Carragee Consulting

A business use case for Multi-Model RAG and OCR

Contract enforcement is a big weakness of mine as a landlord. I’ve had plenty of tenants pay late and I can’t think of one time I actually charged a late fee.

When I hired a property management business they charged it to a tenant in the first weeks after she moved in because she didn’t pay her first month rent on time confusing her security deposit with rent.

Over time my management company messed up various things like rent increase notifications for the leases on different cycles, but they also tried to drive towards simplicity by renewing with their “standard” lease terms.

Naturally, not all negotiations result in a deal with “standard” terms.

The complexity and administrative burden B2B companies face in compliance is huge.

Whether you work in a highly regulated industry, you have a high volume of contracts with unique T&C’s, or you release lots of versions with different warranties or SLAs it can get unwieldy in a hurry.

Organizations miss out on revenue their customers agreed to pay them because it’s challenging and time consuming to bill anything non-standard.

So while one solution is to hold back the creativity of the sales team and constrain them to pre-approved pick lists that fit nicely in a database, the reality is that you’ll have to make a judgement call for which type of error is optimal under imperfect information.

One of the ways companies may be allow for flexibility in negotiations and introduce micro segmented offerings WHILE still consistently collecting their fees is AI.

While you can’t eliminate the manual step of collecting and scanning documents, once they are digitized you can feed them into Unstructured which will put them in a vectorized structure that is LLM friendly.

Then your deal desk can use natural language to query for “create a list of all leases which require renewal notifications to be sent in the next 6 weeks. Including the lease start and end date in the result, current rent and emails address of the tenants.”

Years ago I tried to partner with company called Pramata to be able to propose this type of consulting service.

Now I see with coordinated use of
LangChain, unstructured.io, LlamaIndex and SingleStore you can VERY rapidly turn wild ideas into labor saving apps.

Start more confidently collecting the value you negotiated by structuring the unstructured data and then make it easy for your team to put Analytics in Action!