Auditoria Universal Connector

1. Overview

1.1 What is the Universal Connector?

In simple terms, the Universal Connector is how you get your accounts payable and receivable information into Auditoria when there is no automatic, direct link to your accounting or ERP system. Instead of a live connection, you provide your data as files — much like exporting a report to a file and placing it in a shared folder — and Auditoria reads those files, checks them, and returns the processed results for your team to load into your system.

More precisely, the Universal Connector lets your organization exchange Accounts Payable (AP) and Accounts Receivable (AR) data with Auditoria when your enterprise resource planning (ERP) system has no prebuilt connector. You move data through standardized comma-separated values (CSV) or JavaScript Object Notation (JSON) templates over a secure file channel — either Secure File Transfer Protocol (SFTP) or an Amazon Web Services (AWS) Simple Storage Service (S3) bucket. The template-driven approach keeps data consistent, reduces mapping errors, and supports back-office AP and AR automation without a custom-coded connector.

1.2 When should you use it?

1.3 What it supports

The Universal Connector supports AR Collections, AR Helpdesk, and AP Helpdesk. It does not support AP Invoices, AR Remittances, or AP Accruals. The object-by-object matrix, file formats, and size limits are covered where relevant below.

The email-based invoice workflow in Section 4 is delivered through AP Helpdesk email ingestion. It is separate from the AP Invoices product, which the connector does not support.

1.4 What it is not

• Not real-time. Processing runs on a schedule (every 4 hours by default) or on demand. This is batch integration, not live sync.
• Not structurally customizable. Field names, order, and file naming follow fixed Auditoria templates. This keeps data clean, but an existing report cannot be used as-is without a mapping exercise.
• Not hosted by Auditoria by default. You provide and maintain the SFTP server or S3 bucket (case-by-case exceptions only).
• Not shared. Every client gets a dedicated SFTP endpoint or S3 bucket — never shared across tenants or business units.

1.5 Who does what

You own the file infrastructure (SFTP or S3), the ERP export that produces the files, and the ERP import that consumes the results. Auditoria polls the drop-off folder, validates each file against its template, processes valid data, archives it, and writes results and error logs back for you to pick up. Auditoria never writes to your ERP directly.

2. Prerequisites

Have the following in place before you implement. These are what you need; the interface steps for entering the details are in Section 3.4.

File transfer environment — set up one of the following:

• SFTP:
  • A secure SFTP server with reliable connectivity.
  • Read, Write, and Processed folders that Auditoria can access.
  • Credentials (host, port, username, and password) shared securely with Auditoria.
• AWS S3:
  • A dedicated bucket with Read, Write, and Processed paths, and an optional Attachments folder.
  • Access keys and permissions meeting Auditoria's security requirements, provisioned securely within your AWS environment.

Auditoria platform access:

• Administrator or implementation-level access to the platform.
• An active Auditoria tenant configured for your organization.
• The required Auditoria modules enabled (for example, SmartVendor).

ERP readiness — your ERP or business application must be able to:

• Export the required data (invoices, vendors, payments, and so on) in CSV or JSON.
• Automate and schedule exports into the Read folder.

Supported file formats:

• Accepted: uncompressed, plain-text CSV and JSON files only.
• Rejected: ZIP, GZIP, Excel, encrypted, and image files.
• Size limits: JSON under 150 MB; CSV under 75 MB.
• Attachments (PDFs, images) are archived for reference, not parsed as structured data.

Share all credentials and access keys through secure channels only — never plaintext email. See Section 6 for the full policy.

3. How It Works

3.1 Templates and file naming

Auditoria provides a fixed template for each business object (Invoices, Vendors, Payments, and so on), specifying field names, order, data types, and allowable extended fields. Download the templates and complete them with your source field mappings before you begin:

• AP data mapping template — vendor, vendor bill, purchase order, and payment objects, with line items: AP Universal Connector Data Mapping Template
• AR data mapping template — customer, invoice, payment, and credit memo objects: Generic ERP AR Data Mapping Template

Do:

• Use the exact template for the intended object.
• Fill in all required fields.
• Keep field names and order unchanged, except approved extended fields.
• Validate your export against the template before uploading.

Don't:

• Reorder, rename, or add undocumented fields.
• Remove required columns, even unused ones.
• Save as Excel, PDF, or compressed formats.

Why this matters: Auditoria matches each file against its template automatically. A renamed column, a missing required field, or an incorrect file name fails the entire file — there is no partial load. Treat the template and naming convention as exact specifications, not guidelines.

File naming pattern: <PREFIX>_<yyyyMMddhhmmssAM/PM>.<ext>, where the prefix is one of ENTITY, CUSTOMER, INVOICES, PAYMENTS, or VENDOR. For example, an input invoice file is INVOICES_20251118093000AM.csv, and its error log is INVOICES_error_20251118093000AM.log.

3.2 Folder structure

You store AP and AR data in a single root directory. Within it, the connector uses three folders:

• Read: Your drop-off point — the folder Auditoria monitors for new input files.
• Write: Auditoria's results for you to pick up — processed output, acknowledgments, error logs, and, on SFTP, archived attachments.
• Processed: Archive of files handled successfully; this doubles as your audit and compliance archive.

On AWS S3, you can also configure a dedicated Attachments folder; on SFTP, attachments are written to the Write folder. (One capability — AP email ingestion — routes attachments to the Write folder even on S3; this exception is flagged where it applies, in Section 4.)

You choose the folder paths during setup. Once Auditoria is polling a given path, treat it as fixed: renaming or restructuring it requires re-running setup.

Processing schedule. A component called the AI Agent checks the Read folder on a schedule (every 4 hours by default) or when a sync is triggered on demand, and it processes whatever it finds. As a rule of thumb, once a polling cycle has run, a file still sitting in Read has failed validation — look in Write for its error log, correct the file, and drop it back in Read. (A file can also still be in Read because the next poll has not run yet, so allow for the schedule before assuming a failure.)

Audit trail. For audit and compliance, two distinct records exist:

• The Processed folder — the file-level archive, and the canonical audit trail when you need the actual file.
• The application audit log (in the Auditoria console) — the event history (successes, failures, timestamps) when you need a searchable record of what happened, independent of the files.

3.3 Configuring the connection in Auditoria

Prerequisites: infrastructure and folders prepared (Section 2); administrator access to Auditoria; your connection details to hand.

You set one connection type per instance — either SFTP or an S3 bucket. Provide credentials securely; never send them in plaintext email.

Set up an SFTP connection

1. Confirm the SFTP folder paths you created during infrastructure setup. The paths below are illustrative — use your own, then treat them as fixed once Auditoria is polling them:
   • Root: /Auditoria
   • Read (ingestion): /Auditoria/read
   • Write (results, acknowledgments, error logs): /Auditoria/write
   • Processed (audit archive): /Auditoria/processed
2. Sign in to Auditoria as an administrator and go to Universal Connector > ERP Settings.
3. Set Connection Type to SFTP and complete all fields (see below).
4. Select Save.

Result: The connection is saved and appears under ERP Settings.

Figure 1. ERP Settings page with Connection Type set to SFTP.

Set up an S3 bucket connection

1. Confirm the S3 folder paths you created during infrastructure setup (illustrative; use your own, then treat as fixed once polled):
   • Bucket root: /Auditoria
   • Read: /Auditoria/read
   • Write: /Auditoria/write
   • Processed: /Auditoria/processed
   • Attachments: /Auditoria/attachments
2. Gather the required S3 details: AWS region, bucket name, Access Key ID and Secret Access Key, the four folder paths, and the at-rest encryption type.
3. In Auditoria, go to Universal Connector > ERP Settings and set Connection Type to S3 Bucket.
4. Complete all fields (see below). Enter a hyphen (-) only where a field genuinely does not apply.
5. Select Save.

Result: The connection is saved and appears under ERP Settings.

Figure 2. ERP Settings page with Connection Type set to S3 Bucket.

4. Running the connector

4.1 The end-to-end data flow

This is the canonical description of how data moves from your ERP through the connector and back. Other sections refer to it rather than repeating it.

1. Export. Your ERP exports data using Auditoria's CSV or JSON template, named per convention (for example, INVOICES_20251118093000AM.csv).
2. Upload. Your team places the file in the Read folder. Upload to a temporary location and atomically move it into Read to avoid partial pickups.
3. Validate. The AI Agent polls Read on schedule (or on demand), validates the file name and format, and checks the contents against the template schema at both header and line levels.
4. Error path. Invalid files stay in Read; a detailed error log is written to Write (for example, INVOICES_error_20251118093000AM.log). Fix and re-upload.
5. Success path. Valid files move to Processed; the processed output is written to Write, ready for download and ERP import.
6. Import. Your ERP or middleware reads the validated output from Write and loads it in, ideally through an automated job on a matching schedule.

Re-sending a record: If a record ID is reused or a file is re-sent, Auditoria updates the existing record (upsert) rather than creating a duplicate.

Figure 3. Universal Connector end-to-end data workflow.

4.2 Sync modes

Separately from the polling schedule (Section 3.2), each connection runs in one of two sync modes, set during implementation:

• Incremental: Only new or changed records are processed.
• Full: All records are processed; Auditoria calculates what has changed, including deletions.

4.3 Monitoring day to day

Once live, follow this operating rhythm:

• Check Write regularly for processing results, error logs, and any acknowledgment or summary files.
• Check Read after each polling cycle — a file left there after a cycle has run typically failed validation. Review its error log, fix, and re-upload.
• Periodically verify Processed is intact and complete.

Retain logs and archived files for the period your legal, regulatory, or audit policy requires. Error, acknowledgment, and summary logs in Write are overwritten each cycle, so to keep a permanent record, download and archive them to your own systems.

5. AP/AR Examples

These examples show how data moves through the connector end to end. Use them alongside the field-level definitions in the data mapping templates (Section 3.1).

5.1 AR data exchange

AR follows the canonical flow in Section 4.1 — no email ingestion or AI extraction:

• Objects: Customer, invoices, payments, credit memos.
• File names: the required prefix per object (for example, CUSTOMER_, INVOICES_, PAYMENTS_, CREDITMEMO_) plus the standard timestamp.

5.2 AP data exchange (file workflow)

AP master and transactional data also follows the canonical flow in Section 4.1 (separate from the email-ingestion capability in Section 5.3):

• Objects: Entity, Vendor, Bills, Payments, Purchase Orders.
• File names: the required object prefix plus the standard timestamp (for example, ENTITY_, VENDOR_, PAYMENTS_).

5.3 AP Invoice Workflow

This is an AP Helpdesk email-based capability, distinct from the AP Invoices product. It is structurally different from the file workflow — treat it as an addition, not a variation.

1. Reception. Invoices arrive at a dedicated, monitored inbox. Auditoria identifies emails with invoice attachments by sender, subject, and attachment type.
2. Extraction. The AI Agent uses Optical Character Recognition (OCR) and Natural Language Processing (NLP) to extract vendor details, invoice number, date, and due date, line items and amounts, purchase-order references, and payment terms. It also runs duplicate checks, applies coding (for example, cost center and GL account), flags anomalies, and validates against business rules and schema.
3. Write output. Auditoria writes each invoice to Write as a subfolder named with the invoice number, containing the data file (JSON or CSV) and the original attachment (timestamped). This routes to Write regardless of SFTP or S3 — it does not follow the general S3 Attachments-folder rule in Section 3.2.
4. Logs. Auditoria writes a success log and a failure log to Write (distinct from the Processed folder — see step 6). Each invoice's result also appears in the console.
5. ERP import. Your ERP or middleware polls Write and imports the data and attachments.
6. Reconciliation. Unlike the standard flow, these files are not moved to Processed. Auditoria reconciles the logs for the invoices delivered through SFTP or S3 and displays the results in the console for compliance and recordkeeping.

Figure 4. Universal Connector AP invoice email-ingestion workflow.

6. Operations and security

For how the connector reads, validates, and archives files, see Sections 3.2 and 4.1. This section covers the operational and security responsibilities around that flow. One principle underpins the rest: each organization or business unit uses a dedicated SFTP endpoint or S3 bucket, and infrastructure is never shared across tenants.

6.1 Data handling

• Auditoria transfers all data between your ERP environment and the platform over secure, encrypted channels (SFTP, or S3 with TLS).
• The connector accepts only the plain-text CSV and JSON files described in Section 2, conforming to Auditoria's templates and naming conventions.
• Attachments are archived rather than parsed as structured data: in Write on SFTP, or a dedicated Attachments folder on S3.

6.2 Access controls

• Your IT or cloud administrator manages permissions for each SFTP or S3 directory.
• Grant only the read, write, and move or rename permissions that users and Auditoria service accounts require.
• Review folder access regularly to limit unauthorized exposure.
• Do not distribute credentials or access keys by email or other insecure channels.

6.3 Isolation

• Cross-client isolation comes from dedicated infrastructure: each client provides and maintains its own SFTP server or S3 bucket, never shared with another client.
• Within your own environment, the separate Read, Write, and Processed folders keep the ingestion, output, and archive stages distinct.
• SFTP servers, S3 buckets, and folders cannot be shared between different organizations or business units.

6.4 User management

The Universal Connector does not provide application-level user management or assignable roles for file processing. SFTP or S3 credentials and directory permissions govern all file access, and your IT or cloud administrators control them.

• SFTP: access is governed by SFTP user accounts and password management; only IT can add, disable, or change access.
• AWS S3: access is governed by AWS Identity and Access Management (IAM) users, roles, and policies; consult your IT or security team when personnel or business needs change.

6.5 Backup and disaster recovery

Auditoria does not back up or restore the customer SFTP or S3 directories (Read, Write, Processed, and Attachments). Backup, retention, and disaster recovery remain your responsibility:

• Schedule regular backups of all directories in the deployment.
• Retain Write and Processed contents for audit and compliance. Error, acknowledgment, and summary logs in Write are overwritten each cycle (Section 4.3), so archive anything you need to keep.
• Restore lost or deleted files through your internal data management team, using your own backup solutions.

7. Troubleshooting

8. FAQ

What is the Universal Connector? A file-based integration, over SFTP or AWS S3, that ingests standardized CSV or JSON into Auditoria on a schedule or on demand. It does not push data into your ERP.

When should I use it? When your ERP can export CSV or JSON and you want a prescriptive, template-driven batch integration — typically when no native connector exists.

What ERPs are supported? Any ERP that can produce or consume flat files in Auditoria's required format (for example, SAP Ariba or Infor).

Which products, objects, and record types are supported? See Section 3.2, "What it supports." Line items are supported for Bills/Invoices, Payments, and Purchase Orders.

Who hosts and maintains the file infrastructure by default? The client provides the SFTP server or S3 bucket, folders, and credentials.

Can Auditoria host SFTP instead? Not standard — possible only with case-by-case approval.

Can multiple clients share one file endpoint (multi-tenant)? No. Each client needs a dedicated SFTP endpoint or S3 bucket.

What folders are required, and what permissions do they need? Read, Write, and Processed (Section 1). Auditoria needs read on Read, write on Write, and the ability to move/rename files into Processed.

What file formats and size limits apply? CSV or JSON only; no compressed formats. JSON under 150 MB, CSV under 75 MB (Section 2).

Can we change the folder structure, fields, or file types? No. The connector is intentionally prescriptive — this is what keeps the integration low-maintenance and error-resistant.

How must files be named? <PREFIX>_<yyyyMMddhhmmssAM/PM>.<ext>, prefix one of ENTITY, CUSTOMER, INVOICES, PAYMENTS, VENDOR (Section 3.1).

Can I send multiple files for the same record type in one cycle? Yes.

Are custom or extended fields allowed? Use Auditoria's templates exactly, including field names, order, and data types. Extended fields are supported as provided in the templates (for example, cost center, GL account, division, entity, location).

Where do I get the templates? See Section 9.

How often does ingestion run, and can I schedule it? Every 4 hours by default; can be triggered on demand. Sync windows and force runs are supported operationally — coordinate during implementation.

What sync modes are supported? Incremental (only new/changed records) and Full (all records, with deltas and deletions calculated by Auditoria) — see Section 4.2.

What happens on success and on failure? On success, files move from Read to Processed, and acknowledgments/status may be written to Write. On failure, files remain in Read and detailed error logs are written to Write for correction and re-upload.

What happens if an ID is reused or a file is re-sent? The existing record is updated.

Do you push results back into the ERP? No. Auditoria writes only acknowledgments, results, and errors to Write for your pickup.

Where can we find audit logs? Two places, for two purposes: the Processed folder (the actual archived files) and the application audit log in the console (the event history) — see Section 3.2, "Processing schedule and audit trail."

Are there per-tenant limits on connections, record types, or files per day? No enforced upper limit. Practical throughput depends on file sizes, volumes, and scheduling.

What is the best practice to avoid partial pickups? Upload to a temporary location, then atomically move/rename the file into Read once it's complete.

What if I hit a problem? See Section 7, Troubleshooting.

9. Data Mapping Templates

The data cycle. Your ERP produces the data, Auditoria reads and checks it and hands back the results, and your team loads those results into the ERP. Auditoria acts on files only: it reads from and writes to the file location you set up, and it never writes to your ERP directly.

Field types.

Fields also exist at two levels: header-level (the transaction as a whole — invoice number, total, due date) and line-level (individual items — line description, line amount). Standard, optional, and extended fields can all appear at either level.

Auditoria provides a standardized data mapping template for each product area. Each template defines the fields, field types, and file structure your exported CSV or JSON files must follow. Download the current templates and complete them with your source field names before you begin:

• AP data mapping template — vendor, vendor bill, purchase order, and payment objects, with line items: AP Universal Connector Data Mapping Template
• AR data mapping template — customer, invoice, payment, and credit memo objects: Generic ERP AR Data Mapping Template

10. Terminology

11. Support

For more information or assistance, contact Auditoria at support@auditoria.ai. You can:

• Submit and track support requests through the Auditoria Support Portal.
• Email technical questions, troubleshooting steps, or general inquiries to the same address.