Does AI Narrow the Gap for Mid-Size Specialty Chemical Suppliers?
AI narrows the gap between mid-size specialty chemical suppliers and the majors where the gap is application knowledge, not plant scale. A small team whose AI carries its own data sheets and field history can answer at a depth the majors built large organizations for. The gap narrows for suppliers that begin and widens for those that wait.
In the European Central Bank's survey of euro area firms for October to December 2025, 38% of firms were at an advanced stage of AI adoption, against around 45% of large firms, while the share reporting significant AI use was similar across size classes (ECB, 2026; 5,067 firms across sectors).
Among euro area firms making significant use of AI, small and medium-sized enterprises expected to put 21% of their investment into AI, against 17% for large firms (ECB, 2026).
In 2025, 17% of small EU enterprises, 30.36% of medium enterprises and 55.03% of large enterprises used at least one AI technology (Eurostat, 2025; 157,000 enterprises with 10 or more employees surveyed).
BASF reported approximately 9,000 research and development employees in 2025, and 87% of its 1,995 million euros of R&D expenses went to R&D in its operating divisions, which BASF describes as mainly application- and customer-related (BASF Report 2025, 2026).
Of about 31,000 chemical companies in Europe, 97% are small and medium-sized enterprises, and EU27 chemical SMEs account for 28% of sales and 36% of employment (Cefic, 2022 and 2023 data).
What did the ECB's 2026 survey find about AI adoption by firm size?
The European Central Bank found that larger euro area firms adopt AI more often, but that intensive AI use is spread evenly across firm sizes. The finding appeared in a box of the ECB Economic Bulletin, Issue 2/2026, published on 2 April 2026, based on ad hoc AI questions in the Survey on the Access to Finance of Enterprises (SAFE) for October to December 2025 (ECB, 2026). The round surveyed 5,067 euro area firms across sectors, 92% of them with fewer than 250 employees (ECB, 2026). The survey defines small and medium-sized enterprises (SMEs) as firms with up to 250 employees.
Across all respondents, 38% of firms were at an advanced stage of AI adoption, meaning significant or moderate use, and around 45% of large firms were at that stage (ECB, 2026). The ECB then adds the point that matters for a mid-size supplier: "the share of firms reporting significant use of AI is similar across size classes and ownership types, suggesting that AI adoption is spreading evenly among a core group of firms" (ECB, 2026).
The investment plans point the same way. Firms already making significant use of AI expected to allocate 20% of their investment to AI, and among those firms SMEs led with 21%, against 17% for large firms (ECB, 2026). The ECB reads the pattern as "a reinforcing cycle, where firms already using AI invest more to further develop and integrate these technologies" (ECB, 2026). Read together, the figures suggest size predicts who starts more than how far a committed adopter goes.
Why did the majors' technical service organizations open the gap?
The majors' technical service organizations opened the gap because specialty chemicals are sold on application knowledge, and that knowledge has long scaled with headcount. A specialty product earns its price when someone can say why it will work on this customer's substrate, in this bath, at this line speed, and what to change when it does not. That answer needs people who have seen many similar sites.
The large producers fund that capacity at scale. BASF reported approximately 9,000 employees in research and development in 2025, and said that R&D in its operating divisions, "which are mainly application- and customer-related, accounted for 87%" of 1,995 million euros in R&D expenses (BASF Report 2025, 2026). Application labs, industry specialists and regional technical service teams give a major's customer a deep bench to call.
Most of the industry does not look like that. Cefic counts about 31,000 chemical companies in Europe, 97% of them SMEs, and EU27 chemical SMEs contribute 28% of sales and 36% of employment (Cefic, 2022 and 2023 data). In a mid-size supplier, a handful of application engineers often cover a long list of accounts. Their knowledge sits in their own heads, in email threads and in visit reports nobody else reads, so each answer depends on whether the right person is free that week.
How does AI let a small technical team answer at the majors' level?
AI lets a small technical team answer at the majors' level when it carries the team's application knowledge and reasons over it, so depth no longer depends on how many specialists are on staff. The bottleneck in a mid-size supplier is rarely the chemistry. It is the senior engineer's time, spent rebuilding each customer's conditions and finding the right data sheet revision before any reasoning starts.
Take an inquiry about a two-component adhesive failing to cure on a customer's new composite part. The answer depends on the substrate, the shop's humidity, the mix ratio the operators actually use, the open-time limits in the technical data sheet (TDS) and how a similar bond behaved at another plant. AI that reads the supplier's own TDS, safety data sheet (SDS), certificate of analysis (CoA) and earlier case records can assemble those conditions, set them against the product's limits and draft a reasoned answer with its basis attached. The engineer then reviews and approves the reasoning instead of building it from scratch.
That changes the ratio between people and accounts. One experienced engineer can stand behind many more answers, and a newer colleague can answer from the record a senior colleague left, which is the logic behind turning a senior engineer's way of working into a team skill. Scale used to buy depth. A record that the AI reasons over can supply part of it.
Why does the gap widen for suppliers that wait?
The gap widens for suppliers that wait because AI built on a supplier's own cases improves with use, and the record behind it cannot be bought later. Every inquiry a team answers this way leaves a case: the customer's conditions, the reasoning, the product chosen and what happened on site. The next similar inquiry starts from that case, not from an empty page.
The ECB sees a related pattern in investment plans: firms making significant use of AI expected to put 20% of their investment into AI, against 4% on average for firms not using it (ECB, 2026). Eurostat's 2025 survey of enterprises with 10 or more employees shows where the large firms stand, with 55.03% of large EU enterprises using at least one AI technology, against 30.36% of medium and 17% of small enterprises (Eurostat, 2025).
For a mid-size specialty supplier, the practical consequence is a compounding difference. A competitor that started a year earlier holds a year of reasoned cases its AI can draw on, and its engineers have learned which answers the customer trusts. A supplier that starts later begins with the same tools and an empty record. The distance between them is measured in cases, and cases accumulate only by doing the work. When a senior engineer leaves, a team without that record loses more than one person.
How does a major's technical service model compare with an AI-supported small team?
A major's technical service model and an AI-supported small team differ mainly in where application knowledge lives and what scales it. The table below compares the two for the work of answering a customer's technical question.
Question | Major's technical service organization | Mid-size team with AI over its own record |
|---|---|---|
Where application knowledge lives | Many specialists, application labs, internal reports | A few engineers plus a shared record of cases, documents and field history |
What scales the depth of an answer | Headcount and lab capacity | Cases accumulated and reasoned over |
How an inquiry gets answered | Routed to the right specialist | Drafted from the record with its basis, approved by an engineer |
What happens when a specialist leaves | Others on the bench cover | The reasoning stays in the record if it was captured case by case |
Where scale still decides | Pilot plants, broad portfolio, global supply | The same physical limits, which AI does not change |
The row that decides the question is what scales the depth of an answer. For a major, depth grows with headcount; for a small team, it can grow with cases, provided each answered inquiry becomes part of a record the team and its AI can use again.
The other rows follow from that one. A major routes an inquiry to the specialist who knows the application, and the answer is only as available as that person. A small team with a shared record drafts the answer from earlier cases and the current data sheets, and an engineer checks the reasoning before it goes out. When a specialist leaves a major, colleagues on a deep bench cover the gap. When one leaves a mid-size supplier, the reasoning survives only if it was written into the record case by case, which is why the record, not the tool alone, is what closes the distance.
Where do the majors keep their advantage over mid-size suppliers?
The majors keep their advantage over mid-size suppliers wherever scale is physical or regulatory rather than a matter of knowledge. AI does not build a pilot plant, widen a product portfolio or add production sites in another region. A customer that needs a new molecule developed, qualified on three continents and supplied from two plants will still go to a supplier that has those assets.
The majors are also adopting AI. Eurostat found that 55.03% of large EU enterprises used AI in 2025 (Eurostat, 2025), and among firms not yet using AI, the ECB reports that large firms are less likely than SMEs to see it as lacking usefulness (ECB, 2026). A mid-size supplier that expects the majors to stand still is reading the data wrong.
What the data does support is narrower and more useful. The share of firms making significant use of AI is similar across size classes (ECB, 2026), so a mid-size supplier that commits is not structurally behind. Many specialty sales turn on the technical answer to a specific customer's specific problem, and on that ground a well-supported small team can meet a major's depth. The contest moves from who has more specialists to who has the better record.
What should AI for a mid-size specialty chemical technical team be able to do?
AI for a mid-size specialty chemical technical team should let a few engineers answer with the depth of a large bench and keep every answer as a reusable case. Five conditions follow:
It answers from the supplier's own TDS, SDS and CoA and its field history, and shows that basis next to each answer.
It lays out the reasoning from the customer's conditions to the recommendation, and an engineer approves it before it goes out.
It closes each case as one document that the next similar inquiry can draw on.
It connects accounts, projects, work logs and documents, so a newer engineer sees what the team already knows about a customer.
It does not fill gaps with invention: an estimate carries its basis, and a point without data stays marked as unconfirmed.
The cost of waiting is the record a competitor builds in the meantime. Each inquiry answered from memory and email leaves nothing for the next one, while a team that started earlier answers its next similar inquiry faster and with more evidence behind it.
A mid-size team can start with one workflow, such as replies to incoming technical inquiries on a handful of accounts. The first cases are thin. As they accumulate, the record starts to answer questions that used to wait for the one engineer who knew. That is also how a team can find a similar problem's fix after its engineer has left.