ComputerEase integration options: imports, API, report books, ODBC and RPA
Five ways data moves in and out of ComputerEase, and when each one is the right tool.
Last updated October 2026
The short answer
To read data, use an ODBC extract for anything you analyze or report on regularly, the API for live lookups of jobs and vendors from another system, and report books for one-off questions.
To write data, use the API where it accepts what you're sending, generated import files for batches like purchase orders and AP invoices, and a robotic process automation (RPA) bot only for the gaps, ideally just to press the Import button on files your systems already built.
Decision table
| Method | Direction | Use it for | Avoid it for |
|---|---|---|---|
| Report books | Read | One-off questions, accountant-formatted reports, a small audience | Anything recurring or shared widely: it's manual and stale the moment it's sent |
| ODBC extract | Read | Everything transactional: job cost, POs, AP, AR, inventory, GL. Dashboards and AI | Real-time needs (it runs on a schedule, usually nightly) |
| REST API | Read and some writes | Live job, vendor and employee lookups from a CRM; posting time, expenses and subcontracts | POs, AP, the GL and inventory quantities, which it doesn't expose |
| File imports | Write | Batches of POs, receipts, inventory adjustments, AP invoices matched to POs, budgets, vendor pricing | Anything that needs a decision mid-entry |
| RPA (Power Automate Desktop) | Write, sometimes read | Screens no import or API covers; clicking Import on generated files | Bulk reading (ODBC is faster and sturdier) or anything an import file can do |
Report books
Report books and the query tool are what most ComputerEase shops already use: someone runs a report and emails it, or exports it to Excel. They're the right answer when a controller needs a one-time view, or the report format itself matters, like something going to a bank or a bonding company.
They break down when the same report runs every week for several people. Every copy goes stale the moment it's sent, nobody can drill into it, and the person running it becomes the bottleneck.
ODBC extracts
ComputerEase ships with an ODBC driver that can read most of its tables: jobs, job cost detail, purchase orders, AP vouchers and checks, AR invoices, inventory and the GL. A scheduled extract on the ComputerEase server copies those tables out every night into a cloud database, where Power BI and AI agents can use them without touching ComputerEase during the day.
- The driver is 32-bit and lives on the ComputerEase server, so the extract runs there. On a Deltek-hosted dedicated server that means admin access to the server.
- Use a dedicated login for the extract with data-file access to the modules you need. Check its GL access code, job access code and checking-account password groups: they can filter rows silently, so an extract can look complete and still be missing data.
- Leave sensitive data behind at the source: payroll, Social Security numbers, tax IDs, bank and routing numbers.
This is the route for anything you'll report on, reconcile or ask questions of. More on ODBC and Power BI.
The REST API
The ComputerEase API is good at master data and a handful of writes. It reads jobs, job totals, vendors, employees and the item master, and it can accept timesheets, expenses with receipts, and subcontracts directly. It's the right tool when another system, like a CRM or a field app, needs to look up a job or vendor without waiting for a nightly extract.
It doesn't expose purchase orders, AP, the GL or inventory quantities, so it can't feed financial reporting on its own. Full breakdown of what the API covers.
File imports
ComputerEase can import files for many of the transactions people type in by hand: purchase orders, receipts, inventory adjustments, AP invoices matched to PO lines, vendor pricing, budgets and bank reconciliation transactions. The API's import service also accepts JSON or XML for POs, job transactions, vendors, jobs, AP and AR invoices and change orders, staged for someone to review and import.
- Generate the files from the system where the work starts, like an estimating or field app, so nobody retypes them.
- Keep a person in the loop: imports land as a batch someone reviews and releases, which is a feature for anything financial.
- Watch PO numbering. A PO number that collides with an existing one can stall a PO import, so generate numbers from one source of truth.
RPA bots (Power Automate Desktop unattended)
Robotic process automation means a software robot signs in and clicks through ComputerEase screens the way a person would. Microsoft's Power Automate Desktop can run these unattended on the ComputerEase server.
RPA works for the screens no import file or API covers, and for the final step of an import, where a bot opens the import screen and runs files your systems already generated.
RPA has three drawbacks:
- It's brittle. A ComputerEase update that moves a field, a slow screen or an unexpected pop-up can stop it or, worse, put data in the wrong place.
- It needs an unattended bot license, a Windows session on the server, and monitoring, because a failed run is otherwise silent.
- Reading data through screens is slow and fragile compared to an ODBC extract.
The rule we follow: never use a bot to type what an import file can carry, and never use a bot to read what ODBC can extract.
Combining them
Most contractors end up with a mix. Here's a common setup:
- Nightly ODBC extract into a cloud database for dashboards, reconciliation and AI agents.
- The API for live job and vendor lookups from the CRM.
- Generated import files for POs and job setup, with a short RPA step that only runs the import.
- Report books for the occasional one-off.
We build the read side, landing your ComputerEase data in your own Microsoft cloud, then dashboards and AI on top. See the packages.
Not sure which mix fits your shop?
We'll tell you on a free 30-minute fit call.