Pre-publication review isn’t new. How it works is.
We brought that shift out into the open in our recent webinar, “The Next Era of Pre-Publication Review.” PerformLine’s Ashley Cianci hosted a candid conversation between Katie Daley Infante, who leads enterprise sales and partnerships at PerformLine, and Kuno Tucker, a seasoned Chief Compliance Officer with years of experience building compliance programs at financial services firms. No pitch. Just two people who see this transition from different seats, talking through what’s working and what to watch for.
Here’s what they covered.
Why the shift is happening now
For years, pre-publication review meant running content against fixed if-then rules. You spell out every term and scenario in advance, then check content against that list.
Over the past 24 months, tier-one banks have pushed hard in a different direction, according to Katie: toward AI and contextual review that reads for meaning and intent instead of exact words.
The stakes behind that shift are real. PerformLine’s own benchmark research, which Katie cited on the panel, found that 65% of marketing compliance issues fall into six categories: misrepresented free offers, bait-and-switch tactics, unsubstantiated claims, deceptive promotions, outdated APRs, and deceptive guarantee language.
These aren’t rare, exotic violations. They’re the routine output of marketing teams moving fast, and they’re exactly what pre-publication review exists to catch.
From caution to “a triple win”
Kuno’s first reaction wasn’t skepticism about AI in compliance specifically. When asked how he felt hearing “AI” and “compliance review” in the same sentence, he said his instinct was caution about AI in general: some of his early fears, he noted, “have come to fruition” when AI is used without proper guardrails.
Applied to marketing review, his answer was immediate. He pointed to a previous firm where slow SLAs and back-and-forth email reviews frustrated advisors: “I’m all for it,” he said. “When I first hear about AI and marketing reviews, I’m excited. I think this will be a real game changer for a lot of firms, not just for efficiency and effectiveness, but for compliance and the advisor experience, too.” He called it a “triple win” for advisors, compliance officers, and the firm itself.
A big piece of that win is consistency. The frustration is familiar: advisors calling in and getting one answer from one reviewer, then a different answer from someone else on the same question, Kuno said. AI removes that variability, in his view. It also frees compliance staff from mundane, repetitive reviews, so they can focus on complex, higher-value work: the judgment calls that need a human.
The same inconsistency shows up on the review side, not just with advisors. A 50-page pitch deck reviewed first thing in the morning might get a different read than the same deck reviewed at 4:30 p.m., with 10 other things competing for attention, Katie pointed out. AI doesn’t get tired and doesn’t have competing priorities, she said. That consistency matters for the end consumer, and for how fairly a firm treats its own advisors.
The first line of defense is moving
Ownership of pre-publication review is moving from compliance teams alone to the content creators and advisors producing the material in the first place. The goal is to help creators produce compliant content from the start, cutting down on revision cycles later.
The friction point is what Kuno calls the “rogue advisor”: someone whose content is technically defensible but off-brand, or who has a personal view that doesn’t belong in firm-published material. AI-assisted review at the point of creation, he said, can guide advisors toward consistent brand language in real time without shutting down their voice.
That consistency also removes a subtler problem Kuno had seen firsthand: advisors picking favorite reviewers because one person “gets” their writing better than another. Katie built on the point: it’s human bias built into review, she said, even when reviewers don’t realize it’s there. Replace the person-to-person handover with a consistent standard, and disagreements shift from “I disagree with what this reviewer told me” to “let’s talk about this guideline.” That alone, she said, improves review culture.
Beyond checklists: context intelligence in practice
“Context intelligence,” in Katie’s terms, means moving past checklists. Deterministic logic handles black-and-white rules well: if a risk warning must appear within 50 words of a triggering term, a rules engine checks for that. Large language models and image classification technology extend review further, evaluating whether content communicates the right idea in the right way at scale, and whether a disclosure is legible, not just present.
One example makes the gap concrete. A firm Kuno knew of received a six-figure sanction because a disclosure’s font was too small, even though securities regulations in the U.S., Canada, and Europe don’t specify an exact size, only that disclosures must be legible. Contextual compliance can now compare a slide’s header size, body text size, and disclosure size against each other, Katie noted, catching a disclosure that’s disproportionately small relative to everything else on the page. A deterministic “must be at least 8-point type” rule would miss that.
Reputational risk came up too, particularly around political or opinionated content from individual advisors. Most firms want to stay apolitical, Kuno said, and once something publishes online, you can’t fully retract it. In his view, AI can flag or reword content before it reaches print, catching instances a single human reviewer might miss.
The practical payoff: accuracy and coverage
The tangible benefit, as Katie framed it: more accurate findings, fewer false positives compared to rigid deterministic logic, and coverage of risk categories that traditional checklists were never built to catch.
Kuno hadn’t worked with AI-enabled pre-pub review directly in prior roles. But he’d seen the underlying pattern before: human reviewers who understood context better were the ones advisors gravitated toward. That’s the same value AI aims to deliver at scale.
What to look for in a vendor
The checklist, AI-related or not: data residency and domicile, how the vendor built and tests its AI, and how responsive they are when something needs correcting. That’s Kuno’s list, drawn from years of vendor evaluations. AI can go wrong without a vendor that tests rigorously and iterates responsibly as new versions ship.
What earns sign-off in a regulated environment, Katie added: institutional knowledge, explainability, and auditability you can defend to a regulator. No black-box models. Every flagged decision needs an explanation tied to a specific regulatory statute. And auditability means you can show who approved a piece of content and why, if an examiner asks.
The build-versus-buy question is worth raising directly, too, Katie said. Enterprises that decide to build AI compliance tools in-house often underestimate the capacity problem: can their own team ship a new version every time the underlying models change, and can they match the explainability a dedicated vendor has already built in?
This isn’t abstract for Kuno. At a previous firm, he proposed building compliance tools in-house with a team he’d worked with before. They were direct with him: his project wouldn’t be a priority against existing institutional client work, and even if built, keeping pace with regulatory and marketplace changes would be difficult. The firm went with a third-party vendor instead.
Signals you’re ready, and the mistake to avoid
The signals are familiar, according to Kuno: advisors unhappy with turnaround times, compliance capacity that can’t keep pace, poor audit or examination results, and, most concerning, advisors skipping review altogether to get content out faster. That last one exposes the firm to real risk.
Growth is another trigger, Kuno noted. Whether it’s an increase in assets under management, new advisors, or a merger and acquisition, more volume doesn’t have to mean hiring another compliance officer. In his view, AI can help absorb that workload and smooth out inconsistencies in standards when two firms with different compliance cultures merge.
The mistake to avoid mirrors the vendor checklist above: avoid vendors with no track record, unclear testing practices, or ambiguous data residency.
Rolling it out without losing your advisors
Even a strong tool “can die on the vine” without the right rollout, Kuno said. His approach:
- Communicate early and often about why the firm is adopting the tool: for the firm, for compliance, and for advisors.
- Test internally with the compliance team and a small group of tech-forward “friendly advisors” who’ll give constructive feedback.
- Run a pilot with 10 to 20 advisors for 30 to 60 days, with a fast feedback loop.
- Roll out to the broader advisor population in tranches rather than all at once, especially at larger firms.
Keep everyone aligned on why the firm is implementing this in the first place, Katie added, or “you’re purchasing shelfware”: a tool nobody uses. Pushing early, direct feedback back to the vendor does double duty. It improves the product, and it builds internal champions who can speak to why the firm made the switch when the project team isn’t in the room.
There’s a related point for anyone building a business case. In Kuno’s experience, proving ROI means showing benefits beyond efficiency gains: helping recruit and retain advisors, burnishing the firm’s reputation, and smoothing M&A integration. Katie added PerformLine’s own vantage point on top of that: “seeing this technology implemented across hundreds of institutions” surfaces ROI angles a compliance team might not spot on its own.
The closing thought
The shift isn’t about AI reading faster, Ashley said in closing. “It’s about contextual judgment at scale. Still grounded in your own institutional knowledge, still explainable, and still something you can stand behind and defend to a regulator later.”
Kuno closed with the line “Slow is smooth, smooth is fast.” His point: a frictionless process isn’t slower. It’s faster in the ways that matter. Katie agreed, adding that speed for its own sake isn’t the goal. Controlled, repeatable, and accurate results are what make review fast in a way that holds up.
That’s the real question to sit with, whichever seat you’re in: not whether AI can review faster, but whether it can review the same way every time, explain itself when a regulator asks, and free your team to spend its judgment where it counts.
Want to keep this conversation going? Reach out and we’ll help you think through what a shift from rules-based to contextual review could look like for your team. No pressure, no pitch.
