I see a massive potential for Swarms in the enterprise resource planning industry, and I’m a specialist in processes and technologies.
I want to open up for discussion on a few real-world use cases that represent great potential for this technology. These can make enterprises more efficient and be the “pulse” of an organization’s heartbeat.
From a broader perspective, manufacturers, asset-intensive, financials, human resources, logistics, and any module of a modern ERP can be touched by swarms and automated to an unthinkable level.
Below, I’ll list some processes that can revolutionize and save hundreds of thousands of dollars for companies, especially the ones that rely on older technologies such as invoice processing (yes, companies receive invoices via email and have people manually keying invoices). I’ll start with some basic opportunities, but that can go complex.
I’ll add my insights about it, which may be feasible, but that’s part of the challenge.
Let’s start with the basics:
- Invoice Processing: An agent that receives an invoice document (in any format) then goes for a 1-way, 2-way, 3-way, or 4-way matching. If the invoice can’t be matched against a voucher in the system, send the information to a workbench where the clerks will work on the exceptions. The average cost for an unmatched invoice is $53.00 per invoice!
- Customer Credit Limit: An agent receives Customer Credit Limit approval requests. The swarm would check internally in the system for payments, whether or not those payments were paid on time, check sales orders on hold, unpaid invoices, DSO on receivables, and after the internal due diligence, would go to the internet, on D&B, credit bureaus, banks, tax authorities, social media, etc. to make a reason whether or not the credit is OK for increase.
- Purchase Order Approvals: An agent receives Purchase Order information and, based on certain criteria and internal due diligence, such as vendor compliance (ISO certificates in the expiration dates), checks for delivery reliability, checks with quality, and checks with credit reports and external entities to check whether the supplier can deliver or not.
- Assets work order maintenance: Based on sensor information (IoT via Node-Red) like temperature history or vibration, an agent can constantly monitor an asset’s status and perform predictive maintenance to avoid operation disruptions.
- Homebuilder: An agent monitors lots’ status and issues contracts based on previous agreements and business rules. Of course, the tenant information is checked with external sources for approval or denial of the transaction.
These are a few examples of how companies in many industries can benefit from Swarm technology. I hope this can grow, and I put my company open to build such agents and add resources in this research. Later on, we can build dashboards and logs to lookup all the swarms activities, and my goal on building this, is to log each step in the document record in the ERP system, so we can have a transparent view of the decisions made by the Agent.
There are so many other valuable cases in Financials. Still, several companies are SOC1 compliant, which may make implementing such Agents/Swarms more challenging due to liabilities—yes, people can go to jail if they don’t use such systems responsibly. It would be difficult to appoint a fraud or resolve a wrong deposit of $1M because the AI Agent added an extra zero in the transaction.
I’m super excited to tackle these small use cases with Swarms instead of Custom GPTs, as I’m doing now, even though the GPTs provide an AMAZING user experience!