Data to AI™ / Copilot/Personio
Give managers and finance a Teams agent that answers who is out, how headcount splits across cost centers and which schedules people work, from current Personio data.
Better answers, fewer tokens.
Microsoft 365 Copilot
Which departments took the most approved sick leave days in September?
Approved sick leave in September added up to 96 days. Customer Support took 41 of them, more than Engineering and Sales together.
| Department Name | Days Count |
|---|---|
| Customer Support Absences | 41 |
| Engineering Absences | 23 |
| Sales Absences | 17 |
Days Count is summed over absences with an approved Status; Department Name comes from Employees, matched on Employee ID.
Personio is an HR platform for small and medium-sized businesses that covers personnel administration, time tracking and absence management. Dataddo extracts Employees, Absences and Attendances, so People teams can ask about headcount, time off and working hours in plain language.
@People Desk Which approved Absences in the Customer Support department overlap next week, by Start Date and End Date?
Team leads ask the agent in their Teams channel while planning shifts, instead of messaging HR or opening the absence calendar for every person.
@People Desk Count active Employees per Cost Centers Name, weighted by Cost Centers Percentage.
The finance controller asks Microsoft 365 Copilot during budget planning and gets the split that Personio holds, not a figure retyped from an old slide.
Use the dataddo tools to describe Employees and generate a test fixture of Work Schedule Name, Weekly Working Hours and Holiday Calendar Name.
An internal tools developer uses GitHub Copilot in Agent mode to keep the shift planning service's test data in line with real Personio schedules.
Data to AI™ is priced per seat, like the AI assistants your team already uses, with ~25% off on annual billing. Need tailored contracts or volume pricing? Talk to us about Enterprise.
Serve governed business data to Claude, ChatGPT and other AI tools over MCP.
$10.49/month
per seat
3 seats minimum
Flexibly deploy Dataddo in any cloud or hybrid environment.
Custom
We have a payment model that works for you
| Without Dataddo |
With Dataddo
|
Outcome for you | |
|---|---|---|---|
| Team availability | Managers ask HR who is off, or check absences person by person before they plan the week. | The agent in Teams reads Absences with Status, Start Date and End Date from the latest scheduled extraction. | Planning without chasing HR |
| Cost center split | Finance receives a headcount spreadsheet once a quarter and allocates shared employees by hand. | Employees returns one row per cost center with Cost Centers Percentage, so the agent splits people who work for several cost centers. | Allocation that matches Personio |
| What managers can see | A shared export reaches everyone it is forwarded to, salary columns included. | The agent can only answer from the Personio datasets and columns you extracted, so leave pay and personal columns out of the source. | Scope set at the source |
| Token consumption | Copilot spends tokens exploring raw Personio columns, sampling rows and retrying until it understands the data. | Field definitions and relations are supplied up front, so the model skips the exploration and goes straight to the answer. | Fewer tokens per answer |
| Answer quality | Prone to wrong joins, invented fields and numbers that are hard to check. | Answers grounded in defined fields. In the Dataddo benchmark, 78.5% of questions were answered correctly, against 65.5% with plain CSV files. | Better answers |
The Personio datasets HR and payroll teams use most, with their real field names. Copilot queries them by name through the Dataddo semantic layer.
| Dataset | Metrics and attributes |
Ask Copilot
|
|---|---|---|
| Employees |
|
How is headcount spread over each Cost Centers Name once Cost Centers Percentage is applied? |
| Absences |
|
Which sick leaves still have an open Certificate Status? |
| Attendances |
|
What is the average Break length per Employee over the last month? |
We asked 43 pre-registered questions of three real datasets - Google Search Console, HubSpot deals and Google Ads - and delivered the same rows to the same model in four ways. Only the data layer changed.
of questions answered correctly - the best of the four delivery methods tested
fewer confidently wrong answers - 14.0% of runs against 24.0% and 26.3%
more correct answers on messy CRM data - better on 8 questions, worse on none
Same model, prompts and rows in every condition; 759 graded runs, ground truth frozen before any run. On Google Search Console alone, documented CSV files did as well as Data to AI™ - the advantage comes from messier data such as HubSpot deals.
Read the benchmark →Create a Personio source in Dataddo with your client ID and secret, extract Employees and Absences, and attach it to a Copilot destination or AI Model. In Copilot Studio, add a Model Context Protocol tool for https://headless.dataddo.com/mcp-data with OAuth 2.0 Dynamic discovery, then publish the agent to Teams and Microsoft 365 Copilot.
Through a Copilot Studio agent. Add Dataddo Data Access as a Model Context Protocol tool with OAuth 2.0 Dynamic discovery, then publish the agent through Channels to Teams and Microsoft 365 Copilot.
From the server description you enter in Copilot Studio. Say it holds Personio employees, absences and attendances, so questions about people and time off trigger a Dataddo call.
No. The agent answers from the Personio flows attached in Dataddo, so decide which datasets and columns everyone using the agent may see before you extract them.
Employees returns one row per cost center. Someone split across two cost centers appears twice, each row with its Cost Centers Name and Cost Centers Percentage.
Yes. Add the Dataddo server to .vscode/mcp.json, start it, sign in and use Copilot Chat in Agent mode. Business and Enterprise orgs must enable MCP servers in Copilot policy first.
