Event businesses are built around people. Yet much of the work behind an event happens far from the guest experience: moving information between tools, checking receipts, chasing missing details, and repeating the same administrative steps. In this fireside conversation, Loophole’s Archer Lawrence, Noah Latsch, and Sean Sakaguchi discuss how automation can give that time back without asking a business to abandon the way it works.

Start With The Work, Not The AI Tool

The conversation draws a useful distinction between buying a product and improving a process. An AI application might take notes or draft an email, but its usefulness is limited if the result stays isolated from the rest of the business. The more practical question is what needs to happen between the systems a team already uses.

That might mean moving approved information from an inbox to a spreadsheet, updating a customer record, or triggering the next step after a person makes a decision. The starting point is the handoff: what information is needed, where it comes from, and what should happen next.

Use Rules Where Rules Are Enough

The team describes automation as a spectrum. Some tasks have clear rules and predictable results. Others require interpreting an unstructured document or weighing several pieces of information. AI can help with the second category, but it does not need to be responsible for the entire workflow.

A reliable system can combine ordinary rules, a narrowly defined AI task, and human review. The choice depends on the purpose of the task and the consequences of getting it wrong. A company may want to keep tight control over customer-facing communication while using AI to assist with repetitive internal administration.

Make Adoption Part Of The Solution

Noah emphasizes that every implementation has both a technical component and a human component. A system is not useful simply because it works in a demonstration. The people using it need to understand it, trust it, and fit it into their day.

For seasonal event businesses, timing makes that particularly important. Busy periods leave little room for retraining; quieter periods can bring different financial pressures. The team’s approach is to connect familiar tools and improve the process around them wherever that is practical, rather than treating a wholesale migration as the default.

The conversation returns to a small but telling reaction from an early client: seeing a repetitive task disappear meant having time to enjoy a coffee. The value was operational, but the first response was human. Less friction can make a working day feel different.

What The Receipt Example Reveals

The team walks through a catering expense workflow that began with receipt PDFs in Google Drive and an existing Google Sheet. The system extracted information, organized individual line items, and presented the results in a view the head chef could use to understand the budget.

That example combines three kinds of work: connecting platforms, interpreting receipt content, and designing a useful interface. It also shows why testing matters. Different vendors represent quantities, prices, tax, and discounts differently. The team describes an Amazon receipt format that initially caused undercounting when a unit price was mistaken for a line total.

Receipt review belongs in the workflow. Unusual formats and mismatched totals need to be easy to spot, so a person can check them without redoing every receipt by hand. The team used those exceptions to refine the extraction process and improve the results.

Build Something That Can Improve

The closing discussion separates the model from the system around it. Integrations, data structures, and a clear dashboard remain valuable even when the AI component changes. A better model can then improve one part of an established workflow instead of forcing the business to start over.

For a small or mid-size business, the practical next step is modest: choose one repeated handoff, describe how it works today, and identify the result a person actually needs. Build around that result, test the exceptions, and make the change easy for the team to adopt.