After-sales service needs its own capability standard.

Ojamenta Insights · October 2026 · 6 min read

The automotive industry knows how to judge a process. Suppliers are assessed on the maturity of their development processes before an OEM trusts them with a programme. Production plants are certified on their quality systems before a part reaches the assembly line. Yet the place where most customers meet the brand most often, the service workshop, has no equivalent. That gap is costing OEMs, dealers and customers more than it appears.

How service is judged today

Dealer service is measured, but mostly in two ways. The first is the audit: a periodic visit with a checklist of standards, scored as present or absent on the day. The second is the outcome: customer satisfaction scores, retention and revenue, which tell you that something went well or badly, but not why.

Both are useful. Neither tells a dealer how capable their processes really are, or what to improve next. A workshop can pass an audit and still lose customers every week. Another can score well on satisfaction this quarter because of one exceptional advisor, and fall back the day that person leaves.

A checklist tells you whether a step happened. A capability standard tells you whether it happens well, every time, by design.

Why the sector needs a standard now

• Every OEM measures differently. Multi-brand dealer groups run several service standards, audits and scorecards at once, with no common language between them.
• Pass or fail gives no direction. Dealers need a clear path from where they are to the next level, not a list of findings.
• Outcomes are not traced to causes. When satisfaction falls or comebacks rise, there is no structured way to trace it back to the process that failed.
• The vehicle is changing faster than the workshop. Electric and connected vehicles bring new diagnostics, new skills and new customer expectations, and service processes must mature with them.
• Consistency is the brand promise. A customer expects the same experience at every dealer of the same brand. Without a shared measure of capability, consistency is left to chance.

What a good framework should do

Other parts of the industry have shown what works. A capability framework for after-sales should assess how well each process is performed and managed, not only whether it exists. It should be evidence-based, so that a rating can be shown and checked. It should give every dealer a clear improvement path, so assessment leads to action. It should connect what the customer experiences to the process behind it. And it should be OEM-neutral, so that capability can be compared fairly across brands and networks.

Ojamenta’s framework

Drawing on long experience in automotive process capability, quality and delivery governance, Ojamenta has developed a process capability evaluation framework built specifically for automotive after-sales service. The framework is ready, and the steps to protect its intellectual property are under way. We will share more as that process completes, and we are open to conversations with OEMs and dealer groups who would like to shape its first applications.

When the framework meets the platform

A framework on paper still depends on periodic, manual assessment. Its real value appears when it is built into the tools the workshop already uses every day.

With Valora running the service journey, the evidence a capability assessment needs is captured as the work happens: every booking, reception, repair order, quality check and follow-up. Assessment stops being an annual event and becomes continuous. Gaps are visible the week they appear, each linked to a clear next step. Dealers see exactly which process to strengthen. OEMs see capability across the entire network on one page, and can direct support where it matters most. Customers simply notice that every visit is as good as the last.

Together, a capability standard and an intelligent platform turn after-sales from a function that is audited into one that is continuously improving.

Your engineering got faster. Your management did not.

Ojamenta Insights · October 2026 · 6 min read

Over the last two years, engineering organisations have pushed AI deep into the technical workflow. Code is drafted, defects are triaged and tests are written with assistance. The same number of engineers now moves considerably more work. That was the goal, and it is working. It has also created a problem almost nobody planned for.

More work delivered means more work to oversee

Every extra ticket, defect, decision and risk still lands on a project lead or a programme manager. Their tools and habits have barely changed. They ask leads what is happening, wait for the answer to travel back up the chain, and assemble it into a status report. By the time it reaches them, it is neither current nor complete.

The numbers on this are not flattering. Asana’s Anatomy of Work research found that around six in every ten working hours go on “work about work”: chasing updates, attending status meetings and switching between tools. Gartner has forecast that AI will take over most of today’s project-management chores by 2030, above all collecting data, tracking progress and reporting.

A programme rarely fails because the code was slow. It fails because nobody noticed the one thing that stopped moving.

Perceived speed and real delivery have come apart

In a controlled trial published in July 2025, METR found that experienced developers using AI tools took about 19% longer to finish their tasks, while believing they had been faster. The finding is not an argument against AI tools. It is an argument for evidence. When perception and reality diverge, the only defence is a management layer that looks at what actually happened, not at what was reported.

The information already exists

Everything a manager needs is already written down. Every ticket records two different things: what it claims, such as a status of “In progress”, and what actually happened to it, such as comments, status changes, time logged and attachments. Nobody can fake the second by dragging a card across a board. When a ticket claims to be moving but nothing has happened to it for days, that gap is the finding.

The same applies to risks with no mitigation written, actions with no owner, and work that has sat in “New” for a fortnight. All of it is discoverable. Very little of it reaches anyone’s screen until a review or an audit.

What an augmented management layer looks like

The answer is not another dashboard. Dashboards show everything and decide nothing. A useful management layer does four things every morning:

• Reads the systems the programme already writes to, without asking anyone for an update.
• Checks what each item claims against the evidence, by arithmetic rather than opinion.
• Surfaces only what diverges, ranked, with the person to speak to named.
• Scores each team’s delivery discipline daily, so improvement conversations start from evidence.

AI has a role here, but a narrow one. It can write the morning brief in plain words and turn spoken minutes into proposed actions and risks. It should never be the thing that finds the problem or produces the number. Managers do not act on numbers they cannot check, so every figure must be reproducible by hand from the source system.

The prize is span of control

The limit on how many people one lead can oversee was never courage or capability. It was how much one person could find out in a morning. When the finding-out is done before the lead opens their laptop, the same lead can cover far more ground without being overwhelmed. That is not an argument for fewer leads. It is an argument for leads who spend their week deciding and unblocking instead of chasing.

Engineering got its AI upgrade first. Management is the other half of the system, and it is overdue.

Sources: Asana, Anatomy of Work Index; Gartner, AI and project management forecast (2019); METR, “Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity” (July 2025).

Augment the service advisor. Do not replace them.

Ojamenta Insights · October 2026 · 5 min read

Walk into most dealer workshops on a Saturday morning and you will find the same scene. A queue at reception, service advisors typing into the dealer management system while the customer waits, and a phone that keeps ringing. The advisor is the person the customer trusts most, and they spend much of their day on work that keeps them away from the customer.

Why the advisor matters so much

In after-sales, the advisor is the relationship. They interpret the customer’s concern, decide what goes on the repair order, explain the estimate and recommend additional work. Whether the customer returns for the next service, accepts an extended warranty or tells their friends to go elsewhere is largely decided at that desk.

That is why the question for AI in the workshop is not “how many advisors can we remove?” It is “how much more can each advisor give to each customer?”

The best use of AI at the service desk is to give the advisor back the time they spend typing.

Where the advisor’s time actually goes

• Re-collecting information the customer already gave on the booking call.
• Writing the walkaround by hand: damage, odometer, belongings and the insurance copy.
• Building the repair order line by line, and checking open campaigns and bulletins.
• Phone tag to get approval for additional work found on the lift.
• Answering “is my car ready?” calls through the day.

None of this needs the advisor’s judgement. All of it can be prepared, drafted or automated so that the advisor reviews and decides rather than types.

What augmentation looks like at each step

When the customer arrives, their concerns from the booking call, in their own words, are already on the advisor’s screen. The tablet walkaround marks visible damage and reads the odometer, and the advisor confirms it with the customer standing beside the car. The repair order arrives drafted from the concerns, the vehicle’s history and open campaigns, and the advisor adjusts it. Additional work goes to the customer on WhatsApp with photos, and the advisor steps in only if there is no reply. Progress updates go out on their own.

The advisor still owns every decision. They simply stop being the system’s data-entry operator.

What changes for the dealer

Faster reception means shorter queues and more customers handled at the same desk. Accurate repair orders mean fewer missed jobs and fewer comebacks. Photo-backed approvals raise the acceptance of genuine additional work without hard selling. And an advisor who has time to talk is the strongest retention tool a dealer has.

The technology that wins in after-sales will be the one advisors want to use, because it makes them better at the part of the job that only they can do.

AI on the customer call: disclosure, consent and the human handover

Ojamenta Insights · October 2026 · 5 min read

AI voice agents can now answer every call, at any hour, in the customer’s own language. For businesses that miss a large share of inbound calls, that is a real gain. It also puts a machine in one of the most personal moments of the customer relationship: a conversation. Whether customers trust it comes down to three rules.

1. Say it is AI, at the start

Customers should never discover halfway through a call that they have been speaking to a machine. The agent should introduce itself as an AI assistant in its first sentence, in plain words, in the customer’s language. Disclosure costs almost nothing, and it changes how people feel about the rest of the conversation. Hiding it turns a convenience into a deception.

Trust is built in the first sentence of the call, not the last.

2. Ask before you keep anything

A voice call produces personal data: a phone number, a vehicle registration, a recording and a transcript. India’s Digital Personal Data Protection Act, 2023 sets clear expectations on notice, consent and purpose. In practice that means telling the customer the call is recorded and why, using what is collected only for that purpose, keeping it no longer than needed, and making it easy to ask for it to be deleted. Privacy should be designed into the call flow, not added to a policy page afterwards.

3. Always leave a door to a person

No agent handles every situation well. A customer may be upset, confused or simply prefer a human. Two safeguards matter. The customer should be able to ask for a person at any point and be transferred. And a human supervisor should be able to see how every live call is going and join it the moment sentiment drops, with a short summary handed over so the customer does not have to repeat themselves.

A good handover is not a failure of the AI. It is the system working as designed.

Language is part of respect

In a country with many languages, answering a customer in the language they choose is not a feature. It is courtesy. An agent that detects the customer’s language and switches with them, and that understands mixed-language speech the way people actually talk, earns patience that an English-only menu never will.

The test

Before deploying any AI on customer calls, ask one question: if the customer listened to the recording afterwards, would they feel respected? If the answer is yes, the technology is serving people. If not, no efficiency gain is worth it.