Home · Solutions · Other solutions
Solution · Other solutionsThe chase list becomes an exception list, and payroll questions stop needing a ticket
Timesheets complete before the payroll cut-off, every month
A robot checks completeness and approvals daily against the payroll calendar and nudges the right people in Teams; an agent answers the routine questions with a citation.
Executive summary
Two specialists spend the end of every month writing reminders, and payroll still pays on estimated hours.
Two things are built and they share a calendar.
Fewer estimated payments and off-cycle corrections, because sheets are completed before cut-off rather than reconstructed after.
the provider input file and the payroll system; the HR case tool; the Power BI model
Business problem
Payroll
Timesheet compliance fails in the last mile. The time system records hours, but it does not know who is on leave, whose manager left in March, or that the cut-off moved for a bank holiday. Payroll fills that gap by hand: export the completeness report, match it against the org chart, write the reminders, then escalate. Every unapproved sheet at cut-off becomes an estimated payment or an off-cycle correction, each costing far more than the entry it replaces.
The second failure is information. Payroll rules are documented, but employees cannot find them, so they ask, and the same forty questions arrive every month in different words. Real problems, a wrong tax code, a bounced bank detail, sit buried among questions about the payment date.
It persists for a structural reason: payroll is a small team with an absolute deadline, and the people who create the work are not the people who do it. A manager who approves late pays nothing for it. The specialist who cannot take leave in the last week of the month pays every month.
How it works today
The pattern is the same whether payroll runs in-house or through a provider who needs a clean input file.
- PersonAn employee records hours in the time system, or forgets, and nothing tells anyone which
- SystemPayroll exports the completeness report a few working days before cut-off
- PersonA specialist compares the export with the headcount list and the leave calendar by hand
- PersonReminder emails go to everyone with missing hours, including several people who are on holiday
- WaitingManagers with unapproved sheets get a second reminder, then a Teams message to their manager
- Risk of errorCut-off passes with rows still open, so those people are paid on estimated hours or held
- PersonEmployees write to the HR mailbox about missing overtime and payment dates, the same questions reworded
- Risk of errorA specialist investigates, replies and logs a correction that lands in next month's run
Why the current process costs more than it appears
The cost grows where nobody is looking.
- Estimated payments are not the end of a story, they are the start of one, and every step of it is handled by the most expensive people in the process.
- Bulk approval at month end hollows out a control auditors rely on. A manager who clears forty sheets in one click has reviewed nothing, and the separation of entry, approval and payment becomes a formality.
- Key-person dependency is acute: the specialist who knows which manager responds to which kind of message cannot take leave when it matters, and cannot easily be replaced.
- Reminders that live in mailboxes and chats leave no usable trail, so who was asked, when, and what was decided has no answer that survives an audit.
Cost of inaction
Corrections compound. An estimated payment in month one produces a question in month two and a reconciliation line in month three, so the cheapest failure in the process keeps generating the most expensive work in it. Add an acquisition or a temp-staff population and the list grows faster than the team, because each new group follows rules only one or two people know.
None of this reaches a budget line. It shows up as overtime in a four-person team, as churn when the specialist who knows which manager to call leaves, and as an error rate the board hears about only once a works council raises it.
A plausible organisation with realistic proportions. The figures are there to be recalculated on your data; they are not a client result.
A pan-European engineering and technical services group with 2,600 employees in seven countries; hours are recorded in the time module of the HR platform and payroll runs monthly in-country, partly in-house and partly through a provider needing a clean input file.
About 18% of timesheets are incomplete or unapproved three working days before cut-off, roughly 450 chases a month; the HR shared mailbox takes around 720 payroll questions a month; around 60 corrections or off-cycle payments a month trace back to late timesheet data. A payroll team of four handles all of it alongside the run.
A completeness report is exported and reconciled by hand, reminders go out by email, escalation happens in Teams chats, and at cut-off the remaining sheets are estimated or held.
Four minutes per chase with the follow-up, nine minutes per question, and 35 minutes per correction, concentrated in the last three days of the month.
A robot checks completeness and approvals daily against the payroll calendar, cross-checks master data and leave so reminders reach active employees and current managers, nudges employees in Microsoft Teams and sends managers an actionable approval notification; a conversational agent grounded in the calendar and guidelines answers routine questions with a citation and turns personal ones into tracked cases.
In the modelled case most sheets are complete before cut-off rather than after, corrections fall, and the last three days of the month stop being a chase. The arithmetic is a model and the company is invented.
Proposed solution
Two things are built and they share a calendar. The first is a completeness robot. From an agreed number of days before each cut-off it reads timesheet and approval status daily through UiPath Integration Service, or through UiPath SAP automation activities where the time system has no usable API, and cross-checks HR master data and leave. That cross-check separates a generic reminder from a useful one: nobody on holiday is chased, and nothing goes to a manager who left in March.
Employees with missing hours get a Microsoft Teams message linking straight to their own sheet. Managers with unapproved sheets get a UiPath Action Center actionable notification in Teams, one sheet at a time with employee, period and hours in view, which is a different act from clearing forty rows at month end. Escalation follows a rule set payroll writes and owns, and every reminder and decision is logged in UiPath Orchestrator. At cut-off the robot produces two artefacts: the exception list and the provider input file.
The second build is a payroll answers agent, running as a UiPath conversational agent in Teams and grounded through UiPath Context Grounding in the payroll calendar, guidelines and FAQ held on SharePoint. It answers when and what-is-the-rule questions with a citation, and shows an authenticated employee the masked bank details on file. It does not read individual pay records. Anything personal becomes a structured payroll case with employee, period and evidence attached, routed to a specialist.
Power BI holds what management could not see before: completeness by day before cut-off, chase volumes by department and manager, corrections, and the themes behind the questions.
UiPath Orchestrator triggers, queues, credential store and audit; UiPath Action Center actionable notifications in Microsoft Teams; UiPath conversational agents deployed as a Teams app; UiPath Context Grounding indexes over SharePoint; UiPath AI Trust Layer model allow-list, PII masking and interaction logging; Power BI
The completeness robot and its master-data cross-check, the escalation rule set, the reminder and notification content, the cut-off exception list and the provider input file, the agent's topics, hand-over rules and grounding index, and the Power BI model
Time system and HR platform through UiPath Integration Service connectors or SAP automation activities; the HR case tool for payroll cases; Microsoft Entra ID for employee authentication and group-based access
How the automated process works
- AutomationFrom an agreed number of days before cut-off, the robot checks completeness and approval status daily against the payroll calendar
- SystemMaster data and the leave calendar are cross-checked, so nobody on holiday is chased and no message goes to a manager who has left
- AutomationEmployees with missing hours receive a Teams reminder with a link straight to their own sheet
- PersonManagers act on an Action Center notification in Teams, approving one specific sheet with the period and hours in view
- AutomationEscalation follows the rule set payroll owns, and every reminder and escalation is written to the log
- PersonAn employee asks a question in Teams; the agent answers from the calendar or the guidelines with a citation, or opens a payroll case
- AutomationAt cut-off the robot compiles the exception list and the provider input file
- PersonPayroll decides what happens to each remaining exception: estimate, hold, or pay
Human-in-the-loop model
Automation handles
- The daily completeness and approval check against the payroll calendar and the leave data
- Targeted reminders in Teams, actionable notifications to managers, and escalation per the rule set
- Answers to calendar, guideline and FAQ questions, each with a citation to the approved document
- Payroll cases created with the employee, period and evidence attached, plus the cut-off exception list and provider file
People decide
- What happens to each exception at cut-off: estimate, hold, or pay
- The resolution of an individual payroll case and any corrective payment
- Timesheet approval itself, which stays with managers and is never automated
- Changes to payroll rules, the calendar and the guideline content the agent is grounded in
Before and after
Systems and integrations
Where a rule suffices we do not use a model. Where judgement is needed, a person decides.
Inputs
- the payroll calendar and guidelines on SharePoint
- timesheet and approval status from the time system
- HR master data and the leave calendar
Automation layer
- UiPath Orchestrator
- UiPath Robots
- UiPath Integration Service
- UiPath Agents with Context Grounding
- UiPath Action Center
Target systems
- the provider input file and the payroll system
- the HR case tool
- the Power BI model
Human touchpoints: the Teams reminder to the employee; the Action Center approval notification to the manager; the agent chat; the cut-off exception list
Technologies used
run the daily check, the escalation rule set, the schedule and the audit trail
Areads timesheet and approval status and HR master data through the platforms' APIs
Aactionable notifications so a manager approves one specific sheet inside Microsoft Teams
Aanswers routine payroll questions and opens a tracked case when one is personal
Aindexes the payroll calendar, guidelines and FAQ on SharePoint so answers cite an approved document
Areminders, manager notifications and the question channel, in the tool people already have open
Aauthenticates the employee before a masked bank-detail lookup and drives group access
Acompleteness by day before cut-off, chase and correction volumes, question themes
AIllustrative economic model
Start by questioning the assumptions.
Sixty-five per cent is the share of this desk work the robot and the agent absorb, and it is folded into the volume rather than applied afterwards. Three streams feed the total: 450 timesheet chases at 4 minutes, 720 payroll questions at 9 minutes and 60 corrections at 35 minutes, together 1,230 touches and 173 hours a month for four specialists at €36 an hour fully loaded. The calculator therefore counts 800 touches a month at a blended 8.4 minutes rather than 1,230 at three rates. Manager time, employee time and provider re-run fees sit outside the model, and nothing in it was timed at a client.
Run the numbers on your data
An illustrative estimate from your own inputs. It models released capacity; it is not a promise of savings.
Business benefits
- Fewer estimated payments and off-cycle corrections, because sheets are completed before cut-off rather than reconstructed after
- The last three days of the month stop being a chase, because a robot has worked the list daily since it opened
- Payroll questions are answered in minutes without a ticket, from the calendar and guidelines a specialist would have quoted
- Approval becomes a real control again, because a manager acts on one specific sheet instead of clearing forty at month end
- The trail of who was asked, when, and what was decided exists without anyone maintaining it
The management view
- Completeness by department and manager is on screen each morning before cut-off, with the open escalations beside it
- Chase volume, corrections and question themes become monthly evidence, turning a confusing overtime rule or a manager who never approves into a fixable root cause
- Workload becomes predictable, because the team plans around an exception list, not a chase list
- Adding a country or a shift model stops multiplying manual work, since the rules are configuration rather than one person's habit
Board-level KPIs
Security and governance
Control is not an add-on.
- The robot's service account reads timesheet status and master data and reaches no pay amounts; its secret is held in the Orchestrator credential store or Azure Key Vault
- The agent answers from approved documents only. Pay questions are never answered from payroll records; they become cases, and the bank-detail lookup returns masked values to the authenticated employee alone
- Model calls run under the UiPath AI Trust Layer with an allow-list, PII masking and a logged interaction record, in the EU region of UiPath Automation Cloud, with Microsoft 365 content inside the EU Data Boundary
- Reminders, escalations and cut-off decisions are written to Orchestrator; Teams and SharePoint activity remains auditable in Microsoft Purview
- Managers stay the approvers, so entry, approval and payment remain separated and the control an auditor tests is strengthened rather than bypassed
Why now
Experienced payroll specialists have become slow to hire across European markets, so capacity has to come from removing work rather than adding people
Time systems and HR platforms now expose status and master data through APIs, so a daily completeness check no longer means a screen scrape that breaks on the next release
Works councils and regulators are asking harder questions about pay accuracy, and a logged, documented chase process answers them more easily than a mailbox does
Relevant executive roles
Fewer off-cycle payments, better pay accuracy, and an approval control that stands up when an auditor tests it
A calm cut-off, an exception list instead of a chase list, and leave that can be taken at month end
Employees get an answer about their pay in minutes, and the questions that are real problems surface rather than sink
Volumes become measurable and the process scales across countries without the chase list scaling too
Common questions and objections
Usually one generic reminder to everyone with missing hours, people on leave included, and nothing to managers. What is added is targeting, escalation that follows a rule you wrote, and an exception list at cut-off.
It answers only from approved documents and cites the source, and it never reads individual pay data. Anything personal becomes a tracked case for a specialist, which is more traceable than the mailbox reply it replaces.
The provider needs a clean input file on time, and the chasing that produces it happens on your side. This builds that file and measures the provider's error rate against your own data.
When this is not the right solution
- A few hundred employees, one payroll and a manager who knows everyone by name; a calendar reminder may genuinely be enough
- Timesheets are not a payroll input because pay is fixed with no variable element, in which case only the answers agent is relevant
- The time system is being replaced within the year, so the checks should be built on the successor
- No maintained payroll guidelines exist; write them first, because an agent grounded in stale documents is worse than none
A question for the next management meeting
Move next month's cut-off three days earlier: how many people in our payroll team would notice, and how many managers would?
Implementation approach
We start with one slice of the process and extend only once it is proven.
We deliver
- Analysis of one cut-off cycle and one month of mailbox questions, mapping who chases whom and which questions repeat
- The escalation rule set, written with payroll and HR and signed off before anything goes out
- The completeness robot with its master-data and leave cross-check, the reminders and manager notifications
- The cut-off exception list and the provider file, tested against a past payroll period
- The agent: grounding index, topics, hand-over rules and answer testing on real questions
- A pilot in one legal entity across two cut-offs, the robot first in shadow mode and then live
We need from you
- Test access to the time system and HR platform, and a current payroll calendar
- Payroll guidelines in final form, and an owner who keeps them current
- A payroll specialist as product owner and one legal entity for the pilot
- Whatever works-council consultation applies to automated reminders in your countries
Stages
Discovery
One cut-off cycle, the chase list, the mailbox and the calendar, with payroll
Design
Escalation rule set, message content, exception-list format, agent topics and hand-over rules
Build
Completeness robot, integrations, Teams touchpoints, grounding index and the Power BI model
Shadow run
For one cut-off the robot reports what it would have sent, and the rules are corrected
Pilot
One legal entity live across two cut-offs; the agent starts with the twenty commonest questions
Scale
Remaining countries, further pay elements, the provider file and monthly theme reviews
Departmental. Effort is driven by the number of countries and pay rules, whether the time system exposes approval status through an API, and how current the payroll guidelines are.
Three days before cut-off, 312 rows and two people writing emails.
Send us last month's chase list and the twenty questions your HR mailbox answers most often. We come back with the share a robot and an agent would have closed before cut-off, and the escalation rules that would have done it.
Send us last month's chase listThe neighbouring process usually has the same problem
Payroll errors are found by the employee on payday, not by the team that built the file.
View solution HR & peopleContractor hours from timesheet to agency invoiceYou approve the hours in one place and pay the agency invoice in another, and nothing checks that the two agree.
View solution Other solutionsMissing itemisations and tax lines caught at submissionThe VAT on a hotel night is lost in the week the folio is wrong, not in the quarter the adviser finds it.
View solutionIndustries we deliver this in most oftenManufacturing & industryServices & ITFinance & insuranceShared services