Executive case narrative | Before

The margin meeting kept starting with archaeology.

Innospec had the ingredients of a high-performing specialty-chemicals operator: technical product depth, global customer relationships, disciplined R&D, and a major ERP program in flight. The visible problem was margin pressure. The operating problem was slower: every commercial exception required people to reconstruct the facts across systems, emails, product evidence, regulatory context, and supply-chain constraints.

The situation

Public filings describe exactly the environment where hidden workflow drag becomes expensive: segment volatility, raw-material complexity, regulatory obligations, technical service, restructuring, and ERP transition.

The pressure was not uniform. Fuel Specialties was resilient, Performance Chemicals grew sales but saw operating income fall, and Oilfield Services declined sharply as Mexico production and U.S. stimulation conditions weighed on the business. In the third quarter, Innospec recorded asset and intangible impairments tied to weaker near-term recovery expectations in selected businesses.

Inside a company like this, the same question appears in several forms: Which exceptions can we act on this week, and which ones are waiting on missing evidence? A pricing action may require supply input. A customer request may require technical and regulatory context. A field-services issue may need finance, inventory, and commercial judgment. Hiring for a scarce technical role may sit outside the operating meeting but still determine whether the work moves.

“By the time a margin exception reached the meeting, the first half of the discussion was usually spent rebuilding the story. We had data. What we did not have was a clean operating file: owner, evidence, decision, next step.”

Composite stakeholder voice | Commercial leadership

The finding

The X-Ray focused on the work between systems rather than the systems themselves.

The issue was not a missing dashboard. It was four unmanaged loops.

Loop 01

Margin exception assembly

People manually gathered price, cost, customer, and regional context before leaders could decide.

Loop 02

Technical evidence chase

Product, R&D, regulatory, and customer-trial facts lived in different places and formats.

Loop 03

ERP-adjacent exceptions

The ERP program standardized records, but not every decision path around the record.

Loop 04

Critical-role throughput

Technical and operating roles created hiring and onboarding dependencies that affected execution.

Public-source basis: Innospec FY2025/Q4 results release; Innospec Q3 2025 results release; Innospec 2025 Form 10-K; Innospec About Us. Composite quotes are used to express representative operating pain, not attributed statements.

Solution and after

Deskwise turned the work into a control layer leaders could actually run.

The solution was intentionally narrow: do not replace ERP, do not automate judgment, and do not launch a generic AI program. Build the missing operating layer for exceptions, evidence, ownership, and follow-through. The system gave each case a file, each file an owner, and each owner the evidence needed to move the decision forward.

The intervention

A margin execution cockpit with human review loops

1. Intake and classify

Capture margin, supply, customer, regulatory, ERP, and hiring exceptions through one structured intake rather than scattered email threads.

2. Build the evidence file

Use AI assistance to summarize source material, flag missing inputs, and assemble review-ready packets with links back to the underlying records.

3. Route for judgment

Send the decision to the right human owner, track aging, document the decision, and push clean status back to the operating rhythm.

Before Deskwise

The operating meeting looked backward. Leaders reviewed what had happened, then asked who could find the missing context.

Exceptions had no single owner. Commercial, supply chain, regulatory, finance, and technical teams each held part of the answer.

ERP created a better record, not a better handoff. The work around the record still lived in side trackers and inbox follow-ups.

After Deskwise

The meeting started with the queue. Leaders saw which decisions were ready, which were blocked, and who owned the next step.

Review packets arrived pre-assembled. AI summarized the file; humans made the decision where judgment, risk, or customer context mattered.

Follow-through became measurable. Aging exceptions, missing evidence, rework, and handoff latency became operating metrics.

The change

The most important change was behavioral. Teams stopped asking, “Who has the latest version?” and started asking, “What decision is this waiting on?”

“The system did not tell us what commercial decision to make. It made the decision visible early enough that we could make it with the right people in the room.”

Composite stakeholder voice | Business unit leader

The workflow layer changed the rhythm of execution. A raw-material movement could be connected to affected customers. A regulatory evidence request could be routed with the source documents already attached. A technical-service escalation could show whether the next action belonged to the field, the lab, the commercial owner, or finance. Hiring and onboarding dependencies for scarce roles could be tracked as part of execution capacity rather than treated as a separate HR problem.

Same-day

Routing target for new margin or technical-service exceptions after intake.

50-70%

Target reduction in time spent assembling review packets for recurring exception types.

Target duplicate side trackers for the pilot queue once the workflow layer is operational.

Why it mattered

The business did not need another reporting surface. It needed an execution surface.

In a specialty-chemicals company, the value of a decision depends on timing and confidence. Margin leakage, customer response, regulatory readiness, and technical-service quality are all affected by the same operating problem: the decision file is rarely complete when the decision is first needed.

Deskwise made the hidden work explicit. It separated robot work from judgment work, reduced the administrative burden on high-value experts, and gave management a practical way to see where execution was stuck before the next monthly explanation.

Source links: FY2025/Q4 release; Q3 2025 release; 2025 Form 10-K; Innospec About Us. Composite quotes are not attributed to named Innospec employees.