A small business often works hard while routine work quietly expands: a customer repeats the same information, a handoff sits in an inbox, or a new tool produces a draft someone must repair. Working smarter means finding one repeated source of delay or correction and changing it in a way you can test. The starting point is a real customer request, the people who handle it and the time from beginning to end. Choose one part of that path rather than redesigning the company in a single meeting.

Watch one job travel through the business

Take a recently completed job with a clear beginning and end. For example, a small building contractor might start when a customer requests a quote and finish when the agreed work is completed and confirmed. List the steps that occurred: capture the request, inspect the job, prepare the quote, obtain approval, book labour and materials, carry out the work and check completion. Note where information changed hands, was requested again or waited for someone. Ask the people doing the work what actually happened, not only what a written procedure says.

Write down the elapsed time between first contact and finished service. Also count corrections, repeat calls and handoffs. A two-minute approval step may sit unowned for two days, so measuring only active minutes can miss the problem. A second check may be an important safety or quality control, so do not remove it simply because it appears twice on a drawing. ASQ describes flowcharts as a way to see the sequence of actions, decisions and people and use that picture to understand a process. A numbered list or a few cards on a wall are sufficient for a first look.

Ask a specific diagnostic question: “Which repeated interruption would make the next five jobs easier if it happened less often?” You might discover that customers are asked to repeat a delivery address, that a technician receives an incomplete brief or that a manager rechecks work because the first check is not trusted. Name the problem in observable terms. “Improve communication” cannot be measured. “Reduce jobs paused for missing access instructions” can.

Choose between the two topic areas

Fix Everyday Workflows is the route for handoffs, waiting, duplicate work and repeated corrections. Its guides show how to map a process in thirty minutes, find and remove needless copying while retaining essential checks, and run a fifteen-minute weekly improvement meeting. Choose it when the people already have the tools to do their jobs but the path between them is unclear. A better booking field or assigned owner may be a more useful first change than new software.

Practical AI and Automation is the route for a specific repeated task that might be assisted by a digital tool. Its guides help choose a first low-risk AI task and test an automation on a small sample before broader use. Choose it when you can define a clear input, output and human review, and when the data may be used under your contracts and local rules. A tool can save time, but review, correction, subscription and setup are part of the real cost.

You may move from the first area to the second. Map the work before automating it; an unclear handoff remains unclear when it happens faster. Conversely, a simple repeated transfer with a clear rule can make a good limited technology trial once the quality checks and fallback are defined. Neither path requires a large project. Begin with the one source of friction that appears most often in your own cases.

Test one change on the next few cases

Assume a local contractor reviews ten comparable maintenance jobs and finds that four required an extra call because the customer's access information was missing. The owner asks the person who accepts bookings to confirm access details during the first conversation and place them in the existing job record. The next ten jobs are measured with the same rule: count extra calls caused by missing access details, elapsed time from request to confirmed booking and any wrong instructions given to technicians. The numbers are illustrative, and a ten-job sample is a first check rather than proof of a lasting effect.

Suppose the extra calls drop from four to one but the remaining case involves an urgent booking taken through a different channel. The owner can test the same check in that channel instead of announcing a company-wide success. If extra calls fall but technicians receive incorrect codes, the method of confirming and updating information needs revision. A good change should improve the chosen measure without producing a new problem. The Institute for Healthcare Improvement's testing guidance describes this as planning, trying, studying and adapting a change in its local setting.

Choose the measure before the trial. You could track time from enquiry to booking, duplicate entries per job or calls for missing information. Pair it with one safeguard: corrections, missed appointments, customer complaints or staff time elsewhere. Record the date the change began, who owned it and any unusual jobs. Use a similar set of cases before and after when possible. If the team handles only one relevant case per month, extend the observation period rather than drawing a conclusion from one outcome.

Keep the work human and accountable

A short regular review turns observations into action. Ask what happened to the measure, whether the change was actually used, what exceptions occurred and whether the people doing the work want to adapt it. A change that only one employee can operate may be fragile. Write down where the updated information lives and who maintains it. If a tool sends messages or changes records, identify who can intervene and how the old method works when the tool fails.

Not every repeated action is waste. Independent approval can protect against fraud; a completion check can protect customer safety. Data-sharing changes may affect confidentiality and privacy. Contract terms, employment responsibilities and regulated checks differ by location. Seek qualified help before changing a safety-critical, security-sensitive or legally required step. Likewise, a serious incident needs immediate attention through the appropriate route, not a place in next week's improvement list.

A small firm does not need a fixed benchmark for the right number of handoffs or minutes per job. It needs its own baseline and a clear decision about whether the next change helped. Next step: follow one recently completed request from first contact to delivery. Record the longest wait or most frequent correction, choose Fix Everyday Workflows or Practical AI and Automation, and test one reversible change on the next five similar cases.

Sources