Is AI a Cost Play or a Revenue Play in Specialty Chemicals?
In specialty chemicals, AI is a revenue play first, because a wrong or slow technical answer costs a deal, not just staff hours. Most 2026 surveys, PwC's January one excepted, find AI productivity and cost gains more common than revenue gains. The revenue case holds where sales depend on fitting products to customer conditions and answering technical questions with evidence.
In PwC's follow-up survey of 351 CEOs, run from 15 May to 22 June 2026, 18% reported cost decreases from AI alone, 4% revenue increases alone and 9% both (PwC, 2026).
In PwC's 29th Global CEO Survey of 4,454 CEOs, 30% reported increased revenue from AI over the previous 12 months, 26% lower costs and 56% neither (PwC, 2026).
Of 3,235 leaders surveyed, 66% report productivity and efficiency gains from AI, 40% reduced costs and 20% increased revenue, while 74% hope to grow revenue through AI (Deloitte, 2026).
Respondents most often attribute AI revenue gains to marketing and sales, followed by product and service development (McKinsey, 2026; n=1,719).
McKinsey reports, from cross-industry research, a revenue uplift of 3 to 15 percent and a sales ROI uplift of 10 to 20 percent for companies that invest in AI in marketing and sales (McKinsey, 2024).
What did PwC's August 2026 CEO Snapshot find about AI returns?
PwC's CEO Survey Snapshot, published on 4 August 2026, found that CEOs reported AI cost decreases more often than AI revenue increases. PwC returned to 351 CEOs who had taken part in its 29th Global CEO Survey and surveyed them across 59 countries and 27 sectors between 15 May and 22 June 2026 (PwC, 2026). In PwC's words, the aggregate picture had not changed dramatically, "with a similar proportion of Snapshot CEOs reporting cost decreases (18%), revenue increases (4%), or both (9%)" (PwC, 2026).
The full survey behind the Snapshot gives a broader view. PwC's 29th Global CEO Survey, published on 19 January 2026 and based on 4,454 CEOs in 95 countries and territories surveyed from 30 September to 10 November 2025, found that 30% reported increased revenue from AI over the previous 12 months, 26% reported lower costs, and 56% had seen neither (PwC, 2026). In that survey revenue gains edged out cost decreases, but more than half of CEOs had seen neither.
The Snapshot also found movement beneath the averages. Roughly half (51%) of the CEOs said AI's business impact had shifted over the eight months between the two surveys, migrating between positive impact, no change and negative impact on revenue and costs (PwC, 2026). PwC also reported that 38% of CEOs said their company had used AI since January 2026 to identify and act on new business opportunities, such as shifts in customer demand or chances to adapt products or services (PwC, 2026).
Why do AI gains show up as cost savings first?
AI gains show up as cost savings first because most AI is first pointed at individual tasks, and time saved on a task is easy to count. Deloitte's State of AI in the Enterprise 2026, based on a survey of 3,235 leaders in August and September 2025, found that 66% of organizations report productivity and efficiency gains, 40% report reduced costs and 20% report increased revenue, while 74% hope to grow revenue through AI in the future (Deloitte, 2026). McKinsey's State of AI 2026 found the same gap from another angle: 80% of respondents report that AI has improved their individual productivity, while 37% attribute at least some EBIT impact to AI use (McKinsey, 2026; n=1,719).
The mechanism is simple. A tool that drafts an email or summarizes a document saves minutes a manager can see. A tool that helps win a customer works through a longer chain: a better answer, a trial, an approval, an order. The revenue arrives months later, many people touched it, and it is rarely credited to the tool.
For a specialty chemical supplier, that measurement habit has a consequence. If AI is judged only by hours saved, the business case favors tools for order entry and document handling. The work that decides specialty sales, the technical answer to a customer's question, gets the least attention, even though it is where the supplier's knowledge earns its margin.
Where does AI revenue come from in specialty chemicals?
In specialty chemicals, AI revenue comes from the commercial and technical work that decides which product a customer buys: the recommendation, the technical answer and the evidence behind the price. McKinsey's State of AI 2026 found that respondents most often attribute AI revenue gains to marketing and sales, followed by product and service development and software engineering (McKinsey, 2026; n=1,719). In a specialty chemical supplier, technical sales and application development would fall mainly in the first two of those functions.
McKinsey's chemicals practice makes a related claim. In an October 2024 article on chemical revenues, it wrote that its research shows companies that invest in AI in marketing and sales seeing a revenue uplift of 3 to 15 percent and a sales ROI uplift of 10 to 20 percent (McKinsey, 2024). The figure comes from McKinsey's cross-industry marketing and sales research, not from chemical companies or technical service teams, and the article does not break it down by task.
PwC's 29th survey adds a pattern. Among the one in eight CEOs reporting both higher revenue and lower costs from AI, 44% had applied AI to their products, services and experiences, against 17% of other companies (PwC, 2026). For a specialty chemical customer, much of the experience is technical: how fast the supplier answers, whether the answer fits the line, and whether it holds up on trial. A supplier that applies AI there is working on the part of the sale the customer actually feels.
What does a slow or wrong technical answer cost a specialty chemical supplier?
A slow or wrong technical answer costs a specialty chemical supplier the trial, the order or the account, not just the hours spent on it. That is an inference from how specialty sales work rather than a survey result, and the mechanism is direct. A specialty product is chosen for how it performs under a customer's conditions: substrate, temperature, water hardness, cure schedule, line speed. The customer cannot test every candidate, so it relies on the supplier's answer to decide what goes on trial.
When the answer is slow, the customer's project does not wait. A supplier that replies first with a reasoned recommendation is more likely to get the trial slot, and the trial often decides the order. When the answer is wrong, the cost is larger: a failed trial, a quality claim, possibly a line stoppage, and a customer that checks every later answer twice. The price of a wrong recommendation runs well beyond the product, and a product that meets its data sheet can still fail on site when the conditions behind the answer were never checked.
Price follows the same logic. A supplier that can show why its product fits, with the technical data sheet (TDS) limits, the safety data sheet (SDS), earlier trials and similar sites behind the answer, moves the conversation from price per kilogram to fit. Without that evidence, the comparison falls back to price.
How does a cost play compare with a revenue play in specialty chemicals?
A cost play and a revenue play in specialty chemicals differ in what the AI works on and in what a miss costs. The table below sets the two side by side for a specialty chemical supplier.
Question | AI as a cost play | AI as a revenue play |
|---|---|---|
What the AI works on | Order entry, document handling, routine requests | Product recommendations, technical inquiry replies, compatibility checks |
How the gain is measured | Hours and cost saved | Trials won, orders closed, prices held |
Where 2026 surveys place the gains | Productivity and cost, reported more often than revenue (Deloitte, 2026; PwC Snapshot, August 2026) | Revenue gains most often in marketing and sales, then product development (McKinsey, 2026); PwC's January 2026 survey: revenue 30%, lower costs 26% (PwC, 2026) |
What a miss costs | Rework on one task | A lost trial, a claim or the account |
What the tool must know | The format of the task | The customer's conditions, data sheets and field history |
The row that matters most is what a miss costs. A cost tool that misreads an order causes rework. A revenue tool that gets a recommendation wrong costs the customer's trust, so it has to show its basis before anyone acts on it. That is the real dividing line: the revenue play is open only to AI whose answers a technical manager can check. The measurement row explains why the revenue play is easy to overlook: hours saved appear in a monthly report, while a trial won because the answer was right and fast shows up only in the next order. The same split runs through distribution, as set out in specialty vs commodity chemical distribution and where AI fits.
When is AI in specialty chemicals mainly a cost play?
AI in specialty chemicals is mainly a cost play where the work is high in volume and routine: order entry, requests for standard documents, repeat quotations. A supplier that receives many requests for the same SDS or certificate of analysis (CoA) gains real hours from automating them, and Deloitte's finding that 66% of organizations report productivity and efficiency gains shows how common that first win is (Deloitte, 2026).
The revenue case also deserves a measured reading. The 3 to 15 percent uplift McKinsey reports is a cross-industry consultancy figure for AI in marketing and sales, and in PwC's Snapshot 13% of CEOs reported revenue increases from AI, 4% alone and 9% with lower costs, against 27% reporting cost decreases (PwC, 2026; totals calculated). A specialty chemical executive can treat revenue impact as something to test on a few accounts and watch grow, rather than something to assume from a survey.
The two plays are not exclusive. Time saved on routine requests is time a technical team can spend on the inquiries that decide sales. PwC's full survey found that one in eight CEOs (12%) reported both higher revenue and lower costs from AI (PwC, 2026). The difference is the order of priority. In specialty chemicals, cost savings set the floor, and the technical answer is where the larger value sits.
What should AI aimed at specialty chemical revenue be able to do?
AI aimed at specialty chemical revenue should produce technical answers a customer can rely on and a technical manager can check. Five conditions follow:
Every answer shows its basis: the company's own TDS, SDS and CoA and the field history behind it.
The reasoning from the customer's conditions to the recommendation is visible, and a person approves before it goes out.
Each case ends as one document that the next similar case can draw on.
Accounts, projects, work logs and documents connect in one record, so each answer starts from what is already known about that customer.
Where information is missing, the tool does not make it up: it marks an estimate with its basis or leaves the point unconfirmed.
The cost of waiting falls on revenue, not budgets. Each technical inquiry answered from memory and email leaves no record for the next one, and each slow reply is a trial that may go to the supplier that answered first. PwC found that about half of the CEOs it surveyed twice saw AI's business impact shift, up or down, within eight months (PwC, 2026), which suggests positions are still moving while a team waits.
A specialty team can start with one workflow, such as technical inquiry replies, on a few accounts. Every answered case becomes a document, and every document makes the next answer faster and better grounded. The value builds case by case, which is the argument behind why every specialty chemical case ends in one document.