When Your Software Decides: The Coverage Gaps Hiding in Automated Business Tools

Reviewed by Tom Moore, Agency Partner, CA Agency Insurance License 6003355
Last reviewed: 6/02/2026

Key takeaway: Automated decision-making business insurance in Spokane has a structural problem most owners don't see until they file a claim. When your software makes an operational call — denies a customer, rejects a job applicant, quotes the wrong price, ships the wrong order — most general business policies were written assuming a person made that decision. The gap shows up across E&O, general liability, cyber, and employment practices coverage. This applies to any Spokane business using scheduling tools, pricing engines, hiring software, chatbots, or fulfillment automation without a human reviewing every output.

What "automated decision-making" actually means for your business

Automated decision-making is any moment your software takes an action that affects a customer, employee, vendor, or applicant without a person reviewing it first. Not every automation counts. A spam filter is automation. A scheduling tool that confirms appointments is automation. Neither one is making a decision that creates business risk.

The line gets crossed when the software outputs something a person would normally have to sign off on. A dynamic pricing engine that quotes a customer $4,200 when the right number was $420. A screening tool that filters out applicants based on a flawed algorithm. A fulfillment system that ships the wrong product to the wrong address. A chatbot that promises a refund the business never authorized.

In each case, the software acted. The customer or applicant was affected. And if they push back legally, your insurance policy is going to ask a question it probably wasn't designed to answer: who made the decision?

The line between automation and decision-making

A practical test: if the output of the software directly affects a person's contract, transaction, employment, or money — and no employee reviewed it — that's a decision, not a task. That's where coverage questions start.

Why your business policy may not follow the software

Most general liability and professional liability policies were written when "the business" meant "the people who work there." The forms assume a person signed the contract, a person quoted the price, a person screened the resume, a person took the order.

When software does those things instead, claims adjusters and defense attorneys get to ask whether the policy actually applies. Some policies have caught up. Many have not. The ones that haven't were written before pricing engines, before AI screening, before automated dispatch — and they read that way.

I've seen Spokane businesses assume their existing general liability policy covers anything the business does. It doesn't. Coverage is defined by the policy language, not by what the business happens to be doing this year. If the policy says "wrongful act of an insured" and the wrongful act was technically performed by software your business deployed, that's a coverage argument waiting to happen.

The "human in the loop" assumption baked into most policies

Look at the definitions section of your professional liability or E&O policy. If it defines covered acts in terms of "professional services rendered by" your employees or insureds, ask your agent how that language applies when the service was rendered by an algorithm your business licensed from a vendor. The Washington State Office of the Insurance Commissioner publishes consumer guidance on understanding policy definitions, and that's the kind of question worth raising before renewal — not after a claim.

Five places software is quietly making operational calls

Most owners don't think of these systems as decision-makers. They should.

Pricing engines and dynamic quotes

Any tool that generates customer-facing prices without an employee approving each quote is making a decision. If the price is wrong, the customer's claim isn't against the software vendor in most cases — it's against the business that used it.

Scheduling and dispatch tools

Service businesses using automated dispatch can end up with crews at the wrong address, missed appointments that triggered customer compensation, or contractor scheduling that violated wage and hour rules. Each one is an operational decision the software made.

Inventory and fulfillment systems

E-commerce automation can ship the wrong product, charge the wrong card, or fulfill a fraudulent order that flagged on a human review but cleared on automated thresholds.

Hiring and HR screening software

This is the highest-risk category for employment practices claims. The Equal Employment Opportunity Commission has issued guidance on employer use of algorithmic decision-making tools, and Washington state has its own employment law standards that apply regardless of whether a human or a machine made the screening decision.

Customer service automation and chatbots

A chatbot that quotes a price, promises a refund, or makes a representation about your product is, legally, your business speaking. If the customer relied on what the bot said and lost money, that's a claim — and your professional liability policy may or may not respond to it.

The coverage lines most likely to have a gap

Four policy types are most likely to surface coverage questions when software makes the call.

Errors and omissions (E&O)

E&O is supposed to cover mistakes in professional services. When the mistake was made by a vendor's software your business deployed, expect questions about whether the act qualifies as a covered professional service. The Insurance Information Institute has consumer-facing background on how professional liability coverage applies in different scenarios.

General liability

General liability covers bodily injury and property damage. It rarely covers economic loss from a bad business decision — automated or not. If your software caused a customer financial harm without physical damage, GL is unlikely to respond.

Cyber liability

If the automated decision-making failure traces back to a data issue — bad training data, a corrupted database, an API failure — cyber liability may apply. But standard cyber policies focus on breach response, not on operational errors caused by software acting on the data it was given.

Employment practices liability

EPLI is the most exposed line when hiring software is in play. Discrimination claims arising from algorithmic screening are an active area of regulatory and litigation focus, and not every EPLI policy is current on the language.

What Spokane business owners should ask their agent

Three questions to raise before renewal:

1. Which of my current policies would respond if a software tool my business uses caused a customer or employee financial harm?
2. Are there exclusions for acts performed by third-party software, AI tools, or automated systems?
3. If a coverage gap exists, what endorsement or separate policy would close it?

These aren't trick questions. They're the conversations every Spokane business running on modern operational tools should have before the renewal lands on the desk.

If you run a business in Spokane and you've added software tools to your operations in the last few years — pricing engines, scheduling automation, AI hiring tools, customer service bots, fulfillment systems — your existing coverage may have been written for a business that doesn't quite exist anymore. We sit down with you, look at what you actually use, and find the gaps before a claim does. One call, no pitch, no pressure. Get a quote here at All Lines Insurance or reach the office at (509) 327-1658.

Frequently Asked Questions