Andrew Palosi, CEO of Ad Leverage and CEO of datacube.ai, holds that AI belongs on data analysis, research, and operational support, and does not belong on content, collateral, or anything meant to convey original thought leadership. Palosi’s evidence is arriving at his agency as inbound work: clients whose lead generation and revenue have fallen off a cliff, who do not know why, and whose traffic collapsed after publishing repurposed AI content into a search ecosystem built on authority, experience, and trustworthiness.
Palosi has built and led Ad Leverage for over 18 years, scaling it to a cross-functional team of 70 professionals running traditional and digital campaigns. He incubated datacube.ai inside that agency and now serves as its CEO, running a platform that automates cross-platform computations for custom KPI dashboards for hundreds of clients.
Glenn Gow, The Scaling Executive Coach and host of The Scaling Executive Podcast, identifies Palosi’s function-by-function split as the operating rule other CEOs should copy: AI everywhere the output supports a decision, and nowhere the output claims to be original.
This episode is for CEOs deciding how far to push AI into their marketing function, particularly those whose teams are already generating client-facing content with it.
Key Takeaways
- The test for any AI use is whether the output will be published as original and authoritative, and Andrew Palosi applies that test rather than judging AI by what the tools are technically capable of producing.
- Search engines have accelerated their detection of repurposed content, and Andrew Palosi points to more Google updates in the last three years than in roughly the 13 years preceding them as the measure of that shift.
- Vendors promising volume are the specific risk, and Andrew Palosi names bad actors making big promises about generating X amount of content over X amount of time as the source of the damage he sees in client businesses.
- The consumer detection threshold has already moved, and Andrew Palosi marks it by his own children replacing “that’s fake” with “that’s AI.”
- AI earns its place in a services business on computation rather than creation, and Andrew Palosi’s datacube.ai automates cross-platform KPI calculations in real time for hundreds of clients after one media client described the problem out loud.
CEOs Split AI by Function: Operations Yes, Original Content No
Andrew Palosi runs an agency that uses AI daily and refuses it in one category. Ad Leverage applies AI to data analysis, aspects of research, and operational support, where Palosi describes it as an accelerator and an assistant for his 70-person team. Ad Leverage does not apply AI to content creation, collateral creation, or anything intended to carry real originality or real emotional connection.
The dividing line is not the technology’s capability. It is whether the output is claiming to be original and authoritative. Palosi calls that use of AI misguided and dangerous, and his reasoning is that repurposing someone else’s content is not the route to authority, while authority is precisely what Google values.
| Category | Palosi’s rule at Ad Leverage | Why the line falls there |
| Data analysis and research | Use AI | The output supports a human decision rather than claiming to be original work |
| Internal operations | Use AI | Palosi describes AI as an operational support structure and assistant for the Ad Leverage team |
| Content and collateral | Do not use AI | The output is published as original and authoritative, which is the claim search engines and consumers now test |
| Anything conveying emotional connection | Do not use AI | Human readers actively scan for AI generation, in Palosi’s account, alongside the search engines doing the same |
Search Engines Now Detect Repurposed Content Faster Than Agencies Can Publish It
Andrew Palosi measures the shift in update frequency rather than in algorithm names. He points to more Google updates in the last three years than in roughly the 13 years preceding them, and reads the pattern as an explicit message to anyone repurposing content and having agents rewrite it, then presenting the result as authoritative.
Palosi ties that acceleration back to what Google’s ecosystem is built to reward: true authority, demonstrated experience, and true trustworthiness.
“And so if everybody’s repurposing the same content, that goes away.”
The mechanism follows from that sentence. If every company publishes material derived from the same source, authority, experience, and trustworthiness stop being distinguishable, and the ranking system built to identify them stops finding anything to identify.
Palosi sees the consequence arrive as new business at Ad Leverage. Clients come to the agency after results have fallen off a cliff, describing the lead generation and revenue they used to produce and asking for help understanding where it went.
CEOs Protect Revenue by Screening Vendors Who Promise Content Volume
Andrew Palosi names the specific actor causing the damage he sees in client businesses: vendors making big promises about generating a set amount of content over a set amount of time. He says that work does nothing but harm the business.
The proposal that should trigger a CEO’s scrutiny is therefore the one priced and scoped by volume, because volume is the metric a repurposing operation can deliver and an authority-building operation cannot. A vendor who can commit to a monthly output number before understanding the business is describing a process that does not depend on the business.
Palosi also marks how fast the audience side of this moved. His own children no longer say something is fake. They say it is AI. Detection is no longer a specialist capability held by search engines, and Palosi is direct that human readers are actively looking for AI-generated material at the same time the search engines are.
CEOs Point AI at Computation Their Clients Cannot Do Manually
Andrew Palosi built datacube.ai to show where AI belongs, and the origin explains the rule. A media client of several years standing kept describing the same frustration to the Ad Leverage team: he had KPIs and numbers spread across different platforms and spreadsheets, and he could not identify who in his aggressively growing company was performing and who was not.
Palosi and one team member began working the problem with Ad Leverage’s programming and development staff, asking whether they could pull each disparate data source in and cross-compute the results. They delivered the calculations across platforms in one place and in real time through automation, and the agency’s designers built the display the client put on the TVs on his walls.
“So it was through just troubleshooting and listening that we built the solution that is now Datacube.”
The work datacube.ai performs is exactly the work Palosi permits AI to do: computation across sources that produces a number a human then acts on. It makes no claim to originality and asserts no authorship. That is why Palosi will run his agency’s analytics on it and still refuse to let AI write a client’s thought leadership.
CEOs Apply These Four Principles to Use AI Without Damaging Their Brand
| Principle | What it means in practice | Named evidence from this interview |
| Judge AI by the claim, not by the capability | Permit AI wherever the output supports a human decision, and refuse it wherever the output will be published as original and authoritative work | Palosi runs AI across analysis, research, and operations at a 70-person agency while keeping it out of every client-facing deliverable, and built datacube.ai as the computational side of that same rule |
| Repurposed content destroys the signal the ranking system exists to find | Authority, experience, and trustworthiness cannot be demonstrated with material every competitor also published, so repurposing removes the basis for ranking | Palosi points to more Google updates in the last three years than in roughly the 13 years preceding them, an acceleration he reads as Google closing the gap on content repurposed and rewritten by agents |
| The audience detects it before the algorithm does | Consumer skepticism now arrives ahead of any ranking penalty, so content that reads as generated loses trust before it loses traffic | Palosi’s own children have replaced “that’s fake” with “that’s AI,” and he describes human readers scanning for AI-generated material in parallel with the search engines doing the same |
| AI pays for itself on computation, not on creation | Point AI at cross-source calculation that no person can perform manually at speed, where accuracy is the deliverable and originality is not claimed | One media client’s inability to identify performers across scattered platforms and spreadsheets became datacube.ai, which now automates cross-platform KPI computation in real time for hundreds of clients |
Quotes from This Episode
- “Where my kids now instead of saying that’s fake, they say that’s AI.” — Andrew Palosi, CEO, Ad Leverage and datacube.ai
- “And with every update, they’re getting better and better at sniffing this out.” — Andrew Palosi, CEO, Ad Leverage and datacube.ai
- “And I think that the landscape, the capabilities, the power, the intelligence is evolving so rapidly.” — Andrew Palosi, CEO, Ad Leverage and datacube.ai
- “I don’t think we can have very many conversations amongst colleagues or professionals without the topic of AI coming up right now.” — Andrew Palosi, CEO, Ad Leverage and datacube.ai
- “So now a few years later, hundreds of clients on the platform.” — Andrew Palosi, CEO, Ad Leverage and datacube.ai
Frequently Asked Questions
Where should a company use AI in marketing and where should it avoid AI?
Andrew Palosi, CEO of Ad Leverage and datacube.ai, uses AI for data analysis, aspects of research, and operational support across his 70-person agency, and excludes it from content creation, collateral creation, and anything meant to convey original thought leadership or emotional connection. His dividing line is whether the output will be published as original and authoritative, because that is the claim both search engines and human readers now test. Palosi describes AI as a phenomenal operational resource and an amazing support structure, while calling its use for original thought leadership dangerous.
Why do companies lose search traffic after publishing AI-generated content?
Andrew Palosi points to a sharp acceleration in Google updates, noting more in the last three years than in roughly the 13 years preceding them, aimed specifically at content that has been repurposed and rewritten by agents then presented as authoritative. His explanation is that Google’s ecosystem is built on true authority, true experience, and true trustworthiness, and that if everyone repurposes the same content those qualities disappear from the pool of ranked material. Palosi sees the result arrive as inbound business at Ad Leverage, from clients whose lead generation and revenue fell off a cliff without explanation.
How should a CEO evaluate a content vendor pitching AI-assisted production?
Andrew Palosi identifies volume promises as the warning sign, naming bad actors who commit to generating a set amount of content over a set amount of time as the direct cause of the damage he sees in client accounts. A CEO should treat any proposal scoped by output quantity as a signal that the process does not depend on the specific business, because authority cannot be manufactured at that rate. Palosi’s own agency separates the decision by function instead, permitting AI on analysis and operations while keeping every client-facing deliverable human.
CEOs Work with Glenn Gow to Scale Their Companies by Scaling Themselves First
Glenn Gow is The Scaling Executive Coach. He coaches ambitious executives into the CEO seat and CEOs into successful exits. With 25 years as a CEO and 5 years in venture capital, Glenn helps leaders scale their companies by scaling themselves first. If this conversation was useful, you can apply for executive coaching with Glenn Gow or apply to be a guest on The Scaling Executive Podcast.
