• LinkedIn
• Twitter
• Instagram
• YouTube
• RSS Feed

Apple •Spotify• Pocket Casts •Youtube •Overcast •RSS

What’s up everyone, today we have the pleasure of sitting down with Ankur Kothari, Adtech and Martech Consultant who’s worked with big tech names and finance/consulting firms like Salesforce, JPMorgan and McKinsey.

Summary: Ankur explains how most AI personalization flops cause teams ignore the basics. He helped a brand recover millions just by making the customer journey actually make sense, not by faking it with names in emails. It’s all about fixing broken flows first, using real behavior, and keeping things human even when it’s automated. Ankur is super sharp, he shares a practical maturity framework for AI personalization so you can assess where you currently fit and how you get to the next stage.

In this Episode…

• AI Personalization That Actually Increases Retention
• Step by Step Approach to AI Personalization With a Maturity Framework
• Fix the Stack You Have Before You Buy Another Tool
• How to Stay Sane When Every Martech Platform Thinks It’s the AI Brain
• The Only Team Structure That Actually Scales Personalization
• Stop Pretending Your Data Is Ready for AI
• Balancing AI Personalization With Human Context
• Predictive Loyalty Engines are Coming to a Martech Stack Near You

AI Personalization That Actually Increases Retention

Practical Example

Most AI personalization in marketing is either smoke, mirrors, or spam. People plug in a tool, slap a customer’s first name on a subject line, then act surprised when the retention numbers keep tanking. The tech isn’t broken. The execution is lazy. That’s the part people don’t want to admit.

Ankur worked with a mid-sized e-commerce brand in the home goods space that was bleeding revenue; $2.3 million a year lost to customers who made one purchase and never returned. Their churn rate sat at 68 percent. Think about that. For every 10 new customers, almost 7 never came back. And they weren’t leaving because the product was bad or overpriced. They were leaving because the whole experience felt like a one-size-fits-all broadcast. No signal, no care, no relevance.

So he rewired their personalization from the ground up. No gimmicks. No guesswork. Just structured, behavior-based segmentation using first-party data. They looked at:

• Website interactions
• Purchase history
• Email engagement
• Customer service logs

Then they fed that data into machine learning models to predict what each customer might actually want to do next. From there, they built 27 personalized customer journeys. Not slides in a strategy deck. Actual, functioning sequences that shaped content delivery across the website, emails, and mobile app.

“Effective AI personalization is only partly about the tech but more about creating genuinely helpful customer experiences that deliver value rather than just pushing products.”

The results were wild. Customer retention rose 42 percent. Lifetime value jumped from $127 to $203. Repeat purchase rate grew by 38 percent. Revenue climbed by $3.7 million. ROI hit 7 to 1. One customer who previously spent $45 on a single sustainable item went on to spend more than $600 in the following year after getting dropped into a relevant, well-timed, and non-annoying flow.

None of this happened because someone clicked “optimize” in a tool. It happened because someone actually gave a damn about what the customer experience felt like on the other side of the screen. The lesson isn’t that AI personalization works. The lesson is that it only works if you use it to solve real customer problems.

Key takeaway: AI personalization moves the needle when you stop using it as a buzzword and start using it to deliver context-aware, behavior-driven customer experiences. Focus on first-party data that shows how customers interact. Then build distinct journeys that respond to actual behavior, not imagined personas. That way you can increase retention, grow customer lifetime value, and stop lighting your acquisition budget on fire.

Back to the top ⬆️

Why AI Personalization Fails Without Fixing Basic Automation First

Signing up for YouTube ads should have been a clean experience. A quick onboarding, maybe a personalized email congratulating you for launching your first campaign, a relevant tip about optimizing CPV. Instead, the email that landed was generic and mismatched—“Here’s how to get started”—despite the fact the account had already launched its first ad. This kind of sloppiness doesn’t just kill momentum. It exposes a bigger problem: teams chasing personalization before fixing basic logic.

Ankur saw this exact issue on a much more expensive stage. A retail bank had sunk $2.3 million into an AI-driven loan recommendation engine. Sophisticated architecture, tons of fanfare. Meanwhile, their onboarding emails were showing up late and recommending products users already had. That oversight translated to $3.7 million in missed annual cross-sell revenue. Not because the AI was bad, but because the foundational workflows were broken.

The failure came from three predictable sources:

• Teams operated in silos. Innovation was off in its own corner, disconnected from marketing ops and customer experience.
• The tech stack was split in two. Legacy systems handled core functions, but were too brittle to change. AI was layered on top, using modern platforms that didn’t integrate cleanly.
• Leaders focused on innovation metrics, while no one owned the state of basic automation or email logic.

To fix it, Ankur froze the AI rollout for 120 days and focused on repair work. The team rebuilt the essential customer journeys, cleaned up logic gaps, and restructured automation to actually respond to user behavior. This work lifted product adoption by 28 percent and generated an additional $4.2 million in revenue. Once the base was strong, they reintroduced the AI engine. Its impact increased by 41 percent, not because the algorithm improved, but because the environment finally supported it.

“The institutions that win with AI are the ones that execute flawlessly across all technology levels, from simple automation to cutting-edge applications.”

That lesson applies everywhere, including in companies far smaller than Google or JPMorgan. When you skip foundational work, every AI project becomes a band-aid over a broken funnel. It might look exciting, but it can’t hold.

Key takeaway: Stop using AI to compensate for broken customer journeys. Fix your onboarding logic, clean up your automation triggers, and connect your systems across teams. Once the fundamentals are working, you can layer AI on top of a system that supports it. That way you can generate measurable returns, instead of just spinning up another dashboard that looks good in a QBR.

Step by Step Approach to AI Personalization With a Maturity Framework

The First Steps You Can Take on The Path To AI Personalization

Most AI personalization projects start with a 50-slide vision deck, three vendors, and zero working use cases. Then teams wonder why things stall. What actually works is starting small and surgical. One product. One journey. Clear data. Clear upside.

Ankur advised a regional bank that had plenty of customer data but zero AI in play. No need for new tooling or a six-month roadmap. They focused on one friction-heavy opportunity with direct payoff: mortgage pre-approvals. Forget trying to personalize every touchpoint. They picked the one that mattered and did it well.

They built a clustering algorithm using transaction patterns, savings trends, and credit utilization to detect home-buying intent. From there, they pushed pre-approvals with tailored rates and terms. The bank already had the raw data in its core systems. No scraping, no extra collection, no “data enrichment” vendor needed.

That decision paid off fast:

• The data already existed, so implementation moved quickly
• The scope was limited to a single high-stakes journey
• The impact landed hard: mortgage application rates jumped 31 percent and approval-to-close conversions climbed 24 percent within 60 days

“Start with a high-value product journey where personalization can meaningfully improve both customer experience and business outcomes.”

Ankur framed this not as a hack, but as a discipline. Skip the all-or-nothing mindset. Pick one product, one flow, and get it performing. Once the numbers move, teams lean in. Budget follows. Expansion becomes execution, not theory.

Key takeaway: AI personalization should begin with a single high-impact journey that already has quality data and clear behavioral signals. Focus on a contained use case where tailored communication can drive measurable lift. That way you can prove value early, reduce risk, and build a practical roadmap for broader rollout. Starting small is not a limitation. It is the smartest path to scale.

Personalization Maturity Framework

Most teams love the idea of AI orchestration. They skip steps, throw money at martech vendors, and hope some machine will finally start sending the “right message at the right time.” Meanwhile, basic segmentation is still running on decade-old personas, and every email still opens with the same lifeless copy. What we’ve built with Ankur is a better blueprint, one that earns every step by making it work before moving up.

He’s rolled out a similar framework inside a national bank, and each phase had to hit business goals before advancing. 5 practical stages that compound real value:

1. Behavioral triggers that prompt a clear next action
2. Microsegmentation based on what people actually do
3. Modular content that adapts in tone, timing, and detail
4. Prediction grounded in life events
5. AI orchestration that adapts journeys in real time

Step 1: Behavioral triggers that prompt a clear next action

This starts simple. Trigger an email when a customer completes a key milestone, like setting up an investment account. In Ankur’s case, this single automation increased follow-up deposits by 23 percent. It wasn’t sexy, but it worked. The goal here is momentum. Use behavior to keep the user moving.

Step 2: Microsegmentation based on what people actually do

Forget age ranges or job titles. Ankur mapped customer segments using transaction data and digital activity. This created tailored journeys for conservative savers and high-risk investors. Product adoption improved by 31 percent. This step separates companies that guess from those that learn.

“You cannot personalize at scale until your segments reflect behavior, not assumptions.”

Step 3: Modular content that adapts in tone, timing, and detail

Ankur’s team built a dynamic content system using modular blocks that responded to customer preferences and financial maturity. One user gets a weekly portfolio snapshot. Another prefers quarterly updates with plain language. Respecting communication preferences dropped opt-outs by 37 percent. Content that adapts builds trust.

Step 4: Prediction grounded in life events

Once the foundation worked, they added predictive models. These used transaction patterns to detect life changes like marriage, retirement prep, or a new dependent. That triggered next-best offers tuned to timing and need. Relevance jumped 42 percent. Conversion climbed 28 percent. This layer made outreach feel intuitive instead of interruptive.

Step 5: AI orchestration that adapts journeys in real time

Only after those four layers held up did they introduce orchestration. Machine learning adjusted the entire experience based on ongoing behavior. This step added 24 percent to retention and increased share-of-wallet by 29 percent. The AI worked because the system beneath it already had structure, logic, and trust.

The entire stack only moved forward when the previous layer delivered measurable impact. That discipline is what separates a working personalization program from a flashy one.

Key takeaway: Build personalization like a ladder. Anchor each step with performance before moving up. Start with clear behavioral triggers. Upgrade to segments based on actions, not demographics. Deliver modular content that adapts to preference. Use predictive models driven by life context. Then, and only then, introduce orchestration. That way you can create personalization that compounds in value instead of collapsing under hype.

Stop Automating Tests Before You Learn How to Run Them

Testing is one of those things every marketing team claims to do, but few do well. Some throw in a subject line test at the last minute and call it optimization. Others overcorrect, wiring up fully automated frameworks before anyone knows what makes a headline convert. Ankur has seen this pattern play out across big-budget financial institutions, and his advice is clear: timing matters more than tooling. Start simple. Learn from the mess. Then automate.

He recommends adding manual A/B testing right after you implement behavioral triggers, around Step 2 in the personalization maturity path. This is when you’ve got some engagement to work with, but things are still lightweight enough to learn fast. His team tested subject lines and CTAs in mortgage refinance emails. One variation crushed the other by 31 percent. That wasn’t an accident. It was a data-backed decision that built testing discipline into the culture.

That muscle matters later. By the time you get to Step 4, with predictive models in place and dynamic segments flowing, your tests should evolve too. This is where automation earns its keep. You’ll have enough traffic to hit statistical significance quickly. Your content modules will be clean. KPIs will be sharper. The infrastructure can handle complexity without buckling. At this stage, Ankur’s team used machine learning to reallocate traffic in real time based on performance. That lifted conversions another 37 percent.

“Many teams jump to automation without ever building the muscle to test with intent. You get results, but no one knows how to use them.”

None of that works unless you’ve already trained your team to ask the right questions. What are we testing? Why? What does winning look like? Automating too early skips that entire layer of learning. Then you’re just measuring noise.

Key takeaway: Introduce manual A/B testing early, around Step 2, to build the discipline of hypothesis-driven experiments. Once you reach predictive personalization and have clean segments, measurable KPIs, and modular content, layer in automated frameworks that can optimize at scale. That way you can grow from gut-check testing to systems that drive real lift, without skipping the hard part where you actually learn.

Personalization Breaks Without a Clean Data Core

AI-powered personalization sounds great until your data betrays you. Ankur learned this the hard way while leading a rollout for a major wealth management firm. The mandate was clear: launch personalized investment journeys as fast as possible. The data was… not ready. But with executives demanding quick wins, the team pushed ahead anyway. That’s how clients in their thirties ended up getting retirement offers, and parents with grown kids received college savings nudges. Every message was “personalized” but completely off.

What was needed was a total reboot. Ankur built a Data Readiness Roadmap from scratch and refused to launch anything new until the foundation was stable. That meant:

• Hitting 90 percent identity resolution across all channels
• Cleaning transactional data to reduce latency to under 12 hours
• Standardizing behavioral attributes across both retail and wealth divisions
• Building usable customer profiles with consistent histories

The team spent six months on this before touching segmentation. That patience paid off. When they relaunched, engagement spiked by 43 percent. New investment account openings rose 28 percent. Nothing changed in the tech stack. The lift came entirely from cleaning up the data layer.

“The quality of your data foundation directly determines the ceiling of your personalization capabilities.”

This is the part most teams skip. They race to deploy machine learning before they even trust their own customer records. No model can fix broken identity graphs or misaligned attributes. Personalization that works always starts with a dataset you believe in.

Key takeaway: Treat your data layer like product infrastructure. Before launching segmentation or predictive models, fix identity resolution, standardize attributes, and clean your transaction history. Use latency and completeness as benchmarks. Build behavioral profiles that hold up across systems. That way you can ensure every personalized message actually reflects reality, not a stitched-together guess.

Fix the Stack You Have Before You Buy Another Tool

The worst-kept secret in enterprise marketing is the mess under the hood. Everyone is chasing AI personalization, real-time journeys, and fancy next-best-action flows. Meanwhile, their tech stack looks like a Jenga tower made of duplicate CRMs, legacy banking platforms, and APIs that haven’t worked since the last merger. Ankur walked into this exact scenario with a regional bank while working at Salesforce, and instead of burning it down, he found a smarter way through it.

The CMO wanted personalization across every channel. The problem was that customer data was locked in silos, scattered across wealth systems, CRM, and digital platforms. Each tool handled a different slice of the customer, and none of them were synced. The entire stack had become a cautionary diagram—what Ankur calls the “spaghetti architecture.” Ripping it out would take years and millions. There was no time, and even less appetite, for that.

“The breakthrough came when we stopped trying to replace everything and focused on orchestrating what already existed.”

So the team built what Ankur calls a “Connective Tissue Strategy.” It wasn’t about starting over. It was about layering above. First, they implemented a customer data platform (CDP) to unify fragmented profiles without interrupting day-to-day operations. Then came three critical layers:

• API connections across the banking core, CRM, and digital systems to build a single customer view
• A decisioning engine that could read from that view and make smart, context-aware choices
• Delivery integrations that pushed personalized messages across all relevant channels in sync

Things really clicked when they added the wealth system into the mix. Client data that had been off-grid was now fueling personalized cross-sell campaigns. That single move increased investment product sales by 34 percent. All without a single platform replacement.

Key takeaway: Personalization at scale doesn’t require a replatform. It requires orchestration. Build a connective tissue layer that lets your existing systems work together. Start with a CDP to unify data. Add decision logic to surface the right actions. Then plug delivery into the channels your customers actually use. That way you can unlock personalization fast, without blowing up your tech stack or your budget.

Everyone Wants to Own the Orchestration Layer

How to Stay Sane When Every Martech Platform Thinks It’s the AI Brain

Every martech vendor is now claiming to be the mastermind of your AI customer experience. CDPs, iPaaS, campaign tools, reverse ETL, journey builders; every one of them is shoving “AI orchestration” into their decks like it’s some kind of universal passport to relevance. The homepage copy has changed. But the architecture hasn’t.

Ankur calls this mess the spaghetti diagram. It’s the tangled web of overlapping systems that all claim to be essential, none of which speak the same data language. He sat in a digital transformation steering committee where the CTO finally asked the one honest question that cut through it all: “In five years, which of these platforms still exists as a separate category?”

“The winners won’t be the ones that chose the right vendor. They’ll be the ones who built for collapse.”

Instead of picking favorites, Ankur’s team built what they now call a Convergence Strategy. They stopped thinking in categories and started thinking in fluid layers. The structure prioritized three things:

1. Open APIs everywhere
2. Short contracts that keep switching costs low
3. A data orchestration layer that sits above the chaos

This gave them room to move when categories started to blur. It paid off when their CDP and automation vendor merged mid-project. No one panicked. The team didn’t need to rethink their workflows or rewire their integrations. Their architecture already expected this kind of disruption.

The real shift is from tools to use cases. Teams that still debate whether the CDP or the ESP owns the customer profile are missing the point. AI agents don’t care where your data lives. They care about whether it’s structured, accessible, and connected to decisions.

Key takeaway: Bet on convergence, not categories. Build your stack around modular components with short-term flexibility and open connections. Let the platforms fight for center stage while you build a data layer that can pivot no matter who merges next. The future belongs to those who prepare for the collapse.

Coordinating AI Agents Without Creating Total Chaos

Enterprises keep layering AI agents like toppings on a bad pizza. One for customer service. One for fraud. One for personalization. One to watch the others. Each vendor sells their own logic as the “brain” of the operation. But brains don’t work when they aren’t connected. What you get instead is a mess of overlapping triggers, siloed decisions, and conflicting messages that leave customers confused and teams pointing fingers.

Ankur ran headfirst into this problem while overseeing enterprise architecture for a large financial services firm. Their stack included agents for marketing orchestration, wealth advisory, customer service, and risk management. All deployed with good intentions. All working in silos. It got bad when the marketing agent sent a high-risk investment offer to a client flagged by the advisory agent as needing conservative strategy. Same customer, two systems, totally disconnected logic.

“The moment AI agents start stepping on each other’s toes, you don’t have automation. You have accidental anarchy.”

To fix it, Ankur’s team built an Agent Governance Framework. Three critical components held it together:

1. A centralized knowledge graph that fed all agents a shared, real-time understanding of the customer
2. A decision hierarchy to resolve conflicting recommendations
3. A unified customer context layer that kept every system aligned to the same data

The real breakthrough came with what he calls a “meta-agent” architecture. This was an orchestrator for the orchestrators, a control plane that coordinated how other agents fired, resolved disputes, and timed their actions. That one move alone reduced conflicting actions by 87 percent and brought a fractured tech stack back into alignment.

The real problem isn’t too many AI agents. It’s too few guardrails. Enterprises that want orchestration without chaos need to treat agent collaboration as a foundational architecture problem. That means designing for control, not chasing features. If you rely on vendor promises to sort this out for you, don’t act surprised when your customer gets five emails, three nudges, and a call from your chatbot… all within ten minutes.

Key takeaway: Treat AI agent orchestration as a system design problem. Build a governance layer that defines how agents share knowledge, resolve conflicts, and operate from a unified customer view. Add a meta-agent to coordinate actions across your stack. That way you can scale automation without creating a fragmented experience that erodes trust.

The Only Team Structure That Actually Scales Personalization

Hiring more specialists will not fix your personalization program. Spinning up a shiny Center of Excellence won’t fix it either. Ankur has tried every organizational model under pressure from executive teams who wanted faster results. What worked was a structure that brought data and decision-making together at the edge of the business. No ivory tower. No disconnected strategy decks. Just measurable outcomes owned by people who are close enough to care.

At first, Ankur set up a centralized personalization team to support every business unit. This was supposed to streamline operations. In reality, it created decision bottlenecks, long queues, and one-size-fits-nobody journeys. The team was excellent on paper and useless in context. Product marketers couldn’t wait three sprints to test a CTA variation. Acquisition leads got tired of fighting for priority against the loyalty team. Speed tanked. Relevance dropped. Nobody won.

The breakthrough came with what he calls the hub-and-spoke model. The hub was a centralized data and tech team that maintained the CDP, managed identity resolution, and owned model infrastructure. Around this hub, Ankur built cross-functional pods, each focused on a specific lifecycle stage like onboarding or retention. Every pod included a product owner, a data scientist, a content strategist, and a technologist. They worked as a unit, with full autonomy and a single goal.

“We plugged the holes in our pipeline but we also gave each team ownership over the result. That’s when the numbers moved.”

That structure created velocity and accountability. The acquisition pod lifted new customer conversion by 28 percent. The retention pod cut churn by 23 percent. None of this required a lame reorg memo. It required putting talent where decisions happen and giving them a stake in the outcome. Centralize the data. Decentralize the execution. Align around moments that matter.

Key takeaway: Skip the fantasy charts and build a hub-and-spoke model that works in real life. Keep your data and infrastructure centralized to maintain speed, consistency, and integrity. Surround it with agile pods aligned to key lifecycle stages. Fill those pods with cross-functional talent who own their metrics and can act fast. That way you can drive real results with personalization that actually fits the customer, not the org chart.

Stop Pretending Your Data Is Ready for AI

The phrase “data readiness” gets tossed around like it’s just a checkbox. Got some events firing? Great. CRM is technically connected? Fantastic. Let’s launch the AI. But ask anyone who’s actually tried to operationalize personalization and they’ll tell you the same thing. Most teams aren’t blocked by lack of data. They’re blocked by the wrong data in the wrong shape with zero context.

Ankur has seen this up close. Teams rush into AI thinking quantity and accuracy are enough. Then they wonder why the “personalized” campaign ends up sending retirement planning offers to clients in their thirties or serves ads for college funds to people whose kids already graduated. The root problem isn’t how much data you have. It’s whether the data actually fits the job you’re trying to do.

There are three hidden failure points that break AI personalization before it starts:

1. Inconsistent timelines: Most customer behavior data is a patchwork of disconnected events. You can’t predict a journey if you can’t track one.
2. Context-free signals: Knowing what someone clicked tells you nothing without knowing why. Personalization without intent is just guesswork in fancier clothes.
3. Relationship blindness: Especially in industries like financial services, individual records miss household dynamics. AI thinks it’s helping a solo buyer when it’s actually disrupting a family decision.

“We saw a 37 percent lift in personalization performance just by enriching household relationship data. No model tweaks. Just better inputs.”

Ankur’s team built a “fitness-for-purpose” framework to pressure-test whether the data could support real personalization. Their benchmark was blunt but effective: nine months of consistent behavior, 85 percent identity resolution across channels, and data latency under 48 hours. That’s the bar. You don’t need pristine records across every field. You need strong enough data in the right dimensions to make the specific use case work.

Key takeaway: Data readiness is not a volume game. It’s a relevance test. Build your standard around the use case, not the database. You don’t need perfect data. You need the right blend of behavioral history, identity clarity, and contextual signals. That way you can stop wasting time polishing the wrong fields and start generating personalization that actually connects.

Balancing AI Personalization With Human Context

AI gets a lot of credit for driving personalization, and some of it is deserved. Timing gets sharper. Messages get more relevant. Portals see more logins and dashboards light up. That’s all great until a client reads their 12th robotic update and starts wondering if anyone behind the scenes actually knows their name. Ankur has seen this tension play out in asset management, where precision matters, but so does the sense that someone human is still steering the ship.

His team rolled out an AI engine for high-net-worth clients. It handled everything from portfolio performance alerts to behavioral pattern analysis. It worked. Open rates rose 42 percent. Clients logged into their accounts 37 percent more often. The system hit every quantitative benchmark. But when the quarterly survey results came in, the story changed. Relationship strength scores dropped by 18 percent. Feedback included phrases like “feels mechanical” and “missing the personal touch.”

Digging deeper, Ankur’s team discovered a clear pattern. Messages that paired AI-generated insights with personal input from a human advisor performed far better than those that relied purely on automation.

“We saw 3.5x engagement when we added a short video from the advisor to market alerts. The AI nailed the timing, but the advisor made it real.”

That realization led to the development of a Human-in-the-Loop Framework. The team put strict rules in place. Life events like retirement or inheritance triggered advisor-led outreach. Investment recommendations over certain thresholds required human review. Communications during market volatility had to include commentary from the client’s advisor. With these rules in place, AI handled scale and relevance. Humans added warmth and credibility. Advisors ended up spending 40 percent less time on low-value messages and more time building real relationships.

Key takeaway: Personalization at scale works best when AI drives timing and targeting, and humans step in for emotional context. Use clear rules to define when the human touch is required. That way you can keep personalization fast and efficient without losing the trust that comes from real human connection.

Predictive Loyalty Engines are Coming to a Martech Stack Near You

Personalization is heading into stranger territory. Not just more granular or more automated, but more psychic. Ankur calls it behavioral intent prediction. It is personalization that anticipates what customers will want before they even ask. Sometimes before they even know. It is not science fiction. It is already in play.

His team built early prototypes inside a loyalty program and watched the pattern engine do its thing. The model picked up on subtle behavioral tells—session frequency, gaps in buying cycles, preference drift. Then it responded before a search was ever typed. Customers who hadn’t even realized they were about to bounce got relevant, timely nudges. Retention jumped 53 percent.

Then came a second wave: empathy as a design input. Not just preference scores or clickstream data, but signals of emotion. In one pilot, Ankur worked with a healthcare provider to identify anxious patients before they voiced concerns. The system parsed tone from phone calls, portal habits, and delay patterns. If it flagged a concern, it adapted. Emails slowed down. Instructions got simpler. Sometimes a care coordinator reached out before the patient even knew what to ask.

“The system recognized stress before the patient did. It changed how we communicated without the person needing to ask for help.”

This shift from reactive to preemptive, from mechanical to emotionally responsive, makes every stale CRM strategy look fossilized. It is the difference between checking boxes and genuinely supporting someone in a moment that matters. Predictive loyalty and empathetic AI are not gimmicks. They are strategic weapons. And the teams who operationalize both will build brands customers stay loyal to for reasons they may never even articulate.

Key takeaway: Build systems that act before users do. Predictive loyalty engines detect subtle intent shifts and proactively deliver value that keeps customers engaged. Combine that with empathetic AI that reads emotional context, and you can tailor communication cadence, tone, and support before problems surface. That way you can drive retention, trust, and long-term value without relying on reactive personalization tactics that are already too late.

Protecting Your Sanity While Scaling Martech

Ankur runs multimillion-dollar personalization engines, but when dinner starts, the only metric he cares about is the family engagement rate. He treats his home life like the most valuable A/B test he’ll ever run. Campaign alerts might spike dopamine, but they do nothing for his daughters. If his smartwatch buzzes during dinner, the test is simple. Does this notification improve his relationship with his kids? If the answer is no, it gets ignored.

“I treat my personal life like a high-priority A/B test that always outperforms my work metrics.”

That mindset doesn’t just show up at the dinner table. On weekends, Ankur goes analog. He puts away the phone. No Slack. No inbox. No dashboards. He grabs a paddle and heads to the pickleball court with his family. This is where he decompresses, not optimizes. He’s still a systems thinker, but the system is different. There’s no campaign to fix. The win condition is showing up.

He’s not chasing mythical balance. He’s setting up boundaries that force clarity. And that’s what most professionals in this space get wrong. They wait for the chaos to calm down before they log off. It never does. There’s always another alert. Another AI pilot. Another metric to explain. The work is infinite. Your attention is not.

Ankur makes room for focus by treating his off-time like it’s worth protecting. This isn’t aspirational wellness-speak. It’s a tactical choice. You either design your day with intention, or someone else fills it with pings.

Key takeaway: Respect your off-switch like it’s a business asset. Ask if every notification is making your life better or just louder. The more you treat your personal time as sacred, the more clarity you bring to the work that actually matters.

Episode Recap

Ankur Kothari has spent enough time in the guts of marketing systems to know where personalization actually breaks down. He joined a mid-sized home goods brand while they were hemorrhaging money; $2.3 million a year lost to customers who bought once, then disappeared. Their churn rate was brutal, 68 percent. The products were fine. The prices were fair. What drove people away was the emptiness of the experience. Every message felt like it came from a robot with a clipboard.

Ankur didn’t throw tools at the problem. He started by looking at what customers were already doing (clicking, buying, emailing support) and pulled it into a model that could predict what might actually matter next. Then he worked with the team to build 27 distinct customer journeys. Real ones. With content that changed depending on what people had done, not what marketers assumed they wanted. Within months, retention jumped by 42 percent. One customer went from a single $45 order to over $600 in purchases the next year. Not because of a clever subject line, but because the brand finally made sense to her.

But he’s seen the opposite too. A retail bank spent millions on AI loan models while their welcome emails were arriving late and pitching products customers already had. That kind of disconnect kills trust fast. Ankur paused the AI rollout, cleaned up the basic flows, fixed broken logic, and restructured the automation. Only then did the AI layer start doing what it was supposed to. Product adoption went up, and revenue followed. The algorithm hadn’t changed. The environment had.

There’s a rhythm to getting personalization right, and it doesn’t start with AI orchestration or fancy dashboards. Ankur laid out a five-stage maturity framework that starts with small, clear actions (like sending the right email after a customer sets up an account) and builds slowly from there. When you can predict life events based on transaction patterns and send timely, useful messages, engagement climbs. When you layer AI on top of that solid foundation, the system gets smarter without losing its grip on reality.

Behind all of this is a simple idea: personalization works best when people feel understood. That means data has to be trustworthy. Teams have to own their journeys. Systems have to talk to each other. In one case, Ankur helped a financial services firm stop its AI agents from tripping over each other by building a central knowledge layer and creating a meta-agent to resolve conflicts. The change wasn’t sexy. It just made everything make sense again.

Ankur also shared a quieter truth. Automation can be efficient, but it’s not always personal. In wealth management, clients loved the precision of AI, but it started to feel cold. Adding a 30-second advisor video to the updates brought engagement back up. People don’t want to be managed by machines. They want to feel like someone’s paying attention. And in the end, that’s what good personalization is; paying real attention to what matters and acting on it like you actually care.

Listen to the full episode ⬇️ or Back to the top ⬆️

Follow Ankur 👇

• LinkedIn

✌️

—
Intro music by Wowa via Unminus
Cover art created with Midjourney (check out how)

Apple •Spotify• Pocket Casts •Youtube •Overcast •RSS

Related tags

See all episodes

Future-proofing the humans behind the tech

Apple •Pocket Casts•Google •Overcast •Spotify •Breaker •Castro •RSS

• LinkedIn
• Twitter
• Instagram
• YouTube
• RSS Feed