High Intent MCA Leads From ChatGPT Ads: A Q&A on Budget, Targeting, and Tracking

By the Lendnet.io team

Merchant cash advance lead generation has run on the same few platforms for years. Meta and Google deliver volume, but brokers and small funders know what comes with it: climbing costs per lead, shared leads, and long hours spent chasing merchants who have already talked to a dozen other shops. ChatGPT ads open a new place to reach merchants, and they work differently from anything most funders have bought before.

I run Lendnet.io, and I am now rolling out ChatGPT ads for brokers and small funders who need access to merchants who are looking for funding. While researching the channel, I worked through the questions below: how the intent logic works, what a realistic budget looks like, how targeting compares with Meta, and how to track results with a Conversions API. I have laid them out as a Q&A, with each question as a heading.

Who is Lendnet.io, and why add ChatGPT ads now?

We have been doing MCA lead gen since 2018, and in that time we have gotten clients profitable on platforms like Meta and Google. ChatGPT ads are our newest offering. We built it for brokers and small funders who need access to merchants who are looking for funding right now.

The reason comes down to where the ad appears. Ads on ChatGPT show up while the user is prompting the AI. When a merchant asks the AI about funding options, your company name and your offer can appear right alongside that conversation. You get in front of the lead while it is burning hot, at the moment the merchant is working out who to talk to.

I am only onboarding a few clients for this at the moment, because I want to give each one real attention while the channel is young.

What does intent logic mean in the context of ChatGPT ads?

It helps to compare how each platform finds a prospect.

On Meta, the merchant is scrolling through a feed. Your ad interrupts that, and the platform uses interest and behavior data to guess that the person might care about funding. Meta’s algorithm is very good at making that guess, but the merchant did not ask for anything.

On Google, the merchant types a short search phrase. That is a real signal, but it is a handful of words, and you have to infer the situation behind them.

On ChatGPT, the merchant describes the situation in full sentences. They might explain what kind of business they run, what they need the money for, how fast they need it, and what went wrong with other options. The ad is matched to that conversation. According to OpenAI’s help documentation, ChatGPT ads use conversational context, along with context hints supplied by the advertiser, to match ads with relevant conversations.

So the signal that triggers your ad is a request the merchant made themselves, in their own words, moments ago. That is the intent logic behind this offer: you appear during the research and comparison stage, when the merchant is asking questions and has not settled on a funder yet.

How does the system decide which conversations get an ad?

Through contextual matching. Advertisers provide context hints, which describe the situations where their offer is relevant. The system then looks for eligible conversations that fit those descriptions.

One detail matters a lot here. Context hints are not exact-match keywords, and they do not guarantee delivery. You cannot type a phrase like merchant cash advance and expect to appear in every conversation that contains it. You describe the situation, and the system decides whether a conversation fits.

OpenAI recommends grouping ad groups around distinct customer needs and using specific, accurate hints. For MCA, assuming the product is eligible, we would build themes like these:

• Business funding needs: conversations about comparing financing options to cover business expenses or manage cash flow.
• Financing comparisons: business owners researching funding options, eligibility criteria, repayment structures, and costs.
• Funding eligibility: businesses exploring which financing products may fit their circumstances.

Treat these as example themes. They are not targeting filters, and they do not guarantee placements.

Will the leads actually be better than the ones from Meta?

I expect them to be, which is why I am working with a small group of clients to test it. I will not tell you the quality is equivalent or superior before the data says so. Meta’s financial ad category audience and the ChatGPT audience are different groups of people in different mindsets, and nobody should assume they produce the same lead quality.

What I can lay out is why the case is strong. A merchant who is asking an AI about funding options has raised their hand on the topic before your ad ever loads. On Meta, a lot of the work goes into creating interest in a merchant who may not have been shopping. On ChatGPT, the shopping has already started.

Higher intent should show up in the numbers that matter: submission rates, completed loan apps, and funded deals. We will measure those in your CRM rather than trusting platform reports alone. Neither platform guarantees high-intent prospects, and we will say so plainly to every client.

How much would it cost to get 10 qualified leads per day?

The right way to estimate it is to work backward from the target, using cost per click, landing page conversion rate, and the real cost of acquiring a qualified lead.

OpenAI’s current advertiser guidance recommends starting with a maximum CPC bid of $3 to $5 per click. That is a starting bid recommendation. It is not a guarantee of your actual average CPC, and it is not an MCA-specific benchmark. For planning, I use the conversion rates from our Meta ads funnel: about 30% of clicks become submitted lead forms, and about 10% of clicks become qualified leads, because many merchants are denied at the lead stage. Here is the math:

• 10 qualified leads per day at a 10% click-to-qualified-lead rate requires 100 clicks per day.
• 100 clicks at $3 to $5 each comes to $300 to $500 in daily ad spend.
• Over 30 days, that is roughly $9,000 to $15,000.
• That works out to $30 to $50 per qualified lead, and about $10 to $17 per submitted lead form at a 30% submission rate.

These are modeled scenarios. The conversion rates come from our Meta ads funnel, so they are an assumption for ChatGPT traffic, and they are not observed ChatGPT campaign results.

Where does the 10% qualified rate come from?

Our Meta ads funnel is the source of these rates, and it runs in stages: lead, pre-qual, loan app, bank statement submittal, then stips. A lead is not a loan application. About 30% of clicks submit the lead form, but many of those merchants are denied at the lead stage and drop out, which leaves about 10% of clicks as qualified leads.

That 10% is measured against all ad clicks, so the budget above holds if ChatGPT traffic behaves like our Meta traffic. I expect ChatGPT traffic to qualify at a higher rate, because the merchant is already asking the AI about funding before the ad appears. Here is what the budget looks like for the same 10 qualified leads per day if the qualified rate improves:

• Same as Meta (10% of clicks qualified): 100 clicks per day, $300 to $500 daily, $9,000 to $15,000 monthly, or $30 to $50 per qualified lead.
• 12.5% of clicks qualified: 80 clicks per day, $240 to $400 daily, $7,200 to $12,000 monthly, or $24 to $40 per qualified lead.
• 15% of clicks qualified: 67 clicks per day, $201 to $335 daily, $6,000 to $10,100 monthly, or $20 to $33 per qualified lead.
• 20% of clicks qualified: 50 clicks per day, $150 to $250 daily, $4,500 to $7,500 monthly, or $15 to $25 per qualified lead.

Every gain in the qualified rate lowers the budget needed for the same 10 leads per day, which is why I track qualified outcomes in the CRM from day one.

Until actual CPC, qualified lead rate, and delivery volume are known, I treat every one of these figures as a forecast scenario and not a validated spend recommendation.

A qualified lead is only the first stage, so cost per qualified lead does not tell the whole story. Cost per loan app has historically run around $100 to $130 on our Meta ads, and I would measure ChatGPT results against that number.

Is there anything that could block an MCA offer from running?

Yes, and it is the biggest uncertainty in the whole plan: ad eligibility. OpenAI’s advertising policies say financial services advertisers may be approved on a case-by-case basis, subject to eligibility requirements. Your specific merchant cash advance offer and landing page need to be verified as eligible before you forecast delivery. That check should happen before any budget is committed.

Is there detailed targeting like the Facebook algorithm, to keep costs low?

Not in the same way. ChatGPT Ads has contextual matching, advertiser-provided context hints, and some campaign targeting controls. The main difference is that context hints help the system identify relevant conversations, while Meta gives you interest and behavior targeting that works as hard audience selection.

Here is how the two compare:

• Interest and behavior targeting: Meta offers broad options. ChatGPT works primarily through conversation context.
• Exact keyword targeting: available in some Meta campaign types. ChatGPT context hints are not exact-match keywords.
• Geographic targeting: available on both, where supported on ChatGPT.
• Custom audiences: available on Meta. ChatGPT supports them through documented audience controls.
• Automated delivery optimization: available on both, using relevance and outcome signals on ChatGPT.
• Guarantee of high-intent prospects: neither platform offers one.

Do not assume ChatGPT has the same granular audience controls or optimization maturity as Meta. The real test is whether its contextual matching produces enough qualified prospects at a cost you can accept.

What are the levers for keeping acquisition costs down?

The closest thing to targeting is giving the ad system a very clear description of the situations where your offer is relevant, then letting it match eligible conversations. Beyond that, there are five levers:

• Maximum CPC bid: set a ceiling for what you are willing to pay per valid click.
• Conversion optimization: where supported, optimize toward a conversion event instead of raw traffic.
• Landing page relevance: align the ad with the actual service and the customer need.
• Conversion tracking: measure qualified outcomes, and look past form submissions.
• Geographic and audience controls: use the supported campaign restrictions where they apply.

The $3 to $5 bid guidance is a place to start. It does not prove you will acquire leads at an economical price. The most useful first step is to set your maximum acceptable cost per qualified lead, then compare it against real campaign data once it exists.

Is there a tracking system comparable to Meta CAPI?

Yes. ChatGPT Ads now has its own Conversions API, similar in concept to Meta CAPI. You can send conversion events directly from your server instead of relying entirely on browser-based pixel tracking, and OpenAI recommends using the Conversions API when possible for more reliable measurement.

For a lead generation funnel, the useful part is that you can track the initial form submission along with later outcomes recorded by your backend. The comparison with Meta looks like this:

• Browser-side pixel: available on both.
• Server-side conversion events: available on both.
• Deduplicating browser and server events: available on both. On ChatGPT you use shared event IDs.
• Tracking events beyond the landing page: available on both, as long as your system sends the events.
• Attributing conversions to ad activity: available on both, subject to matching signals and attribution rules.
• Conversion-optimized campaigns: supported for eligible campaigns and standard events.

The same conversion should carry the same event ID from both the browser and the server, so it is not counted twice.

How would you set up tracking for an MCA lead funnel?

Four steps:

1. Ad click. Capture the oppref click reference when it is available, along with UTM parameters. OpenAI specifically recommends preserving oppref through redirects and including it in server-side events.
2. Lead submission. Fire the browser Pixel event and send the matching server event with the same event ID.
3. CRM outcome. Record whether the lead meets your qualification criteria and how far it moves through your sales pipeline.
4. Performance review. Compare what Ads Manager reports against lead quality and downstream results in your CRM.

One caveat deserves attention. Measuring a qualified lead and optimizing delivery toward qualified leads are two different things. OpenAI’s documented conversion-optimized campaigns currently require a supported standard conversion event. Custom events can be measured, but they are not currently supported as the optimization goal. So I recommend running both the Pixel and the Conversions API where feasible, deduplicating events, preserving click attribution, and validating CRM outcomes against Ads Manager reporting before leaning on the numbers for budget decisions. OpenAI’s conversion tracking overview and Conversions API technical documentation are the places to start for setup.

What happens after the merchant submits the form?

Getting the click and the form fill is half the job. The other half is what happens in the next hour. A lot of brokers lose deals here, because the merchant fills out a form, and then twelve other brokers call the same person.

Our approach is to get the merchant to submit their loan app and bank statements before you ever get on a call with them. We also automate the process, so you are not chasing the merchant while competing against a dozen other brokers. By the time you speak with them, you already have what you need to evaluate the deal. The case study and value video walk through exactly how this works.

Why move on this now?

I have seen this pattern before. When Facebook ads first came out, it was easy to dominate and take all the leads for yourself. Later, everyone jumped on the bandwagon, CPMs climbed, cost per lead went up, and being profitable got harder.

AI advertising is at the early stage of that same curve. The advertisers who start now learn how the platform behaves while competition is thin, and they build their tracking, their creative, and their follow-up process before auction costs reflect a crowded field. Waiting means paying to learn the same lessons later, at a higher price.

What if ChatGPT ads do not work out for my business?

Then we move the budget. If ChatGPT does not perform for your offer, we can set you up with my Meta ads. I am still beating the competition on Meta when it comes to MCA lead gen, so you are not betting your whole pipeline on a new channel.

How do I get started?

Start with the case study, which explains how we get your funding offer in front of merchants, how we get them to submit applications and statements before the call, and how the follow-up is automated: https://lendnet.io/proven-system

After you watch it, schedule a call with me. I am taking on only a few clients for ChatGPT ads right now.

A note on the numbers

The budget figures in this Q&A are modeled from OpenAI’s published starting-bid guidance and conversion rates from our Meta ads funnel. They are planning scenarios, and actual results depend on your offer, your eligibility approval, your landing page, and how the platform delivers. OpenAI’s ads platform is still developing, so check its current documentation before committing budget.