claim set.
- Assembles the incident file supporting a regulatory report or an internal safety review.
Study Data Reconciliation
Pre-built AI agent: Clinical / Efficacy Study Data Extraction
External institutes return raw measurement data and draft reports in their own formats. Kolena checks each file against its study report before it reaches your archive.
- Checks the returned data file against the study schedule and report, using your own error codes for measurement order, data types, and correlations.
- Extracts preliminary results as they arrive and routes them to the named recipients for that study.
- Answers project-level questions across the study archive: which studies ran, at which institute, under which study number.
Three separate teams run this workflow independently: clinical affairs, tolerability, and product support.
Four Sources of Return
Four sources of return, in the order manufacturers typically realize them. We build this model with your volumes, blended rates, and launch cycle times during the pilot.
Teams That Use Kolena
Kolena is deployed function by function. Each team keeps its own logic, templates, and approvals.
What CPG Teams Say
Personal care, food, and household manufacturers use Kolena AI agents to cut document turnaround from weeks to days.
"We had a decade of study reports nobody could search. Kolena coded the verbatims to our own attribute framework and gave us one comparable insight set, with every finding traced back to its source page."
"Registration packets for a new market used to take a full quarter of analyst time. Kolena builds the dossier from the formula record and the safety data, and my team reviews it instead of typing it."
"Every claim on the carton now has its study, its protocol, and its page reference attached. Legal review stopped being the reason artwork misses a print window."
Your Data Stays Yours
Kolena is built for manufacturers holding unreleased formulas, consumer data, and regulator scrutiny: reasoning logs on every output, full audit trails, and encryption throughout.
Start With One Workflow
Send fifty study reports, a launch submission, or a month of supplier COAs. We run them, show you the citations, and size the ROI against your volumes in two weeks.
No integration required. Start with the documents you already have.