The Last 20%:  How Human Judgment Wins in the Future AI-Driven Dealership 

Author: Matt Wills, Senior Partner for Sheppard and Company

If AI is the new toolbox, human judgment is the blueprint that still determines whether a project stands or falls. Equipment dealers have watched AI promises with equal parts excitement and skepticism. The reality so far is messier, but more useful than the headlines suggested. AI excels at pattern recognition, prediction and synthesis at scale, yet it does not carry the lived context, judgment and trust that make a dealership valuable. The pressing question for equipment dealers is not whether to adopt AI, but how to attach it to real operational problems so human judgment captures what many others have described as the decisive last 20 percent of value, the work that determines margin, uptime and customer loyalty. 

Think of AI as an accelerator, not a strategy…

Think of AI as an accelerator, not a strategy. The commercial equation is simple and practical: identify a real pain point, collect the right signals and records, apply AI to surface patterns and predictions, and then use human judgment to verify, prioritize and communicate. Too many dealers stop at convenient automation, email drafting, meeting summaries or paperwork reduction, and leave deeper commercial advantage on the table. The more valuable opportunity is to apply AI to front-line problems that directly influence a customer’s jobsite productivity, such as predicting failures before they become breakdowns, preventing parts stockouts, optimizing loaner fleets and improving first-time fix rates. 

The Last 20% Is Where Real Commercial Value Lives

The last 20 percent is not a brand new insight, and many leaders and authors have already called attention to it. Still, the reality matters: AI will likely handle the repetitive 80 percent of work, such as searching manuals, summarizing histories, routing tickets, forecasting demand and producing initial diagnostics. The remaining 20 percent, which others have emphasized, is where nuance lives. Those are the judgment calls about triaging a high-value account with a down machine, promising an ETA that does not erode trust, deciding whether to cannibalize a unit for parts or reallocating a loaner fleet in the middle of a busy season. Those calls require context, relationships and a feel for downstream consequences on margin and uptime. That last 20 percent is not marginal, it determines business outcomes. 

Other Industries Prove the Same Pattern

Lessons from other industries are instructive. In healthcare, ambient AI scribes that capture clinical visits have dramatically cut documentation burdens, freeing clinicians to spend more time interpreting information and building trust with patients. In logistics, route optimization systems delivered measurable gain only after deep integration, training and routine adjustments by drivers and dispatchers. In our own industry, OEMs and OEM-adjacent vendors are shipping tools that give technicians immediate access to massive libraries of manuals and fault histories. Those tools compress time to knowledge, but they do not replace the technician’s judgment about severity, parts substitutions or customer communications. The pattern is consistent: AI handles the heavy informational lifting, humans decide where and how to act. 

Dealers Must Evolve Into Intelligence Partners

So what should an equipment dealership become? The answer is a shift from product seller to intelligence partner. Successful dealers will combine unrivaled equipment access with predictive insight and trusted human judgment. That translates into practical changes, such as selling uptime instead of only units, anticipating failures rather than primarily reacting to breakdowns, and treating parts inventory as continuity planning rather than transactional stock. Sales teams evolve into customer business advisors who use AI-synthesized intelligence to recommend right-sized fleets and service plans. Technicians become uptime strategists who convert AI-suggested diagnostics into prioritized, real-world repair strategies. Parts teams become availability planners who act on stockout predictions. Leadership becomes the architect of an operating model that blends data, algorithms and accountable human decisions. 

AI Fails When Treated Like Software 

AI projects fail when they are treated like off-the-shelf software instead of process redesign. That pitfall is common and avoidable. Dashboards sit unused, vendors promise transformation without understanding branch workflows, one-off training sessions occur and then leadership moves on, AI recommendations go unverified by experienced staff and no KPI is tied to uptime, margin or customer loyalty. The remedy is disciplined. Map where friction actually exists, separate human-critical moments from routine tasks, match an AI capability to those frictions, and run focused pilots with clear ownership and measurable outcomes. Small experiments, human verification and sustained follow-through beat flashy rollouts.

Start With Small, Measurable Pilots That Move the P&L  

Practical pilot ideas for dealers start with what moves the P and L and the customer jobsite. Pilot an AI-assisted service intake checklist in one branch that pulls service history, telematics, manuals and similar faults, then suggests triage questions and a prioritized response plan. Measure callback time, first-touch resolution and customer update cadence. Try a parts demand predictor that alerts planners to potential stockouts and automatically proposes alternate sourcing or redistribution before a job is affected. Test an AI-driven utilization and maintenance scheduler for rental and fleet equipment that optimizes preventive maintenance windows without crippling utilization. For sales, pilot AI-enabled lead scoring and configuration logic that speeds quoting while routing complex, high-value opportunities to senior advisers. 

Leadership: Architect, Translator and Trust Steward

Leadership implications are real and immediate. Dealer principals must change how they think about technology. The job becomes architect, translator, trust steward and scaler. Decide where AI should alter the operating model, rather than simply approving another vendor tool. Translate technology into real dealership pain points and assign accountable owners for behavior change. Protect judgment, relationships and accountability by insisting on human verification of AI recommendations, especially where promises touch customer productivity. Scale what works by turning successful pilots into repeatable routines and tying them to KPIs that matter: uptime, service margin, parts fill rate and customer satisfaction. 

AI Helps, But Judgment Finishes the Job

A personal anecdote underlines the point. Three years ago I believed AI would replace CRMs. Eventually I started using an AI note-taking tool called Plaude. This is not an endorsement. It simplified the most despised part of my day, turning messy meeting notes into searchable summaries and actionable follow-ups. It felt like a small miracle. Yet it was not perfect. Like many AI tools, it made my life easier while still leaving at least 20 percent of the work to human verification, contextual corrections, prioritization and ensuring sensitive client nuances were accurately captured. Some writers say their tools are already 95 percent reliable. Perhaps I am more cautious, but my experience reinforces the same lesson others have highlighted, that AI amplifies human work without fully replacing the critical final steps. 

Final Thoughts About the Final 20%

The competitive edge for equipment dealers will not come from the raw power of an algorithm or the cleverness of a prompt. It will come from integrating AI into the interaction points that create value for customers, and from preserving the human moments that build trust. The final 20 percent, which many have noted, consists of judgment, empathy, negotiation and relationship management. It is not optional, it is the margin. Dealers who design pilots around authentic customer friction, measure outcomes tied to uptime and financials, and insist on human verification will convert automation into sustainable commercial advantage. 

A final question for every leader: where in your operation can a small, measurable AI pilot reduce friction this quarter and amplify human judgment next quarter? Start there, track the outcomes that matter and build the routines that make that last 20 percent your enduring advantage.