Bison Report Fetcher Agent
Espie logs into Azure Virtual Desktop, walks through Pro Contractor by Viewpoint, and exports seven year-to-date reports one by one. Kurtis does the same thing weekly for InSQL over a Barracuda VPN—clicking through 6-8 reports with non-default parameter sets, saving each as Excel. Both flows are repetitive, brittle, and prevent Bison from running additional reports more frequently.
Espie or Kurtis picks the client and an optional date range from an internal page, clicks Run, and the agent fetches every configured report for that client and drops them in Google Drive. An email notification confirms the batch is ready. No more manual exports. No more clicking through Windows apps.
20-30 minutes every month clicking through Pro Contractor. Multiple hours every week in InSQL. Same seven reports, same manual steps.
Espie logs into Azure Virtual Desktop, walks through Pro Contractor by Viewpoint, and exports seven year-to-date reports one by one.
20-30 minutes per client per month spent exporting from Pro Contractor instead of analyzing the data.
Multiple hours per week pulling InSQL reports via VPN instead of moving upstream into analysis work.
Both systems have no APIs. Reports must be extracted manually from Windows apps behind RDP and VPN, one by one.
One button. Seven reports. Google Drive folder. Done.
One-button extraction replaces the 20-30 minute monthly Pro Contractor export cycle and the multi-hour weekly InSQL pull.
Kurtis: 'If we had the ability to automatically extract more reports, there's theoretically little things we could do for them that would help.' Automation makes it cheap to pull additional reports whenever clients need them.
Reports drop into Google Drive with clear names (report type + date) so Bison's existing analysis scripts and HTML generation can pick them up unchanged.
The agent logs into Pro Contractor and InSQL automatically, runs each configured report with the right parameters, exports each as Excel, and drops them into a per-client Google Drive folder named by report and date. Espie and Kurtis get an email when the batch is ready. No more clicking through Windows apps. No more manual exports. The time saved goes to analysis instead of extraction.
How the extraction works
The agent handles two different flows behind a single internal UI. Flow A pulls Pro Contractor reports from Azure Virtual Desktop. Flow B pulls InSQL reports over Barracuda VPN. Both land in Google Drive.
Espie or Kurtis selects the client (Pro Contractor or InSQL), picks an optional date range override, and clicks Run. Defaults are baked in: prior month-end for Pro Contractor, prior Sunday snapshot for InSQL.
For Pro Contractor (Flow A): agent connects to the Azure Virtual Desktop, opens Pro Contractor by Viewpoint, navigates to Reports. For InSQL (Flow B): agent connects to the Barracuda VPN from a Sagan-hosted Windows machine, opens InSQL from the file server.
Each report has its own parameter set (date range, output format, occasional popup option). The agent runs them in sequence, using the configured defaults or the operator's overrides.
Pro Contractor and InSQL both have 'save as Excel' dialogs. The agent clicks through them and stages the files locally.
All downloaded reports upload to a per-client Google Drive folder, named by report type and date. An email notification tells the operator the batch is ready.
Client selection +
Operator picks which client to extract from (Pro Contractor or InSQL).
Date range (optional override) +
Operator can override the default date range before running (e.g., pull last month instead of this month).
Pro Contractor login and Azure Virtual Desktop connection +
Bison has a dedicated Azure Virtual Desktop session for Pro Contractor access. The extraction uses those saved login details to connect and authenticate.
InSQL login and Barracuda VPN connection +
Bison's IT contact set up a view-only service account for InSQL access over the Barracuda VPN. The extraction uses that account to connect without requiring Duo 2FA.
Configured report list and parameters +
Each report has a stored parameter set (date range, output format, popup options). Agent uses these defaults unless overridden.
Connect to Pro Contractor via Azure Virtual Desktop (Flow A) +
The extraction opens an RDP session to the Azure Virtual Desktop endpoint, logs in using the saved connection details, and opens Pro Contractor.
Connect to InSQL via Barracuda VPN (Flow B) +
The extraction connects to the Barracuda VPN from a Sagan-hosted Windows machine using the view-only service account, then opens InSQL from the mounted file server.
Run each configured report with the right parameters +
Agent navigates to each report, applies the configured parameter set (date range, output format), and clicks Run.
Export each report as Excel +
Agent clicks through the 'save as' dialog, selects Excel format, and stages the file locally.
Upload to Google Drive +
Agent uploads all downloaded files to the per-client Google Drive folder, named by report type and extraction date.
Pro Contractor reports (Flow A) +
7-8 year-to-date reports (Income Statement, Balance Sheet, Project Cash Flow Analysis, Project Profitability, and others) exported as Excel files, named by report type and date.
InSQL reports (Flow B) +
6-8 reports per run (AR Aging, AP Aging, General Ledger, weekly snapshot set, monthly close set) exported as Excel files, named by report type and snapshot date.
Google Drive folder with timestamped files +
All extracted reports land in a per-client folder structure, e.g., Bison Drive / Clients / <Client name> / Extracts / <YYYY-MM-DD>/<Report Name>.xlsx.
Completion email notification +
Operator receives an email with a summary of what was downloaded, file count, and a link to the Drive folder.
Run history in the internal app +
Each extraction run is logged with timestamp, client name, files downloaded, any failures, and a link to the destination folder.
Is this for you?
- + Fractional CFO firms - Bison CFO and similar firms that deliver analysis to clients in meetings but spend hours extracting data from proprietary client systems with no APIs.
- + Teams managing multiple client extractions - Espie and Kurtis each handle different clients with different systems (Pro Contractor vs. InSQL). One UI, two pipelines, same outcome: files in Drive.
- + Operators who need to pull reports more frequently - Once extraction is automated, running additional reports becomes cheap. Kurtis: 'If we had the ability to automatically extract more reports, there's theoretically little things we could do for them that would help.'
- + Teams already using Google Drive for file storage - Bison lives on Google Drive. This build drops extracted reports into existing folder structures so downstream analysis scripts and HTML generation can pick them up unchanged.
- - Systems with public APIs - If your ERP or accounting software has a documented API, use that instead. This build is for non-API systems only.
- - One-off extractions - If you pull reports once a year, the setup cost doesn't justify the build. This is for recurring, repetitive extractions (monthly, weekly).
- - Teams without Windows system access - Flow A requires Azure Virtual Desktop access; Flow B requires Barracuda VPN access and a Windows machine. If your client systems are cloud-only SaaS, this pattern doesn't apply.
- - Extractions that need real-time scheduling - v1 is manual trigger only. Kurtis and Espie click Run when they need the reports. If you need fully automated scheduled runs, that's a future phase.
Pricing
This is a scoped build for Bison CFO. Pricing covers the initial build (prototype + full implementation of both Flow A and Flow B), plus setup and configuration of the two extraction pipelines.
- Scoped build: prototype (Flow A only) + full build (Flow A + Flow B).
- Usage-based runs: each extraction run (manual trigger) is included in the scoped build. No per-run fees.
- Future phases (scheduled runs, additional clients, additional reports) are separate engagements.
How much time does this save each month?
Espie spends 20–30 minutes per client per month manually exporting Pro Contractor reports through Azure Virtual Desktop. Kurtis spends multiple hours per week pulling InSQL reports over Barracuda VPN. This agent eliminates both workflows — one click replaces the entire extraction ritual, freeing both to move upstream into analysis work instead of clicking through Windows apps.
What reports does this extract?
Flow A (Pro Contractor) extracts 7–8 year-to-date reports including Income Statement, Balance Sheet, Project Cash Flow Analysis, and Project Profitability. Flow B (InSQL) extracts 6–8 reports per run including AR Aging, AP Aging, General Ledger, and weekly/monthly snapshot sets. The final report list for each flow is configured during setup; you can add or change reports later without rebuilding the agent.
Where do the extracted files go?
All extracted reports land in a per-client Google Drive folder, named by report type and extraction date. For example: `Bison Drive / Clients / <Client name> / Extracts / <YYYY-MM-DD>/<Report Name>.xlsx`. Files are ready for your downstream analysis scripts and HTML generation without any manual re-naming or re-organizing.
Do I need to set up anything on my end?
For Pro Contractor: you'll need your Azure Virtual Desktop login and the list of reports you want extracted. For InSQL: your IT team will need to create a separate, read-only account that doesn't require Duo 2FA — this is a one-time setup conversation with your IT contact. Once those are in place, the agent handles all the clicking and exporting automatically.
Can I run this on a schedule, or do I have to click a button every time?
Version 1 is manual trigger only — you pick the client and date range, then click Run whenever you need the reports. This gives you full control over when extractions happen and lets you override the date range if needed. Scheduled, fully automated runs are a future phase once the manual workflow is stable and you've used it for a few cycles.
What happens if a report fails to download?
If a run fails partway through, any files that were already downloaded successfully stay in your Google Drive folder, and the run is marked as partial in the run history so you know to redo it. The agent won't silently overwrite files — if a file with the same name already exists, it will flag the conflict so you can decide what to do.
Can I add more reports later without rebuilding the whole thing?
Yes. The agent is built to accept any configured report — you're not locked into the initial set. If you want to pull additional reports later (which Kurtis mentioned would help with client analysis), you can add them to the configuration without needing a rebuild. Automation makes it cheap to pull more reports whenever clients need them.
Is this only for Pro Contractor and InSQL, or can it work with other systems?
This build is purpose-built for Pro Contractor (via Azure Virtual Desktop) and InSQL (via Barracuda VPN) because both are non-API systems that require desktop automation. The pattern — connect to a Windows system, run reports, export as Excel, drop into Drive — can be extended to other proprietary desktop applications later if you add a third client with a different system. For now, it's focused on your two current extraction workflows.
Similar multi-step browser automation pattern: validates and enriches data from a source system, navigates multi-step web forms, handles payment/filing submission, and archives proof-of-filing back to the source. Both builds automate repetitive form-filling across non-API systems to eliminate manual re-entry and create audit trails.
Shares the multi-source fulfillment tracking pattern: monitors a task management system for new orders, retrieves data from heterogeneous systems (email parsing, portal scraping, API calls), fuzzy-matches records, writes data back to the system of record, and sends notifications. Both builds orchestrate data flow across systems with no native APIs.
Ready to stop clicking through Windows apps?
If your fractional CFO firm or accounting team spends hours every month extracting reports from non-API systems, this build frees that time for analysis. Let's talk about your extraction workflow.