What Does It Cost When AI Recommends a Competitor Instead?

The cost compounds quickly. Every enquiry that chooses a competitor recommended by an AI assistant represents the acquisition cost you already paid, the customer lifetime value you will not collect, and the referral or repeat business that now goes elsewhere. Multiply one lost customer by those three numbers and the real cost becomes visible.

How many enquiries does AI discovery represent right now?

Nearly half of all consumers used an AI tool for local search in the past month (yext.com), and 71% of those AI users say they are using AI more than they were a year ago. The behavior is no longer experimental. 43% of adults who use AI tools have asked an AI assistant to recommend a local business in the past 90 days (scope.online).

That share varies by income. At $150k+ household income, AI has already overtaken Google as the starting point for local search, with 54.5% starting their journey with AI. These are not tire-kickers. These are the customers with the budget to buy and the willingness to pay for what they need.

The enquiry volume you see today includes people who found you through AI and those who did not. The worksheet below helps you estimate how many enquiries arrived because an AI assistant named you, and how many went to competitors because the assistant named them instead.

What does one lost customer actually cost?

Start with customer lifetime value. The basic customer lifetime value formula is: CLV = Average Purchase Value × Purchase Frequency × Customer Lifespan (twilio.com). If a client spends $500 per engagement, returns twice a year, and stays for three years, their lifetime value is $3,000.

The CLV to CAC ratio should be at least 3:1, meaning each customer generates 3x more value than you spent to acquire them (zendesk.com). If your acquisition cost per customer is $200, each lost enquiry that would have converted costs you $200 in wasted spend plus the $3,000 in revenue that customer would have generated.

Then add the referrals. A satisfied client refers others. A lost client refers no one. The compounding cost of one decision made by an AI assistant in someone else's favor can exceed five figures when you account for the full relationship and the customers that relationship would have generated.

Worksheet: estimating enquiries lost to AI-driven discovery

Use your own numbers. Do not borrow industry averages. The exercise works because it is grounded in what your business actually sees.

Step 1: Estimate AI-sourced enquiries

Take your total monthly enquiries. Multiply by the share of your target market that uses AI for business discovery. If you serve households earning over $150k, use 54%. If you serve a broader consumer base, use 43%. The result is a rough count of enquiries that likely involved AI at some point in their decision.

Not all of those enquiries came to you because the AI named you. Some found you despite what the AI said, or they verified the AI's recommendation with a traditional search. But the number gives you a floor.

Step 2: Identify your website conversion rate

A good conversion rate for a website is between 2% and 5% for most industries (popupsmart.com), with the global average at 2.35% across all industries (greetnow.com). If 100 people visit your site, two to five of them take the next step, whether that is filling out a form, calling, or booking.

Check your analytics for your actual rate. If you do not track it, 2.35% is a reasonable starting assumption.

Step 3: Calculate cost per enquiry

Add up what you spent last month on all lead generation: paid search, local services ads, directories, agencies, your own time managing campaigns. Divide by the number of enquiries you received. On a per-lead basis, you will typically see costs ranging from $50 to $150 for local businesses (newmedia.com), with the average cost per qualified lead across all industries at $198.

Your number might sit outside that range. Use it anyway. The worksheet is about your business, not the average.

Step 4: Estimate customer lifetime value

If you do not have this number already, calculate it now. Average transaction value, multiplied by purchase frequency per year, multiplied by how many years a typical customer stays. A $400 transaction, three times a year, for two years, equals $2,400.

Write that number down. It is the revenue at stake every time an AI assistant names a competitor instead of you.

Step 5: Calculate the cost of one lost customer

Add your cost per enquiry to your customer lifetime value. If it cost $100 to generate the enquiry and the customer would have been worth $2,400, the cost of losing that customer is $2,500.

Step 6: Multiply by the number of enquiries lost

If 20% of your AI-sourced enquiries went to competitors because the assistant recommended them and not you, multiply that lost-enquiry count by the cost per lost customer. That is the monthly cost of not being recommended.

The number will be uncomfortable. It should be. The cost is real, it is recurring, and it grows every month as more people adopt AI for discovery.

What to do with the number once you have it

The worksheet gives you a business case. If the monthly cost of lost enquiries is $5,000 and you spend $200 a month checking whether AI assistants recommend you, the decision is clear. If the cost is $500, you have a judgment call.

The free AI-visibility check (/check) shows whether Claude, ChatGPT, Gemini and Perplexity name your business when someone asks who to hire, and which competitors they name instead. Each assistant is reported separately. The check takes no signup and returns results in seconds.

If the assistants do not name you, the guide (/guide) explains what they look for and how to give them the signals they need. If they name competitors, the same competitors appear in every scored audit on ActiScore's pricing (/pricing) page, from the free plan through to the Growth tier, so you can see their scores on the same rubric and identify what they are doing that you are not.

The cost of inaction is measurable now. The worksheet proves it. The question is whether you will measure it before or after the revenue gap becomes too large to ignore.