In energy and utilities, the most expensive phone call is the one that was never made. A technician drives to an address to read, replace or service a meter, rings the bell, waits, and leaves; the visit is paid for anyway, and the work still has to be done another day.
Utilities rarely think of this as a customer-contact problem. Outbound calling is unusual in the sector, field work is often outsourced, and the cost of a wasted visit sits in a contractor invoice rather than in the contact centre's budget. That is exactly why it persists. In our work with energy companies, the field visit is one of the clearest cases where a short, well-timed conversation pays for itself before anyone has discussed customer experience.
Why utilities hardly call their customers#
Ask an energy company what it would phone customers about and the answer is often a shrug. Outages are announced by notification. Tariff changes arrive by letter or email, and there is little appetite for calling people to tell them prices are going up. Inbound lines take complaints and questions. Outbound is left with a few well-defined processes: collections, satisfaction surveys after a hotline or point-of-sale contact, and contract renewals with offers such as different rates or a price fixed for a period.
Field visits fall between these categories. They are scheduled by an operations or metering team, carried out by an external firm, and communicated to the customer by letter or text, if at all. Nobody owns the question of whether the customer will actually be there.
The economics of an empty doorstep#
Much of the field work in the sector is done by contractors. The contract usually pays for the visit, not for the outcome: the contractor turned up, so the contractor is paid. If nobody answers the intercom, the visit counts, the meter remains unread or unreplaced, and a new visit has to be planned, scheduled and paid again.
The cost is not only the second fee. A missed meter reading means an estimated bill, which generates queries and disputes on the inbound line. A missed meter replacement delays a programme with regulatory or technical deadlines. A missed service visit leaves a fault unresolved. Each failed visit therefore creates work in at least two other departments, none of which sees the original cause.
Which visits to confirm#
Not every field activity needs a call. Where the meter is outside and accessible, the technician does not need the customer. The process pays off where access depends on someone being at home:
- Meter access and readings where the meter is inside a home, a locked basement or a building with controlled entry.
- Meter replacements, including programmes that swap large numbers of meters in a defined period and need a high first-visit success rate to stay on schedule.
- Service and inspection visits for electricity or gas installations, where the visit is both a customer service and, in some cases, a safety or compliance obligation.
- Follow-up visits after a previous no-access attempt, where the customer has already been missed once and the cost of missing them again is higher.
The common thread is that the visit only works if a specific person is in a specific place at a specific time. That is a scheduling problem, and scheduling problems are solved by conversation, not by notification.
How the process runs#
The logic is the same one we use for medical appointments: the goal is not to collect confirmations but to find out early which visits will not happen, so that the route can be changed before the van leaves. In practice it looks like this:
- The schedule arrives from the metering team or the contractor, by file transfer or API: address, time window, type of visit, contact number.
- First call ahead of the visit, early enough to replan. The customer hears who is calling and why, including that the caller is an AI assistant on behalf of the energy company, then the date and window of the visit, and is asked whether someone will be there to give access.
- Easy rescheduling. If the window does not work, the customer can choose another within the rules agreed with the field team, or the request is passed back with their preferences.
- Retries for the unreached, within the calling windows and attempt rules agreed for the process.
- Text message as fallback where calls have not connected, with the visit details and how to change them.
- Daily statuses back to operations: confirmed, moved, cancelled, unreached, uncertain. The route planner can drop or swap visits before the technician sets off.
Speed matters here too. Because conversations run in parallel, the whole list for the coming days can be worked in hours rather than days, including repeat attempts. That gives the field team the most time to replan.
The conversation can also carry practical information the technician needs: where the meter is, whether there is an intercom code, whether a dog will be loose in the garden. Every item is optional and approved in advance, because the call must stay short. A customer who wanted a two-sentence reminder and got a questionnaire will not pick up next time.
Why a conversation, not another letter#
Most utilities already notify customers about visits, by post, email or text. The notification is not the problem. The problem is that a notification is one-way: it tells the customer when the technician will come, but it tells the company nothing about whether anyone will be there. Silence after a letter means nothing. Silence after several call attempts and a text is at least a status the route planner can work with.
A conversation also handles the situations a notification cannot. The customer works shifts and can only do mornings. The flat is rented and the landlord holds the key to the meter cupboard. An elderly customer wants to know that the person at the door is genuine and what identification they will carry. Each of these is a reason a visit fails, and each can be settled in a short call with answers approved in advance and a route to a person at the energy company when the question goes beyond the script.
There is a trust dimension as well. Customers are rightly cautious about strangers asking to enter their homes, and fraudsters do impersonate utility staff. A call that clearly identifies the company and the AI assistant, gives the customer the visit details they already received in writing, and never asks for payment or personal data beyond what the process needs, reinforces the legitimate visit rather than adding to the noise. We design these calls with the client's security and communications teams, not around them.
One service, several processes#
Visit confirmation rarely stays alone. Energy companies that start with one outbound process tend to find the same infrastructure useful for others, because the customer base, the integration and the compliance set-up are already in place. In the sector, the natural neighbours are:
| Process | What the call achieves |
|---|---|
| Field visit confirmation | Access confirmed or visit moved before the technician drives out |
| Satisfaction surveys | CSAT or NPS after a hotline or point-of-sale contact, with complete answers |
| Collections | Early contact on overdue bills, with payment status and next steps |
| Contract renewals | Offers such as new rates or a fixed price for a period, before the contract ends |
Processes in the sector that Vee's managed service covers; scope differs by client.
Surveys are where we have a published reference for the sector. In a CSAT survey for an energy company, we measured how many customers completed the whole survey at the first attempt. The thirty-first call of the day starts with the same energy as the first, and it shows in complete answers.
We do not have a published figure for field-visit confirmation, and we would not borrow one from another process. The effect depends on the share of visits that need access, how far ahead they are scheduled and whether the field team can actually replan once it knows. What we can say is what to measure.
What to measure#
- No-access rate for confirmed visits compared with a comparable set of visits that were not confirmed, on the same visit types and regions.
- Lead time of cancellations: how many days before the visit the no-access risk was discovered, and whether that was enough to replan.
- Repeat visits avoided, measured in the contractor's records, not the contact centre's.
- Downstream effects: fewer estimated bills and fewer related queries on the inbound line where readings were the issue.
The last point is easy to miss. The benefit of a confirmed reading often shows up months later in a quieter complaints queue, in a different department, on a different budget line. If nobody connects the two, the confirmation process will look like a cost.
What to ask before you start#
- Which visit types fail most often for lack of access, and what does each failure cost including the repeat?
- Can the route planner or contractor act on a status received a few days before the visit?
- Who owns the outcome: metering, operations, customer service or the contractor manager?
- How will the schedule reach the calling process, and how will statuses come back?
- What will the customer be able to change in the call, and within which rules?
- Against what comparison will the no-access rate be measured?
The technology for this is not the hard part. The hard part is getting metering, the contractor and customer service to agree on one status and one owner, and then keeping the process tuned as the visit programme changes. That is the work we take on for energy companies: a short call, made early, so that the van only goes where the door will open.
Follow The First Second
Occasional issues on the mechanics and economics of AI-powered customer contact. No noise, no vendor news.
Follow via RSS