Will AI Let Chemical Manufacturers Skip the Distributor?
AI is more likely to change what chemical distributors are paid for than to let manufacturers skip them. Order entry, reorders of known grades and shipment updates are the work automation absorbs. A distributor that carries application knowledge to the customer's line, through technical answers, trials, field follow-up and documented reasoning, keeps a role a manufacturer cannot easily copy.
On June 4, 2026, Univar Solutions announced plans for an enhanced partnership with BASF's Petrochemicals business that would make it the exclusive distribution partner for Hexamoll DINCH in the United States and Canada (Univar Solutions, 2026).
In a summer 2025 survey of 170 chemical distributors in ten European countries, 51% were testing or actively using generative AI and a further 23% were evaluating it (BCG and FECC, 2026).
FECC reports that chemical distributors in Europe serve more than 1 million downstream users, in sectors from pharmaceuticals and construction to paints and coatings, agriculture and automotive (BCG and FECC, 2026).
Azelis reports more than 70 application laboratories and more than 65,000 customers (Azelis, 2026).
EU REACH defines a distributor as a person who only stores and places a substance on the market for third parties (Regulation (EC) No 1907/2006, Article 3(14), consolidated 2026).
Why did BASF choose a chemical distributor for Hexamoll DINCH in 2026?
BASF chose a chemical distributor for Hexamoll DINCH in 2026 because, in its own words, the distributor brought market presence, technical selling and logistics to customers in the region. On June 4, 2026, Univar Solutions announced plans to enter into an enhanced strategic partnership with BASF's Petrochemicals business, making Univar Solutions the exclusive distribution partner in the United States and Canada for Hexamoll DINCH, a non-phthalate plasticizer (Univar Solutions, 2026). The release describes the agreement as a plan and does not report its completion.
Dr. Michael Limbach, BASF's Vice President for the Business Management Industrial Petrochemicals, gave the reason in the release: "Univar Solutions' strong market presence, technical selling capabilities, and established logistics infrastructure make them a highly effective partner as we continue to support customers with reliable access to our non-phthalate plasticizers portfolio throughout the region" (Univar Solutions, 2026). Univar Solutions added that the expanded relationship strengthens its ability to provide localized service, dependable logistics and application driven insights.
One agreement for one product line proves nothing about the industry as a whole. What it records is the reason a large producer gave in June 2026 for routing a product through a distributor: technical selling and logistics, named side by side. The rest of this article asks which of those two AI is likely to change, and which it is likely to make more valuable.
Which chemical distributor work can AI automate?
The chemical distributor work AI can automate is the order-and-ship layer: taking orders for known grades, confirming stock, sending documents, booking delivery and reporting where a shipment is. EU REACH uses a narrow legal definition that matches this layer, describing a distributor as a person "who only stores and places on the market a substance" for third parties (Regulation (EC) No 1907/2006, Article 3(14), consolidated 2026). The definition sets regulatory duties, not a business model, but it names the part of the job that runs on structured data: product codes, prices, stock and delivery dates.
Structured, repeatable work is where digital tools spread first, and European distributors are investing in them. In a summer 2025 survey of 170 chemical distributors in ten European countries, 63% ranked digitalization among their top three investment areas, ahead of supply chain and logistics at 55% (BCG and FECC, 2026). Specialty distributors ranked digitalization as their top priority. The same survey found 51% of distributors testing or actively using generative AI (BCG and FECC, 2026).
The consequence for manufacturers follows from that. When ordering and tracking become software, a manufacturer can offer the same convenience through its own portal, and a distributor whose only service is order-and-ship has little left that the manufacturer cannot run itself. That is an inference from how the work is built, not a survey finding.
Why do chemical manufacturers still need a distributor's application knowledge?
Chemical manufacturers still need a distributor's application knowledge because the questions that decide a specialty sale are asked at the customer's line, and there are far more lines than a manufacturer's technical staff can visit. According to FECC, chemical distributors in Europe serve more than 1 million downstream users across pharmaceuticals, construction, paints and coatings, agriculture, cosmetics, food and automotive (BCG and FECC, 2026). FECC lists distributor services that range from logistics and packaging to formulation, research and development, training and regulatory advice.
Large distributors describe their value in the same terms. Azelis reports more than 70 application laboratories where its teams develop formulations and provide technical guidance throughout the product development process, and says it serves more than 65,000 customers (Azelis, 2026).
At the customer's site, application knowledge is specific. A coatings plant asks whether a dispersant tolerates a new pigment load. A cable compounder asks whether switching plasticizer will change how a jacket compound behaves in cold weather, and what its technical data sheet (TDS) does and does not say about that. The answer depends on the customer's conditions, the current TDS and safety data sheet (SDS), earlier trials at that site and results at similar sites. Someone has to gather those facts, plan the trial and come back after it. That loop of answer, trial and follow-up is the part of distribution that ordering software does not replace, and it is what BASF named alongside logistics when it cited technical selling.
How does AI change the two chemical distributor roles?
AI changes the two chemical distributor roles in opposite directions: it makes the order-and-ship role cheaper to copy and the application role more valuable to keep. The table below sets the two roles side by side to show which parts a manufacturer could take in-house.
Question | Order-and-ship role | Application role |
|---|---|---|
What the customer pays for | Stock, credit and delivery of a known grade | An answer that the product works on its line |
What AI does to the work | Automates order entry, tracking and document dispatch | Speeds up answers built from the customer's conditions and history |
Where the know-how sits | Price, stock and delivery data | TDS and SDS revisions, trial results, visit notes, past inquiries |
Can a manufacturer copy it with a portal? | Largely yes | Only with field staff at every site |
What the manufacturer gets back | Volume and payment | Field results and the reasoning behind each recommendation |
The row that settles the question is whether a manufacturer can copy the work with a portal. Order entry and shipment tracking can be rebuilt by any producer with an online channel, so a distributor that offers only those competes with its own supplier. Application work cannot be rebuilt that way, because it needs someone who knows the customer's line, has the trial history and returns after the change. The last row adds a second reason to keep that distributor. A manufacturer learns from field results it would never see on its own, and a distributor that writes down what happened on each line gives its supplier something no portal produces. Consolidation pushes in the same direction, as the consolidation wave in industrial chemical distribution shows for technical service.
Can chemical manufacturers sell direct without a distributor?
Chemical manufacturers can sell direct where the customer buys large volumes on specification and has its own technical staff, and AI is likely to make that direct channel cheaper to run. A large account that orders the same grade every month needs a reliable order channel more than an application chemist, and a portal serves it. Digital ordering is likely to lower the cost of serving mid-sized accounts directly, which narrows the band of customers a manufacturer must hand to a distributor. That is an inference from how ordering work is built, not a measured result.
The survey does not measure order-and-ship work directly, but its split by segment points the same way: in a summer 2025 survey of 170 European chemical distributors, 56% of commodity-focused distributors reported worsening results, against 30% of specialty distributors (BCG and FECC, 2026). One respondent summed up the demand side: "Customers request lower prices while ordering less".
The opposing view is right about part of the market. A distributor that offers only stock, credit and delivery competes with a channel the manufacturer can build itself. The view is weaker for specialty products sold into many small and varied applications, where a wrong recommendation fails on the customer's line and the price of a wrong recommendation is paid in trust as well as product. In that market the manufacturer's choice is not between a distributor and a portal. It is between a distributor that carries application knowledge and one that does not.
What will chemical distributors be paid for once ordering is automated?
Once ordering is automated, chemical distributors are likely to be paid for evidence: answers that hold up on the customer's line and a record that shows the manufacturer what happened there. That is an inference from the reasons BASF gave and the services FECC lists, not a published forecast.
Manufacturers depend on what a distributor sees in the field. When a product fails at one customer, the manufacturer's technical team needs the line conditions, the sequence of events and what was already tried, and the gap between a vague complaint and a usable one is set out in what manufacturers need from a distributor complaint. The same record decides when a distributor should escalate an issue to the manufacturer and what the manufacturer's lab is asked to examine.
Customers judge distributors on the same work. A plant that runs a trial wants a plan with success and stop criteria, a visit report it can act on and a clear answer when the result is mixed. A manufacturer comparing distributors can see that work only if it is written down, which is why scoring a distributor's technical support starts from records of answered questions, trials and visits. Documented reasoning turns field work that used to be invisible into something both the customer and the manufacturer can check.
What should a chemical distributor's technical work run on?
A chemical distributor's technical work should run on a tool that keeps each answer tied to its evidence and turns each case into a record the next case can use. Five conditions follow from the role described above:
Every answer shows its basis: the TDS, SDS and certificate of analysis (CoA) it used and the customer's field history.
The reasoning from the customer's conditions to the recommendation is visible, and a person approves it before it reaches the customer.
Accounts, projects, work logs and documents connect in one record, so a new question at a site starts from what is already known there.
One case can be written up for different readers, such as the customer contact or the manufacturer's technical team.
Where information is missing, the tool marks an estimate as an estimate or leaves the point unconfirmed.
Waiting has a cost that builds case by case. Each technical answer sent from an inbox without a record is field knowledge the distributor cannot show a manufacturer when the next exclusive line is awarded, and each trial written up only in email is one the next rep repeats. A team can start with one workflow, such as replies to customer technical questions. Every answered case becomes part of the record, and the record grows into the evidence that keeps the distributor in the channel.