Technology · Aftersales
For decades, automotive aftersales has operated through a predictable sequence: the customer identifies a problem, contacts the dealer, books an appointment, brings in the vehicle, waits for a diagnosis, approves an estimate, and eventually receives the repaired car. Every step depends on someone initiating the next one. Digitizing that process (a website instead of a phone call, a form instead of paper, a chatbot instead of a receptionist) makes it more convenient, but it doesn't change its nature: it's still fundamentally reactive.
Edward's original article argues that's about to change. Artificial intelligence already helps generate content, analyze information, and answer questions. Agentic AI represents a deeper shift: from systems that inform to systems that pursue an objective and take action. A chatbot waits for a question. An agent monitors an objective, and acts within predefined limits of authority.
"A chatbot might tell a customer that a service is due. An agent could determine that the service is approaching, calculate the most convenient date based on mileage patterns, verify workshop capacity, reserve the required parts, and propose a confirmed appointment before the customer even thinks about calling."
Three stages, not a leap
The article describes an evolution in three stages: reactive service (the vehicle fails and the customer initiates the process), predictive service (the system identifies that the vehicle may fail and notifies someone), and agentic service (the system identifies the risk and coordinates the response within predefined authority limits). The repair order, Edward argues, should begin before the vehicle arrives at the workshop: symptoms, photos, history, and parts/technician availability already prepared in advance, instead of being assembled at reception.
A multi-agent environment, not a single system
The aftersales of the future wouldn't depend on one giant AI, but on a network of specialized agents collaborating around the same case: a customer agent (communication and appointments), a technical agent (symptoms and diagnosis), a parts agent (inventory and logistics), a warranty agent (evidence and compliance), a workshop agent (capacity and assignment), and a quality agent (verification before releasing the vehicle). The greatest value isn't in speeding up one department, but in reducing the wait and information loss between handovers: waiting for the diagnosis, waiting for approval, waiting for the part, waiting for someone to tell someone else.
The service advisor, more important, not less
It's easy to conclude that AI will replace the service advisor. The article argues the opposite, at least for the highest-value, most complex, or emotionally sensitive cases: what will disappear is administrative friction (searching for information, requesting updates, re-entering data, explaining internal delays), not human judgment. The advisor's value shifts to what technology doesn't easily replace: building trust, understanding customer emotion, explaining complex technical decisions, and managing exceptions.
The governance principle
Not every action should be autonomous. The article proposes a principle that Consultenos has already adopted as the basis for any future agent in this roadmap:
"Automate the routine. Recommend the uncertain. Escalate the consequential."
An aftersales agent may access vehicle telemetry, repair history, warranty data, and manufacturer systems, and could be authorized to order parts, schedule work, or communicate with customers. That requires verified identities, role-based permissions, explicit decision limits, complete action logs, and mandatory human approval for high-impact actions.
Parts: from forecasting demand to anticipating repairs
Parts planning today depends on historical demand, seasonality, and commercial forecasts. The article adds a new layer: probable repair demand at the individual vehicle level. A parts agent could detect that a group of vehicles is developing the same anomaly and anticipate rising demand for a specific component before the corresponding repair orders are even opened, recommending stock rebalancing between branches or consolidating urgent orders.
Trust, cybersecurity, and accountability
An aftersales agent connected to vehicle telemetry, repair history, payment data, and manufacturer systems represents a very different risk profile than a chatbot that only answers questions. The article cites NIST warnings about risks specific to AI agents (indirect prompt injection, insecure tool use, agent hijacking) and concludes that the automotive industry will need a governance model comparable to financial authorization: verified identities, role-based permissions, spending and decision limits, complete action logs, and immediate suspension mechanisms.
The original article's conclusion is the most direct: the future of automotive aftersales is not a workshop without people. It's a service ecosystem where people stop chasing information, correcting disconnected processes, or reacting too late. AI observes, agents coordinate, technicians solve, advisors build trust, and the most competitive companies act before the customer has to ask.
This article is a summary and adaptation of Edward Burgos's original text, published in English on LinkedIn on July 28, 2026 — not a full reproduction.
← Back to Blog
Comments
Comments are saved in this browser — visible only to you, not to other visitors.