• AI visibility

Guide · AI search

When a formulator asks AI who supplies a chemical, does our company come up?

It can, when an assistant can find and check your grades, specifications, applications, regulatory status and where you ship. Chemical buyers are technical, cautious and slow to switch, so AI search rarely closes a sale. What it can do is decide which suppliers a formulator or buyer requests samples and documents from first.

By Underneath Team Guide Updated

This guide is for chemical producers and distributors that sell to industrial customers: specialty and basic chemicals, ingredients, additives, coatings raw materials and water treatment products. It covers legitimate commercial discovery only. It is not safety, regulatory or legal advice.

The short version

  1. Chemicals lead industrial sourcing: on Thomasnet, chemicals were the most sourced product category of 2025 (opens in a new tab), led by specialty formulations, coatings and cleaning compounds, and buyers increasingly evaluated alternative and secondary suppliers.
  2. AI is now routine inside chemical companies: in a Journal of Business Chemistry survey (opens in a new tab) of 124 German chemical and pharmaceutical companies, active AI use rose from 34% in 2020 to 76% in 2025, with generative AI tools such as ChatGPT part of daily work.
  3. Industrial customers dominate demand: more than 80% of basic and specialty chemicals are consumed by the industrial sector, the American Chemistry Council said in SCI’s 2026 outlook (opens in a new tab), and US output rose only 0.7% in 2025, so growth often means winning share from other suppliers.
  4. Buyers want technically capable partners: specialty chemical manufacturers in L.E.K. Consulting’s 2026 survey (opens in a new tab) favored end-market-specific and regional distributors that offer technical knowledge and formulation support.
  5. Regulation is part of every search: the TSCA Inventory (opens in a new tab) lists more than 86,000 chemicals, and the EU’s REACH rules (opens in a new tab) require registration of substances above 1 tonne per year per company, so buyers ask about regulatory status as often as about price.

Who buys industrial chemicals, and how do they decide?

Formulators, engineers and procurement teams at manufacturers, who qualify a supplier slowly and then buy for years.

Several people at the customer judge a new chemical supplier. An R&D chemist or formulator chooses the ingredient and grade. A process or plant engineer cares about handling, compatibility and consistency. A procurement or category manager negotiates price, supply security and terms, and a regulatory or environmental, health and safety specialist checks documents. Across business purchases generally, Forrester (opens in a new tab) finds procurement professionals are decision-makers in 53% of buying cycles.

The decision follows a familiar path. A buyer identifies a need, such as a new product, a reformulation, a supply problem or a tariff change. They research candidate suppliers, request a technical data sheet and a safety data sheet (SDS), order samples, run lab and plant trials, and only then add a supplier to the approved list. Once a material is specified into a formulation, switching means requalifying, so a supplier that wins a spec can keep the volume for years. Building product makers run a similar race, covered in our guide to how architects find materials through AI.

The market is under pressure, which raises the stakes of each new account. The ACC expected growth in only nine of the 20 key chemistry end-use industries it tracks in 2025, rising to 12 in 2026. Global share has shifted too: according to the European chemical industry group Cefic, quoted by SCI, China now accounts for 46% of global chemical sales. Distributors feel the same squeeze: Brenntag (opens in a new tab), the large chemical distributor, reported 2025 sales of EUR 15.2 billion, down 3.7%. In a flat market, the supplier that gets into more early comparisons gets more chances to win.

There is no public benchmark for what a typical chemical account is worth. Our inference: because qualification is slow and switching is costly, one spec-in can be worth far more than the first order suggests.

Where does AI search already sit in chemical buying?

In early research and comparison, used widely by technical buyers but checked against documents and supplier sites.

Evidence on chemical buyers specifically is thin, so we separate what is measured from what is inferred.

  • Chemical companies use AI at work. The German survey found 91% of companies rated digitalization relevant or very relevant in 2025, and active AI use more than doubled to 76%. It describes generative AI tools, such as ChatGPT and company assistants, as increasingly embedded in daily operations. It does not measure supplier searches.
  • Technical buyers use generative AI in purchasing. The 2026 State of Marketing to Engineers research (opens in a new tab) by GlobalSpec and TREW Marketing, which surveyed engineers across industries, found 69% of technical buyers use generative AI during the purchasing process, but rate their trust in its answers at 4.7 out of 10.
  • Independent sources are gaining weight. In the same survey, online technical publications (76%) edged out supplier and vendor websites (74%) as the place technical buyers routinely research purchases.
  • Specialty chemical firms are going digital. L.E.K. reports significant increases since 2022 in the use of AI, generative AI and lead generation software among specialty chemical companies.
  • Buyers still want people. Distributor ChemPoint’s 2025 industry survey (opens in a new tab) of over 150 chemical industry professionals frames its findings around why AI has not replaced the human touch in B2B buying.

Our reading: chemists and buyers use AI to find candidate suppliers and grades faster, then verify everything against technical data sheets, SDSs and trusted publications before they contact anyone.

Which supplier questions do chemical buyers put to AI?

Questions that join an application, a grade, a regulatory status, a pack size and a region.

We wrote the examples below for illustration; they are not records of real buyer prompts.

What the buyer needsExample question (our wording)
Reformulation“Alternatives to fluorinated surfactants for an industrial floor coating, with US supply”
Food and personal care grades“Food-grade sodium citrate distributors in Texas, tote quantities, kosher certified”
Sustainable inputs“Bio-based plasticizers compatible with flexible PVC that are registered under REACH”
Documentation“Where can I get the technical data sheet and SDS for this grade?”
Local supply“Water treatment chemical distributors in the Midwest with bulk delivery”
Comparison“Compare these two distributors on technical support, lead time and minimum order”

Each question works as a filter. If your product pages do not state the grade, the typical specifications, the applications, the certifications, the pack sizes and the regions you serve, an assistant has nothing to match. If your documents sit behind a login with no public summary, a buyer may never learn you carry the grade.

Reformulation questions deserve special attention. L.E.K. found roughly half of specialty chemical companies have reformulated or phased out products containing chemicals of concern in the past three years, with a similar share expecting to continue. Every reformulation starts with a search for an alternative ingredient and a supplier who can document it.

How does an AI answer turn into a sample request and a contract?

Through a candidate list, a document check, samples, qualification trials and, finally, a place on the approved supplier list.

  1. Candidate list. A formulator or buyer asks an assistant, a directory such as Thomasnet or a search engine for suppliers of an ingredient for an application.
  2. Document check. They look for the technical data sheet, the SDS, regulatory status and certifications on supplier websites and trusted publications.
  3. Sample request. The supplier receives a request for samples, usually alongside other suppliers. This is the first moment the supplier sees the opportunity.
  4. Qualification. Lab work, plant trials, quality audits and customer approvals follow, often over months.
  5. Spec-in and supply. The approved material goes into production, with repeat orders as long as quality, supply and price hold.

An assistant’s answer reaches only the candidate list and the document check. Product performance, documentation quality, technical service and price decide the rest. A supplier that is missing from the candidate list never gets to prove any of those. Component makers face the same early cut when engineers ask AI for a part.

What makes an assistant name one chemical supplier over another?

Information it can find and corroborate; the platforms document how they search, not how they rank suppliers.

Documented by the platforms. OpenAI says (opens in a new tab) ChatGPT search typically rewrites a question into one or more targeted queries that it sends to search providers, and that sites should allow OpenAI’s search crawler, OAI-SearchBot, to be eligible for inclusion. Google says (opens in a new tab) AI Mode issues multiple related searches across subtopics and data sources, a technique it calls “query fan-out.”

Observed in our studies, which covered buyer questions in several industries but not chemicals:

  • ChatGPT ran a mean of 3.7 searches per buyer question before answering (hidden searches study).
  • Each tenfold increase in independent sites naming a brand in the cited pages went with 4.7 times the odds of being recommended (brand entity study).
  • Of 269 numbered “best” lists cited by AI surfaces, 24.2% ranked their own publisher first (self-promoting lists study), so a distributor’s own “top suppliers” page is unlikely to be the only list an assistant reads.
  • Only 25.2% of the brands ChatGPT named for a question appeared in all five repeats (consistency study).

Our inference for chemicals. The trust factors are the ones a regulatory specialist already checks: product identity with CAS numbers, grade and typical properties, applications, regulatory status by market, certifications such as ISO 9001, kosher, halal or food-safety schemes with the issuing body named, packaging sizes, plants and warehouses, and up-to-date SDS access. Coverage in trade publications, technical papers, association listings and manufacturer line cards gives an assistant independent confirmation.

Why do safety data and regulatory facts matter so much in AI answers?

Because buyers filter suppliers by compliance first, and an assistant can only repeat what is published.

Chemical buying runs on documents. In the US, OSHA’s Hazard Communication standard (opens in a new tab) gives safety data sheets a specified 16-section format, and the agency’s SDS guide explains what each section contains. TSCA requires EPA to keep the Inventory of chemical substances made or imported in the United States. In the EU, REACH entered into force in 2007, and a company that receives a consumer inquiry about a substance of very high concern in an article must reply within 45 days.

For AI visibility, the lesson is practical, not legal. Buyers ask regulatory questions, and assistants answer from whatever pages they find. A supplier that states, accurately and with dates, which grades are TSCA-listed or REACH-registered, and where the current SDS can be requested, gives the assistant a reliable source. A supplier that says nothing leaves the answer to distributors, directories or outdated pages. And a supplier that overstates a safety or sustainability attribute risks having that claim repeated to buyers and regulators, which is why every such statement should pass your regulatory and legal review first.

Nothing here changes existing duties. Suppliers of regulated or dual-use substances screen customers and restrict sales as the law requires, and AI visibility work should never make restricted products easier to obtain.

What does a chemical supplier lose when assistants leave it out?

Sample requests it never hears about, during a period when buyers are actively testing alternative suppliers and ingredients.

We found no public measurement of chemical sales lost to AI absence, so this is our reasoning, with the evidence that supports it:

  • Buyers are diversifying. Thomas’ 2025 data shows procurement teams evaluating alternative and secondary suppliers to reduce risk. A supplier absent from the first list is absent from that evaluation. Packaging buyers keep second suppliers too, as our guide to winning packaging quote requests through AI explains.
  • Research happens before contact. In the engineering survey, 62% of the buying process happened online before technical buyers engaged with sales. Suppliers learn about an opportunity only at the sample request.
  • Wrong facts disqualify. An assistant that says you do not carry a grade, do not ship to a region or lack a certification removes you quietly. In our study of business facts, 18.9% of AI answers about local businesses stated at least one fact that differed from the business’s Google profile. That study did not cover chemicals, but it shows how easily details drift. If an assistant already misstates one of your grades or certifications, our guide to fixing wrong brand information in AI answers shows how to find the distributor page or directory listing it came from.

How does GEO work for a chemical company?

Generative engine optimization (GEO) makes your products, documents and credentials easy for AI assistants to find, describe and verify.

For a chemical producer or distributor, the work usually includes:

  1. Product and grade pages in text. Name, CAS number, grade, typical properties, applications, compatible systems, pack sizes and regions served, written on the page rather than only in a PDF.
  2. Document access that machines can see. A public description of each technical data sheet and SDS, with a clear way to request the current version, even when the document itself sits behind a form.
  3. Regulatory status by market. Plain, dated statements of TSCA, REACH and other listings, reviewed by your regulatory team.
  4. Consistent identity across channels. The same product names, grades and facts on your site, manufacturer line cards, distributor pages, Thomasnet and association directories. Our guide on how brands build authority for AI search explains why matching product facts and outside coverage reinforce each other.
  5. Independent technical coverage. Trade press articles, conference papers, application notes co-published with customers, and association memberships.
  6. Readable, crawlable pages. Make sure product pages load without logins and that search crawlers can reach them. Our guide on what happens when AI agents can’t read your site covers common blockers.
  7. Measurement. Ask a fixed set of application, grade, regulatory and regional questions across ChatGPT, Gemini, Perplexity, Copilot and Google’s AI features, repeatedly, and compare the answers with the sample requests you actually receive.

No provider can guarantee that an assistant will name your company. The aim is to make your company the easiest supplier for an assistant, and then a chemist, to verify.

What can’t the current data tell a chemical supplier?

It shows AI spreading through chemical companies, not how often AI answers create sample requests or contracts.

  • No chemical-specific attribution. We found no public data linking AI answers to chemical sample requests, qualifications or sales.
  • Usage surveys are broad. The German survey covers AI use at work across chemical and pharmaceutical companies, not supplier searches. The engineering survey covers many industries.
  • Some results are gated. ChemPoint’s survey numbers are not public, so we did not use them.
  • Our studies are cross-industry. Applying them to chemicals is our inference.

Toll and custom manufacturers face a similar buyer path; our guide on how contract manufacturers win RFQs through AI search covers it.

Where should a chemical company begin?

Begin by asking assistants the application, grade and regulatory questions your best customers would ask.

That check shows whether your company is named for the products and markets that matter, whether grades, documents and regulatory status are described correctly, which distributors, directories and publications the answers rely on, and which suppliers appear in your place. The work then is to publish the facts buyers need in a form assistants can read, and to earn the independent coverage that confirms them.

If your growth depends on new sample requests turning into qualified accounts, ask us to review how AI search describes your products. We put your application, grade and regulatory questions to the main assistants, show which distributors and competing producers are named for them, and point to the product, document and coverage gaps that most limit qualified sample requests. How the follow-on work is run, from grade pages and dated regulatory statements to distributor consistency, is described on our generative engine optimization service page.

Frequently asked questions

Do chemists and buyers really use ChatGPT to find suppliers?

Many use generative AI at work: 76% of German chemical and pharmaceutical companies reported active AI use in 2025, and 69% of technical buyers across industries used it in purchasing. Most still verify suppliers through documents and trusted publications.

Should we publish safety data sheets openly?

That is a decision for your regulatory and legal teams. At minimum, describe each product and grade in text and make it clear how a buyer can get the current SDS, so an assistant can point buyers to you.

Does this matter for distributors as well as producers?

Yes. Distributors are often the supplier a buyer contacts, and L.E.K. found manufacturers favoring distributors with technical knowledge and regional reach. Distributor pages need the same grade, document and regulatory facts as the producer’s.

Can GEO help sell restricted or controlled chemicals?

No. GEO is about accurate commercial information for legitimate industrial buyers. Customer screening and legal restrictions apply as before.

Sources

Free strategy call

Some questions are easier to answer about your own business.

Bring the one that matters most. On a free 30-minute call we’ll take a first look at it and send you a short written read afterward.

Ask about my businessHow we work