Automation only earns its place in an MSP if you can prove what it delivers. Most MSPs have automations running in Rewst, but few can say how many tickets they resolve, how many hours they free up, or whether the team is actually working on higher-value tasks as a result.
In this post we walk through five KPIs that answer those questions. Each one is measurable with data you most likely already have in your PSA, your time registration and Rewst. Together they give you a clear picture of what automation is worth, and where to invest next.
1. Tickets handled through the Self Service Portal
The share of tickets that end users start and complete through the Self Service Portal is the clearest sign that your automation is customer-facing and not just a back-office tool. A request such as a password reset, a new mailbox or a software request that runs end to end without a technician is a ticket your team never has to touch.
How to measure it
- Count the tickets created through the portal or a Rewst form, and divide by the total number of tickets in the same period.
- Split the result by ticket type, so you see which request types are fully automated and which still need a human.
- Track the completion rate as well: how many portal requests finish without manual follow-up?
What good looks like: the share grows month over month, and the request types with the highest volume are the first ones to be moved into the portal. A flat line usually means users still call or email instead of using the portal, which is a communication problem rather than a technical one.
2. Time saved, and making sure Rewst reports it correctly
Time saved is the KPI that management cares about most, and the one that is most often wrong. Rewst can report the time saved by your automations, but only if the time saving is set correctly per workflow. A workflow without a value, or with a guess that nobody has checked, makes every ROI report unreliable.
How to get it right
- Measure the manual effort first. Ask a technician to perform the task by hand a few times and take the average, including the small steps around it such as looking up the customer, documenting and closing the ticket.
- Enter that value as the time saved per execution for the workflow in Rewst. Use minutes, and be conservative: a number you can defend is worth more than an impressive one.
- Review the values every quarter. When a workflow changes, the manual effort it replaces usually changes too.
- Check for workflows with no time saving set. They run, but they are invisible in your reports.
How to report it
- Hours saved per month, in total and per client.
- Hours saved per workflow, so you see which automations pay off most.
- The equivalent in full-time employees, which is easier to communicate to a client or a board than a number of minutes.
Time saved is an estimate by nature. Used consistently, it still shows the trend and the relative value of each automation.
3. Staff hours: worked, available, paid and billable
Automation changes how your team spends its hours, so you need a baseline of those hours before you can show the effect. Four measures give you that baseline. Definitions differ between companies, so write down yours and use them consistently.
| Measure | What it means | Where to find it |
|---|---|---|
| Worked hours | Hours actually worked by the employee, including internal work | Time registration |
| Available hours | Contract hours minus leave, sickness, training and internal meetings: the hours that can be spent on customer work | Time registration and HR planning |
| Paid hours | Hours for which salary is paid, including leave and public holidays | Payroll |
| Billable hours | Hours that are invoiced, or covered by a managed services contract | PSA time entries |
From these you can calculate two ratios that tell the real story:
- Utilization = billable hours divided by available hours. It shows how much of the time that can be sold is actually sold.
- Billable ratio = billable hours divided by paid hours. It shows how much of the payroll cost turns into revenue.
Why this matters for automation: when automation takes over repetitive work, technicians should spend more of their available hours on projects, advice and other work that creates value. Compare the ratios before and after you introduced a new automation. If time saved goes up but utilization does not, the freed-up hours are being lost somewhere, and that is worth investigating.
4. Automation and AI versus human: share of hours
This KPI answers a simple question: of all the work done on tickets, how much was done by automation and AI, and how much by people? Expressed as a percentage of hours, it makes the effect of your automation program visible in one number.
How to measure it
- Take the total effort on tickets in a period: the hours your technicians registered, plus the hours that automation and AI handled in their place.
- For the automated part, use the time saved per execution from KPI 2, multiplied by the number of successful runs. Count AI-assisted steps, such as ticket triage, summaries and suggested replies, in the same way.
- Divide the automated hours by the total, and report the result as a percentage.
Tips
- Tag tickets as fully automated, partly automated or manual. The middle category is where most of the growth is, so do not hide it.
- Report it per ticket category. Password resets may be close to 100 percent, while server incidents stay human work, and that is fine.
- Use the same method every month. The absolute percentage matters less than the direction.
This KPI depends on the time saved values being correct, which is the reason to get KPI 2 right first.
5. Automations in use per endpoint
An automation that nobody uses saves no time. Counting how many automations touch each endpoint, such as a workstation, a user or a tenant, shows where your automation coverage is strong and where it is thin.
How to measure it
- Count the distinct automations that ran at least once against an endpoint in the last 30 or 90 days.
- Report the average per endpoint, and the distribution: how many endpoints have none, a few or many?
- Compare clients. A client with a low number usually has a gap in onboarding, in monitoring or in the service catalogue.
What to do with it
- Endpoints with zero automations are your quick wins. Standard crates for onboarding, patching and offboarding usually cover them.
- A high number on a few endpoints and nothing elsewhere points to automations that were built for one client and never reused.
- Track the trend per client when you roll out a new crate, and you see within weeks whether adoption is real.
Putting it together
Five KPIs, one story: the Self Service Portal share shows adoption, time saved shows the value, staff hours show where that value lands, the automation and AI share shows the shift from manual to automated work, and automations per endpoint show your coverage.
Our advice is to start small:
- Fix the time saved values in Rewst first, because KPIs 2 and 4 depend on them.
- Put all five numbers on one dashboard, and review it monthly with the team.
- Set one target per KPI for the next quarter, and pick the KPI with the biggest gap first.
Want help setting this up in your own Rewst environment? Flownetic builds the crates, the reports and the dashboards for MSPs. Get in touch and we will look at your numbers together.