"We thought we needed a €2M platform. Turns out we needed five €15K tools."
A specialty chemicals manufacturer under board pressure to modernise wanted an independent assessment before committing to a seven-figure 'digital transformation.' What we found changed their entire approach to AI.
Tier
Tier 1 + Tier 2
Interviews
17
Duration
5 weeks
Organisation
420 employees · 3 facilities
■The Situation
A specialty chemicals manufacturer with three production facilities had been sold hard by two enterprise AI platform vendors — both proposing 18-month, seven-figure 'digital transformation' programmes. The COO was sceptical but under board pressure to modernise. He wanted an independent, objective assessment before committing.
The COO chose Supply Chain & Operations as the starting point — the area bleeding the most time and margin. The request was clear: map the workflows, identify where AI would genuinely help, and put hard numbers on it. No vendor alignment, no implementation bias.
If the Supply Chain findings were compelling, he would commission the full organisation audit. If they weren't, he would use the process maps internally and move on.
■What We Did
One week in Supply Chain & Operations, then four additional weeks across all departments. 17 interviews across production, procurement, quality control, and management:
Supply chain & planning (n=7)
We mapped the end-to-end production planning and raw materials procurement workflow — from demand signal to purchase order to delivery confirmation. Every manual step, every spreadsheet dependency, every step that required a phone call to complete.
Quality control (n=5)
We mapped the inspection, data-recording, and defect-escalation workflows. We documented where data was recorded, where it was re-entered, and what happened between a defect being detected and a corrective action being logged.
Production & maintenance (n=5)
Interviews on the equipment maintenance scheduling process and the production scheduling workflow. Machine downtime data was cross-referenced against both the production planning system and the quality management system.
The Supply Chain findings from week one were presented before the full audit began. The COO immediately commissioned the full engagement — the initial findings had identified 6 opportunities in Supply Chain alone worth an estimated €560,000.
The Opportunity Matrix was built incrementally as each department was completed. By the end of week five, the complete matrix covered 17 opportunities across all 8 departments.
■What We Found
17 AI opportunities totalling €1.6M in annual savings — none of which required the platforms the vendors were selling.
The two enterprise vendor proposals — both pitching full-scale AI platforms — would have addressed only 4 of the 17 opportunities Kerf identified, at 12 times the cost of the targeted solutions. The winning strategy was five focused, off-the-shelf AI tools costing a combined €74,000 to implement, not a single monolithic platform.
The biggest single opportunity was AI-driven demand forecasting. The current process required a planning team to manually build monthly demand forecasts from sales history, seasonal patterns, and customer order signals — a three-day exercise prone to error that resulted in raw material overstock running at 22% above optimal. An AI forecasting tool could reduce this to under 5%, saving an estimated €280,000 per year. Implementation cost: under €20,000.
Quality control was the second major finding. The inspection process required QC technicians to record results in one system and then re-enter summary data into the ERP — 22 minutes of double-entry per inspection batch, 11 batches per day across two production lines. An AI document extraction layer would eliminate this entirely. Estimated saving: €190,000 per year.
The Opportunity Matrix flagged both enterprise vendor proposals as 'Deprioritise' — the effort-to-impact ratio didn't justify the cost when targeted solutions could deliver better results for a fraction of the investment.
“Kerf saved us from spending €2 million on the wrong thing and showed us how to get better results for €74,000.”
— COO, Scandinavian Specialty Chemicals Manufacturer
■What Changed
Both enterprise vendor proposals were declined. The COO presented the Opportunity Matrix to the board as the company's AI strategy — the first time the board had seen a concrete, numbered plan rather than a vendor pitch deck.
Four Quick Wins were deployed within 60 days: AI-powered quality inspection (defect detection time cut from 45 minutes to 3 minutes per batch, saving €190,000/year), automated regulatory compliance documentation (report preparation cut from 14 hours/week to 2 hours/week across 3 facilities, saving €115,000/year), predictive maintenance for production equipment (unplanned downtime reduced 34%, saving €210,000/year), and AI-driven demand forecasting (raw material overstock reduced 22%, saving €280,000/year).
Total implementation cost for all four Quick Wins: €74,000. Total annualised saving from the four Quick Wins: €795,000.
€795,000 in annualised savings from Quick Wins alone, at a total implementation cost of €74,000. Two Strategic Bets representing an additional €800,000 annually entered planning. Both enterprise vendor proposals declined, saving an estimated €2M+ in platform costs.
Decision Impact
5-week engagement · 17 AI opportunities identified · €1.6M savings identified · €74K implementation cost · 4 Quick Wins deployed in 60 days · €2M+ in vendor costs avoided
Next case study
Building Materials Manufacturing · 03