A four-person aftermarket team at a global OEM paid for itself in its first year. Here’s how the company plans to scale that program without adding more headcount.
I recently sat down with the aftermarket leadership team of a global industrial OEM. They had already done something most OEMs only talk about doing: they assigned real people to uncover business opportunities from the installed base.
A year ago, they funded a four-person aftermarket capture team spread across the Americas, EMEA, and APAC. The team’s job was to comb through five to seven years of project shipments, identify the customers who would soon need parts, and reach out to them. In practice, that meant sending spare-parts recommendations, following up on open quotes, and handing qualified opportunities to field sales.
It worked. Opportunities came in, and the program more than paid for itself.
More people or more automation?
Now they’re facing the question every successful pilot eventually raises. Do they grow the program by hiring, or by automating the work?
Their own numbers made the decision easy. The team spends about half its time mining data and the other half selling. If software could handle the mining, the four people they already have could do the selling work of six. No new hires would be needed for now.
One executive put it plainly:
I don’t want our people digging through spreadsheets and CRM records. They need to be on the phone with customers and building go-to-market strategy.
In his view, they should have started automating the manual work five years ago.
Three components behind the automation
Taking the data mining off this team’s plate requires a platform with three components that work together.
Enterprise Context Engine™
Connects installed-base, CRM, ERP, and service data into one shared view
Development Studio
Where AI agents are built and tested against real customer data
Operational Harness
Runs agents in production with monitoring, governance, and human review
The Enterprise Context Engine™ connects shipment records, CRM, ERP, and service history into one view of each installed unit, so every AI agent works from the same data the sales team sees. The Development Studio is where our team and the customer’s own experts build those agents and test them against real data, starting with the use cases a manual pilot has already proven. The Operational Harness runs the agents in production, monitors their output, and routes each recommendation to a salesperson before anything reaches a customer.
Five lessons for OEMs
At Entytle (now part of RapidCanvas), we’ve worked with dozens of companies on automation projects. From that experience, we’ve found five lessons that make projects successful:
- Test the approach manually before you automate. This OEM’s first year wasn’t wasted. It showed them where data was missing, where the process broke down, and what a good opportunity looks like. You can’t automate a sales motion you haven’t codified and proven first.
- Start with data you trust. Their shipment data had been clean and centralized since 2021. Older records were an unknown, and nobody knew what problems were buried in them. In response, they limited the scope of automation to recent shipments. Even inside that window, they found process errors, including multiple units shipped under the same serial number. Bad data going in produces bad recommendations coming out, so the fix belongs at the source.
- Treat top accounts and the long tail separately. In the top 30 to 50 accounts, the goal is deeper penetration and cross-sell. Importantly, even there, the team has blind spots it can’t see on its own. In the B and C accounts, sales has no capacity and little interest in digging, yet those accounts hold significant latent revenue. Possibly even a few future A accounts. The two groups need different plays built on the same data foundation.
- Run pilots that can become programs. One leader was blunt about this. He won’t run a proof of concept that dies three months later for lack of funding or strategic fit. A pilot should connect to the enterprise data strategy from day one, so separate teams don’t pay twice to solve the same problem. The solution should work with existing data sets and platforms through a shared context layer, since many solutions can benefit from the same data foundation. For example, the installed-base data that the aftermarket team uses can be just as relevant to service, finance, and product teams.
- Plan for adoption time. Standing up a pilot takes about four weeks. Running it properly takes three to five months, because sales cycles, onboarding, and daily habits all need time to change.
Most OEMs are still early in the maturity cycle
Back to my recent meeting. When I walked this team through the progression from reactive to proactive to prescriptive aftermarket sales, they put themselves on the first or second rung. Most OEMs I meet are in the same place.
Getting to the next rung starts with installed-base data the team can trust, even if that means working only from recent shipments initially. The OEM also needs a defined path from a flagged opportunity to a quote, with a named person responsible for follow-up. Once that process works by hand, AI agents can take over the data mining, and the aftermarket team can spend its time with customers. Entytle sets up that combination for OEMs, pairing Entytle’s installed-base data with proven AI agents and the existing sales team.
If your aftermarket team is spending half its week in spreadsheets, get in touch for a consultation now.
