When you deploy an Excel spreadsheet as a web app with EASA, you get something most AI integrations struggle to find on their own: clean, structured, consistently labeled data. Every time a user fills out a form or runs a calculation in your EASA app, the output lands in organized fields like product names, pricing inputs, risk scores, and engineering parameters. These are exactly the kind of inputs that help a language model generate useful, accurate responses.
This guide walks through how to connect your EASA web apps to AI models via API, how to write prompts that make the most of your structured data, and how to keep AI outputs aligned with your business rules.
Why EASA Data Works Well with AI
Most AI integrations pull from messy, unstructured sources like emails, freeform notes, and inconsistent spreadsheets. EASA apps produce the opposite. Because your web app controls what users can input and how it gets stored, the data that flows into an AI prompt is already labeled, validated, and consistent.
This matters because language models generate better responses when they receive clear, well-organized inputs. A prompt that includes a labeled field like “Customer Industry: Manufacturing” gives the model more to work with than a raw data dump.
Start With the Task, Then Pass the Data
The most common prompting mistake is listing data fields first and leaving the instruction until the end. Lead with what you want the AI to do, then provide the data from your EASA app.
For example:
“Based on the product information below, write a two-sentence summary for the sales quote.”
Product Name: [ProductName] Customer Segment: [CustomerSegment] Configured Price: [FinalPrice] Lead Time: [LeadTime]
Without labels, the model has to guess what each value represents. Adding a descriptive label before each field removes that ambiguity and produces more accurate outputs.
Be specific about format and length. If you need two sentences, say so. If you need a minimum word count, include it. Language models follow clear format instructions when you provide them.
Connecting EASA to AI Models via API
EASA supports two-way API integrations, which means your web app can both send data to an AI model and receive the response back into the app. A typical setup works like this:
- A user completes a form or triggers an action in your EASA app
- EASA sends the relevant field data to the AI model via API
- The model generates a response based on your prompt and the data
- EASA writes the response back into a designated field in your app
This creates a seamless experience for the end user and stores every AI-generated output alongside the data that produced it, which is useful for auditing, version control, and quality review.
Use Reference Documents to Keep Responses On-Policy
If your AI outputs need to follow specific business rules like brand voice guidelines, underwriting policies, compliance requirements, or SLA commitments, you can pass reference documents into your prompt alongside the field data.
Attaching a document alone does not change the model’s behavior. You need to include explicit instructions telling the model how to use the document. For example:
“Use the attached tone guidelines when writing this response. Do not promise delivery timelines outside the agreed SLA. Do not include financial commitments not covered in the policy document.”
Teams using EASA for insurance pricing, engineer-to-order quoting, and financial modeling often have strict rules about what an AI-generated summary can and cannot say. Writing those rules into the instructions, rather than assuming the model will infer them, produces outputs that are consistent with your business requirements.
Practical Examples by Use Case
Insurance pricing: After an underwriter completes a rating model in EASA, trigger an API call that generates a plain-language summary of the quote for the agent or policyholder. Pass in the risk factors, coverage amounts, and premium as labeled fields. Include your underwriting guidelines as a reference document. EASA can also use AI to generate prefilled values and push them directly into rating forms before the underwriter begins, pulling from existing policy data or prior submissions to reduce manual entry and improve consistency across quotes.
CPQ and engineer-to-order: When a sales rep finalizes a configuration in your EASA CPQ app, generate a tailored quote narrative that explains the pricing and product selection in clear language for the customer. The structured output from your Excel pricing model gives the AI everything it needs.
Financial modeling: After a user runs a scenario in a financial modeling web app, generate an executive summary of the results. Pass in the key outputs as labeled fields and instruct the model to write for a non-technical audience.
Quick Reference: Prompt Best Practices
- Lead with the instruction before listing any data fields
- Label every field with a descriptive name and colon before the value
- Specify output format and length directly in the prompt
- Pair any reference documents with explicit instructions on how to use them
- Test the same prompt with different model settings and compare results
- Store AI responses back in your EASA app so outputs are tied to the data that generated them
Getting Started
EASA’s API integration layer gives your Excel-based web apps a direct connection to any AI model that supports REST API calls. You bring the structured data and business logic; the AI handles the language generation. To explore how to set up an API integration in your EASA deployment, contact the EASA team for a walkthrough.