AI for service businesses: planning usable office capacity

A growing service business can become constrained by office work even when demand is there. Every additional enquiry needs a customer connection, a next step and the right information. If the owner, office and delivery team repeatedly pick up the same task, coordination grows along with demand. AI becomes economically interesting when the business can complete the same reliable workflow with less of its own handling time. Whether that creates room to serve more customers depends on the work that remains and the capacity to deliver the actual service.

Comparing a faster draft with the previous writing time is not enough. Customer support also involves missing information, scheduling, checks and exceptions. This article develops a capacity calculation for a fictional service business and connects it to a specific office workflow. Every volume and time is an invented modelling assumption, rather than a typical result or a webRichtung promise. Apply the method using your own observations before committing additional customers to the schedule. The decision should reflect complete work, including the less convenient parts.

Identify where growth adds office work

Start with a completed working week. What did the office actually do for a handled case? Reading the enquiry, identifying the customer, checking the service scope, arranging an appointment and preparing the right document may all belong together. Counting them separately can make each tool look efficient while the overall case remains unfinished. A useful unit is therefore a clearly defined case handled through to its agreed result. Repeated contacts about the same request belong in that case’s handling time, rather than appearing as additional completed work.

In the example, the fictional Nord Service coordinates recurring on-site services for business customers. Lea handles the office, owner Deniz decides on unusual service requests, and the delivery team performs the work. A case in this calculation starts with a specific customer service request. It ends when the next service appointment has been clarified and the team has the current document it needs. Delivering the service itself is planned separately. This boundary prevents spare office time from being presented as additional time available on site.

Lea finds that she repeatedly searches for customer information and service descriptions. Some appointment requests arrive before it is clear whether they match the agreed scope. Deniz then receives questions without the full context. The first improvement is consequently specific: take a normal service case from a reliable information base to a usable handover. If that path requires less internal time, it creates a concrete basis for more office capacity. The business can then decide how it wants to use the benefit, instead of assuming that every saved minute must become growth.

Explore the relevant platform areas and start with the office workflow that limits your growth.

Treat the enquiry, appointment and document as one workflow

For a normal case, Lea needs the correct customer relationship, the confirmed service scope and the actual appointment status. A preferred time is not yet a confirmed booking. Finding a service description also does not establish that this particular version belongs to the order. Those connections must remain visible through the workflow. A colleague taking over should understand the distinctions from the material provided. If Lea has to explain everything orally again, the effort has moved elsewhere rather than disappeared from the business.

The webRichtung platform page describes areas for conversations, documents, appointments and text tasks. Documents supports finding information in documents; Assist supports working on texts and tasks. These functions can help with individual steps of the workflow being examined. Check which data and actions are actually connected for your business in the intended use. The process described here does not mean that a complete automated chain is already available without configuration. The value must be established in the working arrangement that your team will really use.

Appointment planning distinguishes released availability from whether the proposed service is appropriate. According to its module description, Calendar offers appointment management and public booking based on availability you release. That does not imply route optimisation or an automatic decision on unclear service scope. Nord Service starts with cases that clearly fit its existing service. Special requests go to Deniz before they become binding delivery plans. The reduced office workload is thus tied to a result the team can actually use without relying on an invented promise.

Improve one recurring case first

Lea selects appointment preparation for an already agreed standard service. She defines which current service description is needed and where incomplete information should become visible. In a normal case, she finds the document, prepares the scheduling step and checks the final handover. An additional requested service remains explicitly open. It must not appear as already agreed simply because a draft message reads smoothly. The new workflow needs to preserve precisely this distinction, since otherwise it could save preparation time while creating correction work for the delivery team.

During the trial, a colleague takes over using only the supplied information. They can identify the customer, the applicable service and whether the appointment is confirmed or still unresolved. If the link to the current document is missing, Lea improves the preparation. If a special request is indistinguishable from a commitment, the wording is corrected. The trial therefore tests the usability of the complete handover. A polished message alone is insufficient when the recipient cannot continue the work correctly from what it says.

Deniz limits the introduction to this case type. New services, complex exceptions and other customer groups stay outside the initial calculation. Preparation and training receive their own time in the implementation plan. They do not become invisible simply because routine work may later take less time. After the first run, the office records actual handling time, including checks and rework. Only then can it judge whether repeating the new workflow uses less internal capacity at a comparable level of quality. A single well-prepared demonstration cannot answer that question.

Calculate today’s capacity completely

For the numerical example, Lea has 24 hours a week available for this area of work. Three hours remain reserved for unexpected work. That leaves 21 hours, or 1,260 minutes, for planned activity. The reserve will not be counted again later as a growth benefit. It deliberately stays uncommitted. Other office responsibilities outside this area are not included here. In your own calculation, first check that the time you call available is genuinely free from those other obligations, or the starting capacity will already be overstated.

In the model, Lea currently handles 60 comparable cases each week. Each needs an average of 18 minutes of internal handling, including questions and review. Sixty multiplied by 18 gives 1,080 minutes, or 18 hours. There are also three hours of fixed coordination for this work area. Today’s total is therefore 21 hours. Planned capacity is fully used, leaving only the deliberate reserve. Further regular cases would have no scheduled space without changing the workflow or the time available to it.

This starting point says nothing about typical office productivity. It simply makes the conditions behind the decision visible. A business with a different case mix, more questions or less available time will get a different answer. Do not count only the quickest cases. An average covering complete normal and exceptional handling is more useful for planning than one successful demonstration. Record the observation period as well. Later comparisons should not accidentally place different service scopes or different operating conditions side by side and call the difference a productivity gain.

Calculate additional capacity after introduction

Suppose, hypothetically, that the same case type now requires an average of eleven minutes of internal handling. Checks, questions and rework remain included. The existing three hours of fixed coordination continue. Two extra hours a week are allocated to maintaining and supervising the new workflow. For 60 cases, that means 660 minutes of case work, plus 180 minutes of coordination, plus 120 minutes of new supervision. The total is 960 minutes, or 16 hours. The separate reserve is still outside this calculation, just as it was before.

Compared with the 21 planned hours, five hours become available: 300 minutes. At an unchanged eleven minutes for each additional case, at most 27 further whole cases fit mathematically. The theoretical total ceiling is 87. Subtracting the 300 fixed minutes from 1,260 leaves 960 for case work; dividing by eleven accommodates 87 complete cases with three minutes remaining. This ceiling leaves almost no flexibility within planned time. It is a boundary of the model, rather than a recommendation to accept exactly that workload immediately.

Deniz initially plans a total of 85 cases, which is 25 more than before. Eighty-five multiplied by eleven gives 935 minutes. Add the unchanged 300 fixed minutes to reach 1,235 minutes, or 20 hours and 35 minutes. Twenty-five minutes remain within the 21 planned hours, and the separate three-hour reserve stays intact. Under its assumptions, the calculation supports this limited expansion of office work. It does not claim 25 additional customers or a corresponding revenue increase. One customer may generate several cases, and the service still has to be delivered.

Check whether the released time is usable

The calculation is useful only when the released minutes fit the actual working day. Small reductions between unpredictable interruptions cannot automatically become an additional fixed customer support block. Lea therefore observes when work arrives and who can handle it. If the extra cases are expected precisely during already occupied peaks, the workflow needs a different distribution. A weekly total can hide this constraint. Check the real working windows before converting the calculated case volume into customer commitments, especially when an appointment requires several people to be available together.

Deniz and the delivery team must also be able to carry the additional workload. Less searching in the office does not create another skilled person on site. If more exceptions now reach Deniz, his decision-making window could become the next constraint. The assumption of constant coordination time would then be wrong. Expansion is plausible only when the next stage of service and the necessary specialist review are available. Office capacity is one part of the decision, rather than a complete statement of the company’s ability to deliver.

Released time does not have to support growth immediately. The business could use it to clear a backlog, improve handovers or create predictable time for existing customers. Those are economically relevant aims without being labelled additional revenue. Define the intended benefit before the evaluation. If the first benefit is more reliable support for current customers, observe that specifically. An unchanged staffing level by itself proves neither sufficient capacity nor a successful new workflow. The question remains what the team can complete, with what reliability, in the time actually available.

Test the calculation with more rework

A second hypothetical outcome shows how sensitive the result is. If cases actually need 14 minutes after introduction rather than eleven, 60 cases require 840 minutes of case work. Including the unchanged 300 fixed minutes brings the total to 1,140 minutes, or 19 hours. Only 120 minutes remain inside the 21 planned hours. They accommodate eight further whole cases at 14 minutes each, giving a total of 68. More rework substantially changes the manageable expansion, even though the workflow is still faster than the original 18 minutes per case.

The planned 85 cases would instead require 1,190 minutes of case work and 1,490 minutes in total at 14 minutes each. That exceeds the 1,260 planned minutes by 230 minutes, or three hours and 50 minutes. Even using the entire separate three-hour reserve would not cover the overrun. Under those conditions, Deniz should not retain the same volume and hope that daily work will somehow improve. He needs to reduce the workload, address the causes of rework or organise genuinely additional available time.

Specific findings are more helpful here than broad product judgements. Incorrect document versions need a better source; unclear customer requests need better clarification; overly restrictive approval rules can create unnecessary questions. After correcting the cause, examine the same case type again. Actual software costs also belong in the economic decision and should be established from the applicable terms. This capacity comparison inserts no product price and claims no automatic cost recovery from valued time. Time released and money saved remain separate questions until the business has evidence connecting them.

Decide on a verifiable expansion

For Nord Service, the sensible decision in the first model is to continue the new workflow and test no more than the specifically planned expansion, provided the delivery team and owner can support it. Value comes from complete handling in usable time. If quality and actual effort remain stable, the next workload can be recalculated. If observations differ, the plan changes. One-off setup remains separate from the recurring weekly comparison and must be scheduled additionally during introduction. It cannot be funded with time that becomes available only after the workflow is established.

On the platform page, you can explore the areas relevant to your particular constraint and use the existing account creation route. Before starting, describe the single case, its completed result and your available time. This turns the wish to serve more customers into a testable business step. The best first application is one whose benefit your team can recognise across a complete case and use when making its next capacity decision. That gives expansion a practical basis beyond the speed of one isolated task.

Choose the relevant use on the webRichtung platform and create your account for a limited, measurable start.