0:06 — Welcome and presenter introduction — Sarah Steele opens the session on nonprofit strategy and AI agents for financial management. She introduces her background — about 15 years in client management and nearly a decade in custom software development — and her location on Pairsoft’s Indiana-based team. 1:23 — What PaperSave is — An overview of PaperSave as a natively integrated AP automation solution built on document management, including optical character recognition for scanning. The stated goals are speeding up AP work, increasing accuracy, protecting security, and keeping the ERP as the central source of truth, backed by more than 20 years of experience in the space. 3:06 — Housekeeping and how to get help — Steele explains that recordings will be sent out and that the account manager is the primary contact for information, demos, and one-on-one sessions. Anyone who does not know their account manager can reach info@pairsoft.com. A list of natively supported ERPs is shown. 4:56 — Agenda for the session — The plan for the webinar: why AI and why now, what Pairsoft has already built and a live demonstration of it, the AI roadmap, keeping the ERP as a single source of truth, developments in Pairsoft Pay, and Q&A at the end. Questions are collected through the chat and Q&A controls. 6:12 — Why AI, and the assist-not-assume principle — The case for AI is framed as helping finance teams make better decisions faster rather than replacing jobs, with AI agents already reshaping approvals and procurement workflows in spend management. The guiding principle stated is that AI should assist but not assume, with a framework governing when agents act, when they suggest, and when they stay out. 8:20 — Starting with GL coding — GL coding is described as the first area Pairsoft developed for AI because not every task is a good fit. The approach is to summarize past activity and suggest a coding rather than automatically switching it, since users find it uncomfortable when AI overwrites their history. Viewers are asked to keep PO matching and approver selection in mind as upcoming areas. 10:24 — Live demo: AI GL coding suggestions — Inside a PaperSave demo environment, an invoice in the AP review inbox has no distributions populated and shows an AI marker indicating the agent is active for that workflow. Clicking it returns three suggested distributions, and selecting the top one populates the correct distribution — staff development — along with fields from the connected ERP. 12:34 — How the suggestions are generated — The agent looks at what the organization has done with that vendor previously and returns the three most likely distribution candidates for the invoice. Because the history already exists for current PaperSave customers, enabling AI makes that data accessible; a UPS vendor the presenter had never coded returned postage and shipping as its top suggestion. New vendors require the system to learn, and users can still add, split, and edit distributions manually. 15:52 — Roadmap: PO matching and approver suggestions — Planned agents include PO matching, which would match invoice line item details to purchase orders to speed approvals, and invoice approver suggestions based on how a vendor has been routed before, dollar amount, and other record attributes. Both would suggest rather than force a routing, and are positioned as reducing errors and giving newly onboarded staff added confidence. 17:24 — Roadmap: reporting, fraud detection, cash forecasting — A reporting agent is under consideration to make PaperSave’s existing complex search and reporting easier to use, with report suggestions welcomed through account managers. Further out, the team is exploring AI-assisted fraud detection — fraudulent invoicing and fraudulent vendor information — and cash forecasting that uses historical invoices to help build future budgets. 19:37 — ERP data hygiene as a single source of truth — Using a gardening metaphor, Steele urges removing inaccurate information from the ERP and adding accurate information that the tool can pull in. Demonstrating vendor records in an NXT environment, she identifies correct vendor addresses as the most important item because OCR matches the printed invoice address to the vendor record, and notes Pairsoft has documentation of which ERP fields can be connected. 22:45 — Pairsoft Pay and closing — Pairsoft Pay is covered briefly: no change to the existing banking relationship is required, funds move from the customer’s primary bank accounts, batches are approved inside the PaperSave tool, and instructions are then sent to the bank for processing. Reported benefits include time savings, cost savings, and revenue share for some clients, and the closing directs all follow-up to account managers or info@pairsoft.com.

Home / Resources / Webinars

Demo: Using Financial AI for Your Nonprofit Tech Stack

Key takeaways

Pairsoft’s PaperSave AP automation tool now includes an AI agent that suggests GL coding distributions inside the invoice workflow.
The GL coding agent works by looking at how your organization has coded invoices for that vendor in the past and offering the three most likely distributions.
Suggestions are never applied automatically — the stated design principle is that AI should assist but not assume, and should not overwrite a user’s history.
The AI feature draws on data you already have in PaperSave, so existing customers get suggestions without having to build up new data first; brand-new vendors have to be learned over time.
On the roadmap: a PO matching agent, an invoice approver suggestion agent, a reporting agent, and further out, fraud detection and cash forecasting.
Suggestion quality depends on clean ERP data — the session urged ‘gardening’ your ERP, especially vendor addresses, which is how OCR matches invoices to the right vendor.
Pairsoft Pay was covered briefly: it does not require changing your existing banking relationship, and approved payment batches are sent to your own bank for processing.

Overview

This webinar, presented by Sarah Steele of Pairsoft, covers how AI agents are being added to the PaperSave accounts payable automation product, with a focus on what nonprofit finance teams can use today and what is planned. Steele introduced herself as having roughly 15 years of client management experience and close to a decade in custom software development, working primarily with the PaperSave tool from Pairsoft’s Indiana-based team. She framed PaperSave as a natively integrated AP automation solution built on document management, with optical character recognition that can handle scanning, and emphasized that the product sits inside the customer’s ERP system so the ERP remains the single source of truth. Pairsoft has over 20 years of experience in this space, and the tool integrates natively with a set of ERPs listed during the session.

The central argument for AI was framed as decision support rather than job replacement: the goal is to help finance teams make better decisions faster while keeping people in control. Steele repeatedly returned to the principle that AI should assist but should not assume — the product team built a framework for deciding when agents act, when they merely suggest, and when they should stay out of a process entirely. Some tasks were described as simply not appropriate for AI, and human nuance and judgment are explicitly not targets for replacement.

The first area Pairsoft built into is GL coding, and it is available now. In a live demonstration inside a PaperSave demo environment, Steele showed an invoice sitting in her AP review inbox with no distributions populated. An AI marker on the record indicated the agent was active for that workflow; clicking it surfaced three suggested distributions. Selecting the top suggestion populated the correct distribution — staff development in the example — along with the associated fields pulled from the connected ERP. The suggestions are derived from the organization’s history with that vendor: the system looks at what has been done with that vendor previously and returns the three most likely distribution candidates. Steele noted that because the underlying history already exists for current PaperSave customers, turning on the AI simply makes that data accessible, which lowers the barrier to adoption. She gave an example of a UPS vendor she had never personally coded, where the top suggestion returned was postage and shipping. New vendors, by contrast, require the system to learn the organization’s normal treatment over time. The AI button does not restrict normal behavior — users can still add distributions, split them, and edit fields as usual.

On the roadmap, Steele described a PO matching agent that would match line item details from invoices to purchase orders to speed up approvals, and an invoice approver suggestion agent that would recommend routing based on how a vendor has been routed before, the dollar amount, and other attributes on the record. As with GL coding, these would suggest rather than force a choice, and she pointed to error reduction and added confidence for newly onboarded staff as a benefit. A reporting agent was described as another idea, intended to make PaperSave’s existing complex search and reporting capabilities easier to use. Further out and less defined, the team is looking at AI-assisted fraud detection — fraudulent invoicing, fraudulent vendor information, vendors attempting to get paid for work not done — and cash forecasting that uses historical invoices to help build future budgets. Feedback and report suggestions were invited through account managers.

The session closed with two practical topics. First, ERP data hygiene: Steele used a gardening metaphor for pulling out inaccurate information and planting good information, demonstrating vendor records in an NXT environment. Accurate vendor addresses were called the single most important item, because OCR matches the address printed on the invoice to the vendor record in the system. Pairsoft maintains documentation listing which ERP fields can be connected and pulled over, available through an account manager. Second, Pairsoft Pay was covered briefly: customers do not need to change their banking relationship, funds move from the customer’s primary bank accounts, batches are approved inside the PaperSave tool, and the payment instructions are then sent to the bank for processing. Reported benefits included time savings, cost savings, and revenue share back to some clients. For all of these topics, the recommended contact path is the account manager, or info@pairsoft.com for anyone who does not know who their account manager is.

Chapters

0:06 — Welcome and presenter introduction — Sarah Steele opens the session on nonprofit strategy and AI agents for financial management. She introduces her background — about 15 years in client management and nearly a decade in custom software development — and her location on Pairsoft’s Indiana-based team.
1:23 — What PaperSave is — An overview of PaperSave as a natively integrated AP automation solution built on document management, including optical character recognition for scanning. The stated goals are speeding up AP work, increasing accuracy, protecting security, and keeping the ERP as the central source of truth, backed by more than 20 years of experience in the space.
3:06 — Housekeeping and how to get help — Steele explains that recordings will be sent out and that the account manager is the primary contact for information, demos, and one-on-one sessions. Anyone who does not know their account manager can reach info@pairsoft.com. A list of natively supported ERPs is shown.
4:56 — Agenda for the session — The plan for the webinar: why AI and why now, what Pairsoft has already built and a live demonstration of it, the AI roadmap, keeping the ERP as a single source of truth, developments in Pairsoft Pay, and Q&A at the end. Questions are collected through the chat and Q&A controls.
6:12 — Why AI, and the assist-not-assume principle — The case for AI is framed as helping finance teams make better decisions faster rather than replacing jobs, with AI agents already reshaping approvals and procurement workflows in spend management. The guiding principle stated is that AI should assist but not assume, with a framework governing when agents act, when they suggest, and when they stay out.
8:20 — Starting with GL coding — GL coding is described as the first area Pairsoft developed for AI because not every task is a good fit. The approach is to summarize past activity and suggest a coding rather than automatically switching it, since users find it uncomfortable when AI overwrites their history. Viewers are asked to keep PO matching and approver selection in mind as upcoming areas.
10:24 — Live demo: AI GL coding suggestions — Inside a PaperSave demo environment, an invoice in the AP review inbox has no distributions populated and shows an AI marker indicating the agent is active for that workflow. Clicking it returns three suggested distributions, and selecting the top one populates the correct distribution — staff development — along with fields from the connected ERP.
12:34 — How the suggestions are generated — The agent looks at what the organization has done with that vendor previously and returns the three most likely distribution candidates for the invoice. Because the history already exists for current PaperSave customers, enabling AI makes that data accessible; a UPS vendor the presenter had never coded returned postage and shipping as its top suggestion. New vendors require the system to learn, and users can still add, split, and edit distributions manually.
15:52 — Roadmap: PO matching and approver suggestions — Planned agents include PO matching, which would match invoice line item details to purchase orders to speed approvals, and invoice approver suggestions based on how a vendor has been routed before, dollar amount, and other record attributes. Both would suggest rather than force a routing, and are positioned as reducing errors and giving newly onboarded staff added confidence.
17:24 — Roadmap: reporting, fraud detection, cash forecasting — A reporting agent is under consideration to make PaperSave’s existing complex search and reporting easier to use, with report suggestions welcomed through account managers. Further out, the team is exploring AI-assisted fraud detection — fraudulent invoicing and fraudulent vendor information — and cash forecasting that uses historical invoices to help build future budgets.
19:37 — ERP data hygiene as a single source of truth — Using a gardening metaphor, Steele urges removing inaccurate information from the ERP and adding accurate information that the tool can pull in. Demonstrating vendor records in an NXT environment, she identifies correct vendor addresses as the most important item because OCR matches the printed invoice address to the vendor record, and notes Pairsoft has documentation of which ERP fields can be connected.
22:45 — Pairsoft Pay and closing — Pairsoft Pay is covered briefly: no change to the existing banking relationship is required, funds move from the customer’s primary bank accounts, batches are approved inside the PaperSave tool, and instructions are then sent to the bank for processing. Reported benefits include time savings, cost savings, and revenue share for some clients, and the closing directs all follow-up to account managers or info@pairsoft.com.

Questions this webinar answers

What does the AI GL coding agent in PaperSave actually do?

When an invoice is in the AP review workflow without distributions populated, an AI indicator appears on the record if the agent is enabled for that workflow. Clicking it returns three suggested GL distributions for that invoice. Selecting one populates the distribution along with the related fields pulled from the connected ERP. The user can still add additional distributions, split them, or edit fields manually.

Where do the GL coding suggestions come from?

The agent looks at the organization’s processing history in PaperSave for that specific vendor — what distributions have been used for that vendor’s invoices previously — and returns the three most likely candidates for the current invoice.

Do I need to build up new data before the AI suggestions are useful?

Not if you have been using PaperSave for a while. The historical data already exists in the system whether or not AI was enabled, so turning on the AI feature makes that existing data accessible for suggestions. Organizations new to the tool will not have that history yet, and any newly added vendor requires the system to learn how that vendor’s invoices are normally handled.

Will the AI change or overwrite coding on its own?

No. The stated design principle is that AI should assist but not assume. The agent summarizes past activity and offers suggestions, but does not automatically switch or overwrite existing coding, because users find it uncomfortable when AI overwrites their history. The product team built a framework specifying when agents act, when they only suggest, and where AI is not appropriate at all.

What AI capabilities are planned beyond GL coding?

Four areas were described. A PO matching agent would match line item details from invoices to purchase orders to speed up approvals. An invoice approver suggestion agent would recommend who to route an invoice to based on how that vendor has been routed before, the dollar amount, and other attributes of the record. A reporting agent would make the existing search and reporting capabilities easier to use. Further out, the team is exploring fraud detection — including fraudulent invoicing and fraudulent vendor information — and cash forecasting that uses historical invoices to help build future budgets.

What can I do to improve the accuracy of the automation?

Keep ERP data clean: remove inaccurate information and add accurate information the tool can pull in. Accurate vendor addresses were identified as the single most important item, because optical character recognition matches the address printed on an invoice to the vendor record in the ERP. Pairsoft maintains documentation listing which ERP fields can be connected, and an account manager can provide it and confirm whether all available fields are hooked up.

How does Pairsoft Pay handle banking?

There is no need to change your existing banking relationship. Funds move from your primary bank accounts to fund payments, batches are approved inside the PaperSave tool, and the payment instructions are then sent to your bank to begin processing. Reported benefits include time savings, cost savings, and revenue share returned to some clients.

Who should I contact to see a demonstration or ask follow-up questions?

Your Pairsoft account manager is the primary contact for scheduling demonstrations, arranging one-on-one sessions in your own system, and answering questions about AI features, efficiency, and Pairsoft Pay. If you do not know who your account manager is, emailing info@pairsoft.com will get you routed to the right person.

Table of Contents

Get A Demo:

Simple solutions. Powerful results. Seamlessly integrated.

Get a PairSoft Demo