Mike True, Co-Founder & CEO of Prescient AI talks about predictive AI and why marketers need to use it effectively to power measurement tactics in this MarTech catch-up:
______
Hi Mike, tell us about yourself and more about Prescient AI in brief?
I’m Mike True, Co-Founder and CEO of Prescient AI. Before starting the company, I spent my career in enterprise AI and analytics sales at companies like IBM, Oracle, and App Annie, where I helped clients generate millions in revenue through AI-powered solutions [1]. I hold a B.S. in Marketing from Salve Regina University.
I founded Prescient AI in 2019, and honestly, it didn’t start where it is today. We originally set out to build a predictive model for the music industry — helping artists figure out optimal tour schedules and venue recommendations. When COVID-19 hit, and live events disappeared overnight, we had to pivot fast. That’s when we turned our focus to solving marketing attribution and measurement challenges in e-commerce [4].
Today, Prescient is an advanced Marketing Mix Modeling (MMM) platform serving 100+ omnichannel brands — including Saatva, Hexclad, Jones Road, MaryRuth’s, and Coterie — and we raised $20M from investors including Headline and Blumberg Capital [9]. We work without pixels or cookies, deliver actionable insights within 36 hours, and can forecast future campaign performance three months out with around 90% accuracy [3]. In July 2025, we launched what we believe is the first fundamentally new MMM framework built entirely from scratch since the technology was first introduced in the 1960s [17].
How are brands today using predictive AI to power measurement and attribution tactics in modern marketing workflows?
What I’m seeing is a real shift — brands are moving away from piecing together siloed ad platform reports and toward unified, AI-driven measurement. At Prescient, our approach is built on dynamic MMM, which is a statistical, probabilistic model that ingests data from ad platforms, Shopify, Amazon, Google Analytics, and even offline sources to help brands understand which channels are truly driving revenue [14]. Unlike the old Nielsen-style annual MMM studies, our model refreshes every single day, so marketers can reallocate budgets and optimize campaigns in near real time [12].
The brands I work with are using predictive AI in a few powerful ways:
Measuring halo effects: We help brands quantify how upper-funnel channels like YouTube or CTV indirectly drive Amazon or retail sales. BrüMate, for example, discovered that nearly 20% of its CTV-driven revenue came through Amazon — something its traditional tools had completely missed [11].
Running budget simulations: Our Optimizer tool lets marketers model different budget reallocation scenarios and get predictive media plans back in seconds [12].
Triangulating measurement: I always tell our clients, don’t rely on just one source. Combine MMM with incrementality testing, post-purchase surveys, and MTA to validate performance from multiple angles [15].
Marketing Technology News: MarTech Interview With Fredrik Skantze, CEO and Co-founder of Funnel
What’s wrong with most marketing attribution norms today, and why should marketers move away from click-based attribution?
The core problem with click-based attribution — whether that’s last-click or multi-touch attribution (MTA) — is that it can only measure what can be tracked through a click. That means entire categories of high-performing channels get systematically undervalued or ignored entirely. I’ve said this many times: “MTA is great for click-based channels, but it’s not very good for something like linear TV, connected TV, podcasts, YouTube, or TikTok — where they’re more view-based channels.” [12]
There are a few other issues I see constantly:
Platform-reported ROAS is inflated and overlapping: Every channel claims credit for the same sale. I’ve seen a brand running nine channels with search ROAS ranging from 298% to 1,750% — that’s clearly impossible when channels overlap that heavily [16].
iOS 14+ and privacy regulations have broken pixel-based tracking: The signal loss is real, and it’s only going to get worse. Relying on MTA or last-click in a post-iOS 14 world means you’re making decisions on incomplete data [19].
No single tool should be your “source of truth”: I’m very direct with all of our clients about this — “The MMM is not your source of truth. The MTA is not your source of truth. Your incrementality test is not your source of truth. The source of truth is the marketer.” The best practice is triangulation [15].
My best practices around attribution:
Use MMM as your backbone for planning and forecasting.
Run incrementality tests to validate specific channel performance.
Use MTA only for bottom-funnel, click-heavy channels where it’s appropriate.
Layer in post-purchase surveys to capture what data simply can’t.
Trust the ensemble of models plus your own judgment — never a single platform’s numbers.
What are most marketers getting wrong in the AI/martech implementation process today?
From working with hundreds of brands, I see the same mistakes come up repeatedly:
Building on outdated foundations: Many martech tools out there are just modernizing old math — legacy 1960s-era regression models dressed up with a new interface. Our CTO Cody Greco and I made a deliberate decision early on that we weren’t going to do that. As Cody put it, “Building on old technology would limit our ability to solve the complex measurement challenges facing today’s marketers.” [17] That’s why in July 2025, we launched a completely new MMM framework built from the ground up.
Over-indexing on bottom-of-funnel Google and Meta: I see this all the time — the majority of budgets flowing into easily trackable, bottom-funnel channels while upper-funnel brand investments go unmeasured and underinvested [7].
Trusting a single measurement source: Any marketer who relies on just platform-reported ROAS or a single attribution tool is seeing a dangerously incomplete picture.
Accepting slow time-to-value: Legacy solutions have been clunky, expensive, and take months to onboard. That’s not good enough anymore. Brands should expect insights within 36 hours and point-and-click integrations [13].
Ignoring Amazon and retail revenue in media models: For omnichannel brands, a massive portion of revenue has historically been completely disconnected from paid media measurement. That’s a blind spot we’ve been on a mission to fix [18].
As martech evolves and old marketing models are replaced, what trends do you think will reshape B2B SaaS marketing and the martech ecosystem going forward?
This is something I think about a lot. Here’s where I see things heading [9]:
Predictive models will replace cookies as the primary measurement lens: The demise of third-party cookies and the tightening of GDPR and CCPA regulations mean that privacy-compliant, statistical models like MMM will become the default infrastructure. The privacy-first internet isn’t a future scenario — it’s already here [19].
Advertising on autopilot: The era of semi-manual media buying is fading fast. AI-driven models will increasingly recommend and execute optimizations automatically, with back-tested confidence. I see a future where brands are running fully automated, dynamically optimized campaigns across every channel [8].
The rise of “Compound AI”: The next phase of MMM involves multiple specialized AI agents — for forecasting, creativity, audience analysis, and saturation analysis — collaborating continuously to deliver adaptive recommendations in real time [14].
In-house brand teams replacing agencies: As automation handles media buying, I believe brands will build lean, data-empowered in-house teams and reduce their dependence on traditional agencies. The tools are now accessible enough to make that happen [8].
AI agent advertising: This one fascinates me. As more search and commerce shifts to LLMs and AI agents rather than Google, entirely new attribution models and ad formats will have to emerge for the “agent world” [7].
Consolidation in martech: We’re already seeing the big players like Publicis actively looking to acquire AI measurement companies — Prescient was cited as a potential target — which tells you how seriously the industry is taking this shift [1].
Five martech and marketing best practices you’d leave us with before we wrap up?
These are the things I come back to again and again with every brand I work with:
1. Triangulate your measurement — Never put all your trust in a single attribution model. Use MMM, incrementality testing, MTA, and post-purchase surveys together. Each one tells part of the story [15].
2. Give upper-funnel channels a fair shot — Stop penalizing CTV, podcasts, or YouTube for not generating clicks. Use view-based and halo-effect measurement to reveal what they’re really contributing to your bottom line [12].
3. Future-proof your martech stack against data loss — Build infrastructure that doesn’t depend on cookies or pixels. That transition is already underway, and brands that wait will be caught flat-footed [19].
4. Demand speed from your measurement tools — If your MMM only updates quarterly, it’s too slow to be useful. Modern platforms should recalibrate daily and deliver insights within 36 hours. Agility is the whole point [12].
5. Remember: you are the source of truth — AI models are incredibly powerful advisors, but they don’t replace the marketer. You’re the one synthesizing data, culture, creativity, and business context to make the final call. Lean on the models, but trust yourself [15].
Marketing Technology News: The Death of Third-Party Cookies Was Just the Start. Are You Ready for Consent Orchestration?
Prescient AI’s platform delivers fast, daily marketing insights across the full funnel: quantifying halo effects, isolating seasonal impact, and identifying the real levers of performance across DTC ecommerce, marketplaces, and retail.
About Mike True
Mike True, is Co-Founder & CEO of Prescient AI
Source References
[1] Prescient AI – About Us & Meet the Team – November 7, 2025
[3] Prescient AI LinkedIn
[4] EP007: Michael True | Equation of Excellence / Fermat Commerce
[7] 5YF Episode #28 Transcript – Focal VC
[8] 5YF Episode #28 – Focal VC Founder Resources
[9] Future of Advertising with Prescient AI CEO Mike True | 5YF #28 – April 14, 2025
[11] What is Media Mix Modeling (MMM)? – Prescient AI Blog – September 23, 2025
[12] S12 E6: Redefining Media and Marketing Measurement – Limited Supply Podcast – May 14, 2025
[13] Investing in Prescient AI: Next-Gen MMM – Headline VC
[14] MMM Explained – Darkroom Agency Observatory
[16] Marketing Mix Modeling: How Multi-Channel Brands Stop Wasting Ad Spend – Ecommerce Coffee Break – August 13, 2025
[17] Prescient AI Unveils First New MMM Since 1960s – PPC.land – July 16, 2025
[18] Prescient AI LinkedIn – Amazon Measurement Launch
[19] Future of Advertising with Prescient AI CEO Mike True | 5YF #28 – April 14, 2025