The Frank Reality of AI’s Impact on Marketing According to Two Agency Executives
We sat down with two agency executives—Chris Parrett, CEO at Social Hustle, and Rupesh (Pesh) Sharma, VP Client Services at Realtime Agency—to talk about the reality of AI, how they’re using it at their agencies, and their advice on how to balance a mandate to innovate with the pressure to perform.
AI is here to help, but marketers need breathing room
Budgets are shrinking, and marketing teams are leaner. With consumer confidence still uncertain, companies have continued to seek every opportunity to cut costs.
For many, AI is an obvious solution. But an impulse rush into AI dependency comes with its own risks.
“There's a lot more scrutiny on budget and pressure to get performance out of it,” says Pesh. “Marketers know AI tools can help them, but their focus is on short-term results.”
Why? Because it has to be. While AI adoption is increasing—56% of marketers say their company is actively implementing and using AI—there’s also no denying that the emphasis on short-term results is discouraging marketers from experimenting as much as they’d like. At the same time, many CMOs are hesitant to greenlight AI due to concerns around security, control, and transparency.
It’s a tough spot: Marketers see AI as a tool to help them meet their performance goals, but they lack both the resources to test AI use cases and the organizational buy-in they need to keep momentum going.
Using impact-to-revenue to prioritize AI adoption
With tighter budgets, thinner margins, and rising pressure to prove ROI, leadership teams have their sights set on one thing: the bottom line.
“Leadership teams get excited about anything that speaks to bottom-line benefit,” Chris shares. “If we can get to the point where we're talking about impact-to-revenue versus impact-to-platform, that's where people start to lean in.”
In other words, AI won’t get a seat in the boardroom if you only frame it as a button you can press to automate surface-level tasks like keyword research or analyzing campaign performance. For AI to stick internally, it needs to be connected to real business impact specific to revenue growth, cost-efficiency, or other financial outcomes.
How Chris and Pesh are using AI to improve marketing outcomes
With smaller teams, tighter deadlines, and mounting pressure, marketers have little room for guesswork—especially when it comes to adopting something as new as AI. Chris and Pesh are no exception, which is why they both lean into proven use cases.
“At Social Hustle, we’re using AI to spot trends, analyze data, understand traffic and interests, and make strategic recommendations,” Chris explains. “These are tasks we’d typically assign to a team of analysts, but with AI we can make decisions in a quarter of the time.”
When applied to the right processes, AI can become an enabler and accelerator that helps already-strapped teams do more.
Here are a few more use cases worth considering:
Faster Time to Insight
Pesh: “We sometimes get seemingly simple questions that end up taking us quite a long time to answer. Even just gathering the right data from various platforms can take hours - let alone time spent on cleaning it, using vlookups and pivot tables to combine and analyze… it’s a lot.”
Which is why those questions often go unanswered, or even un-asked.
One client asked whether the weather in their primary geographies (UK) affected app downloads.
So, simple question: “Do people download our app more when it’s rainy?”
Surprisingly difficult to answer. Unless you have:
- Access to the right data
- A way to easily combine those data sources so they’re ready for analysis
- A way to perform the analysis
Which, it turns out, Pesh’s team did have.
Pixis’ Prism accesses Meta Ads performance data and weather data in real time to not only perform the analysis, but also propose a game plan to optimize bids and budget by geography depending on the forecast.
Audience experimentation
“It used to be that you could build custom audience segments and lookalikes based on user lists,” Chris says. “But Google and Meta are moving away from that. Now, it’s all about broad campaigns, data, signals, and letting their algorithms optimize on that.”
That’s precisely the problem. As Google, Meta, and others tighten privacy restrictions, marketers are left with weaker and less reliable signals to build audiences.
According to both Chris and Pesh, it’s time to rethink your audience targeting playbooks, starting with data enrichment tools like Elevar. “When I’ve launched Elevar with brands, we’ve seen a 15–20% lift in incremental return on ad spend (ROAS) within 24 hours,” he explains.
Then, with the enriched data built from product and order history, customer profiles, and attribution insights, you gain a more comprehensive view of your audience. From there, you can layer in AI to analyze patterns, spot trends, uncover untapped audience segments, and pinpoint the people most likely to convert.
Creative testing
Rising advertising costs, limited visibility, broken attribution, and walled gardens have forced marketers to question certain best practices of campaign design—and inspired a wave of experimentation.
Marketers are now testing everything they can, but Pesh recommends starting with ad creative.
“Creative is the most important element,” he says. “As audiences get broader, engaging ads become the way to define and reach them.”
Since brands can no longer rely on in-platform targeting to reach niche audiences, ad creative has to do the heavy lifting—and it’s on marketers to figure out the best way to encourage clicks.
AI can assist in experimenting with different creative formats—videos, static ads, or user-generated content (UGC)—as well as headlines and calls to action (CTAs). If you want to take it a step further and eliminate ad fatigue for good, you can use Dynamic Creative Optimization (DCO) to automatically tailor ads to specific users in real-time.
Channel testing
Facebook, Instagram, and Google still consume the majority of paid ad budgets, but relying too heavily on one or two platforms is risky business—especially as costs rise, control tightens, and performance visibility wanes.
Chris points to Google as the prime example. “With AI-generated results becoming more commonplace in search results, paid search traffic and conversions are dropping.” The result? Marketers are paying a premium to reclaim the same real estate on Google that they used to get for a discount.
As channels become more expensive and locked down, your best bet will be to spread your spend across channels. AI tools like Pixis can help by analyzing performance history, creative trends, and other external signals to pinpoint emerging opportunities like TikTok in 2020, monitor performance across channels, and reallocate spend to the channels that are moving the needle.
3 keys to AI adoption: Attribution, incrementality, and ROI
As American composer and economist W. Edwards Deming once said, “Without data, you're just another person with an opinion.”
When it comes to AI adoption, his quote holds true. Bold plans and big promises won’t win over leadership. Here’s what will:
Attribution
Diversifying channels, creative, and audiences can set up for better marketing results. But they can also make those results harder to measure.
“Savvy marketers know they need to diversify to grow,” Pesh says. “But without a measurement strategy, there’s no way to know if their budget is going to the right places.”
And that challenge is only intensifying, according to Chris. “We're watching the degradation of data signals,” he explains. “Marketers need better signals, better attribution, and better data.” That’s why attribution platforms like Rockerbox, Domo, Triple Whale, and Northbeam have become staples of Chris’ agency playbook.
But attribution platforms alone aren’t enough. You also need the right strategic foundation:
- Start with clear goals: Define your business objectives upfront and select key performance indicators (KPIs) that align with them.
- Connect the right sources: Ensure your tech stack, including social ads, email, website analytics, and your CRM, all work together to tell the same story.
- Keep your data clean: Attribution is only as accurate as the data behind it. Regularly audit your marketing channels to check that tracking pixels, UTM parameters, and conversion events are working as intended.
Pesh also points out that attribution is as much about depth as it is visibility. “The most important thing is having a measure of quality,” he says. “Even if you're looking at CTR, you have to push deeper to figure out what people are doing after the click.”
Incrementality
Attribution tells you where conversions happened; incrementality shows the lift your tactics create beyond what would have happened anyway.
With tighter budgets and more experimentation, isolating real growth from the noise is key. As marketers move further up the funnel, Pesh argues that clarity is even more important.
“If you're an advertiser and you want to grow, you have to go further up the funnel,” Pesh shares. “The problem with that is it's tough to understand the incrementality from a lot of mid- or upper-funnel activities.”
It’s not the flashiest topic, Pesh admits. But if you want to grow and invest in broader strategies, you need to pair incrementality testing with marketing mix modeling that uses regression analysis to estimate how marketing, sales, and external factors relate over time.
Incrementality can also help you flag instances of campaign cannibalization when one tactic takes credit for results another would have delivered on its own.
Return on investment
You can have airtight attribution and perfect incrementality models. But if your efforts aren’t tied back to business outcomes, it’s just noise.
As Pesh puts it: “If there's no return on the investment, the investment’s going to get cut.”
That mindset is a guiding principle of Chris’ agency, too. “It’s mind-blowing how many companies are investing in activities that don’t generate revenue,” he says. “If it doesn't add money to the bottom line, it just doesn't matter.”
That holds true when it comes to AI adoption. Every decision, test, and tactic needs to roll back into ROI—even if it goes against conventional wisdom, like pausing a viral UGC ad that’s not converting, replacing “trendy” creative with “simple” Dynamic Product Ads (DPA), or cutting ads with high CTR because they haven’t translated to any sales.
Today and tomorrow: The role of AI in marketing
When deployed to the right marketing use cases, AI becomes an accelerator that increases the speed of decision-making, makes experimentation more meaningful, and helps teams allocate their budgets in the most effective way possible.
“AI is freeing my team to work on the stuff that I want them working on,” Chris says. “We're using it to give us decision-making power to go and chase budgets as fast as humanly possible.”
But moving fast with AI doesn’t mean much if the outcomes don’t move the business forward. That’s why the future of AI is less about automating every lever possible, but about intentionally adopting it in a way that’s sustainable, measurable, and embedded into your playbook.
That’s exactly what Pixis helps performance marketers do: tap into their own data to experiment at scale, uncover practical insights, and take confident AI-powered actions—with human oversight—that move their business forward.
Want to learn more about how AI can help you run experiments at scale across audiences, creatives, and channels? Let’s talk.
