In the face of AI, marketers are rethinking their tech stacks
Sarah Kim / Shutterstock / The Current
Marketers have long faced pressure to streamline their tech stacks, but today that urgency has reached a new level. The economic climate is one factor; and AI’s rapid evolution is forcing teams to scramble, juggling client demands and C-suite expectations.
The challenge is the number of platforms in a typical stack and how to integrate them into one cohesive system. In a recent Forrester survey, 65% of marketers reported using 16 or more solutions, while 70% said finding audiences across multiple touchpoints is harder than ever.
“Layering more tech on top of an already fragmented ecosystem typically makes the problem worse,” TransUnion EVP Matt Spiegel wrote in an op-ed on The Current earlier this year. “More tools mean more data silos, more inconsistencies and more time wasted reconciling conflicting information.”
Or, as one agency lead tells The Current, “The tech can be amazing, but if it can’t measure right, it’s pointless.”
Heather Macaulay, president of data consultancy firm MadTech, estimates that fewer than 20% of platforms are truly integrated and interoperable.
This has been a long-standing problem, but AI is compounding the problem.
“One of the hard things about the AI explosion and landscape is it’s moving so fast and everyone wants answers,” Razorfish Chief Strategy Officer Eddie Gonzalez tells The Current. “Agencies have to be one step ahead on understanding those tools as they’re evolving.”
Brands are seeking guidance on which AI tools have staying power, so they don’t buy, train and integrate a new platform only for a better option to supersede it six months down the road.
“The hardest part is this isn’t new technologies and old paradigms; this is new paradigms and new technology,” Jeff Geheb, CEO of VML Enterprise Solutions, tells The Current. “And I think a lot of our clients have almost not even fully felt the blow of it yet.”
Tracking business transformations
Geheb led a report that studied 4,000 business leaders across eight countries on their digital transformations. Across industries such as retail, auto and telecom, companies reported spending an average of $10.9 million per year, juggling two digital transformation initiatives annually.
Still, 37% of those efforts fail, often due to weak leadership and vague road maps, says Hugh Fletcher, VML Enterprise Solutions’ marketing director and co-author of the study. He says the same issues crop up across industries.
AI’s emergence is now a main tenet in companies’ plans, with 77% of respondents saying AI forced them to revisit their strategies. Sixty-four percent of respondents said the pace of AI’s evolution makes it difficult to make long-term decisions.
“This catalyst came in and almost picked up what they were doing and then smashed it into the future in a profound way,” Geheb says. “My general view is most of the companies we work with are not even in mile one of a 26-mile marathon of figuring out what this impact is.”
AI isn’t a magic wand
MadTech’s Macaulay warns that without a strong data backbone, AI won’t deliver. Inaccurate data creates inaccurate outcomes.
“We expect AI to serve as a magic wand that can fill any hole for us,” she wrote in an op-ed for The Current. “It doesn’t work that way.”
The solution? Properly structured data.
“Without [it], retailers make multimillion-dollar real estate decisions based on flawed data that tracks consumer patterns, quick-service restaurants allocate inventory using incomplete purchase histories and financial services companies design products that miss the mark on customer needs.”
Fixing that data foundation involves making it privacy-safe, consented, standardized and interoperable. All those steps help turn data silos into accessible, unified profiles that can be immediately activated. From there, she says integrating new tools can shrink from months to days.
Identity is the connective tissue
A key piece of the puzzle is identity. Earlier this year, LiveRamp executive Travis Clinger compared identity solutions like Unified ID 2.0, RampID and ID5 to the Rosetta Stone: a tool that decodes fragmented data — hieroglyphs or otherwise — across platforms into a common, actionable language.
Spiegel calls identity resolution “the connective tissue between disconnected systems, enabling brands to unify customer data across platforms and channels.”
“Instead of just collecting data,” he wrote, “you’re activating it to power smarter, more precise marketing strategies.”
Don’t fit a square peg into a round hole
As marketers consider the elements of their tech stack, from discovery to testing to integration and implementation, Razorfish’s Gonzalez has simple advice: Make sure there’s a business problem that the platform solves with specific use cases. In short, don’t try to fit a square peg into a round hole, he says.
A case in point: Razorfish deployed AI to solve a core problem for its pharma clients: legal compliance around creative messaging. By training AI on past legal feedback, the agency shortened feedback loops between creative and legal by 15%.
The breakthrough started with identifying the initial problem — strict regulatory friction around the pharmaceutical industry.
“Love the problem, and you’ll figure out the solution,” Gonzalez says.