From 50% turnover to 72% retention

The company

Ditech is a southern Indiana manufacturer running two plants across three shifts, producing precision parts for automotive Tier 1 suppliers with roughly 160 production employees operating 24/7. Nearly all frontline hiring comes through community partnerships that employ people rebuilding their lives after difficult circumstances.

Ditech invests heavily in these workers: van transportation from regional pickup locations, paid orientation, milestone celebrations, and a culture built around giving people a real shot. The problem wasn't the mission. It was the math.

The problem

Ditech hired 60 to 100 people per year, losing 50% within 60 days and 70% within 180 days. The top cause wasn't performance or policy violations. It was job abandonment: roughly 80% of all separations.

A two-person HR department handled all onboarding manually. The HR manager spent an estimated 10 hours per hiring batch entering the same data into Plex, Paycom, PDK badge access, and sometimes a fourth direct deposit system. A batch of 13 consumed a full week before anyone reached the floor.

Once there, new hires were largely on their own. Supervisors were supposed to conduct 10, 30, and 60-day reviews, but the cadence collapsed under constant turnover: no structured coaching, no milestone tracking, no system flagging critical retention windows.

The approach

Most consulting firms would start with multi-month discovery, stakeholder interviews, process mapping, a 40-page deck, then maybe a pilot six months later. We took the opposite approach: a 2-day, on-site proof of concept.

Day one began with a three-hour deep dive into onboarding and retention with the people closest to it: the HR manager, the VP of operations, and the owner. What we heard changed the direction of the build. Owner Tim Dillingham described a "compass": a system guiding new hires through their first 180 days with encouragement, accountability, and milestone tracking. He compared it to an Apple Watch closing activity rings, employees feeling like someone cared about their progress without another management layer.

By that afternoon, our team was building. By end of day two, a working prototype was in the client's hands.

The solution

The tool we built is Ditech Pathfinder: a hybrid AI coaching agent, notification engine, and onboarding tracker for both employees and supervisors. It does four things:

1. Tracks milestones. Progress through 10, 30, 60, and 180-day milestones, with coaching checkpoints triggered automatically when someone approaches a milestone or falls behind.
2. Coaches proactively. Tailored encouragement, attendance reminders, shift prep tips, and milestone reinforcement, with a tone built around progress, not punishment.
3. Preps supervisors. A one-page prep sheet before every check-in covering attendance, milestones, and coaching interactions, so conversations are informed, not improvised.
4. Closes the loop. A real-time view of where employees drop off, which shifts churn hardest, and whether coaching interventions are actually moving retention.

Before Pathfinder, none of that data existed in one place.

What came next

The early results from the 30-person beta were strong enough that Ditech rolled Pathfinder out to all 160 production employees across both plants and all three shifts. Production efficiency has continued to climb as employees stay longer, reducing the constant retraining cycles that had been dragging down output and increasing scrap risk. Supervisors report that check-in conversations are more productive because they walk in with context instead of starting from scratch.

The 180-day number is the one that matters most to Ditech's leadership. At 25%, the company was losing three out of four hires before they ever became reliable, trained operators. At 54%, the math of the entire hiring model changed.