Instruction Manual for OfferFit models including: AdAge AI Decisioning Platform, AdAge, AI Decisioning Platform, Decisioning Platform, Platform
4 days ago — Manual segmentation. Predictive models group customers into segments based on their propensity scores. However, customers with similar scores might have vastly ...
File Info : application/pdf, 7 Pages, 1.10MB
DocumentDocumentThe future of AI is here and it's bringing fresh ways to create, automate and achieve one-toone personalization
How AI decisioning winitlol barninegwmeararketing
Sponsor content
Custom white paper
How AI decisioning will bring marketing into a new era
The future of AI is here and it's bringing fresh ways to create, automate and achieve one-to-one personalization
Personalization has become the cornerstone of effective modern marketing. Today's marketers face unprecedented pressure to deliver customized experiences that resonate with individual customers. Traditional methods like manual segmentation and A/B testing, once considered cutting-edge, are now inadequate for handling the complexities and scale of contemporary consumer demands. Marketers are turning to a new technology, AI decisioning, for one-to-one personalization.
"Predictive models have been better understood for a much longer period of time and are simpler, so it's easier to set up and operate them than AI decisioning," said George Khachatryan, CEO of OfferFit, an AI decisioning platform for lifecycle marketing. "AI decisioning is based on advances in a branch of artificial intelligence called reinforcement learning, which has been rapidly advancing in recent years."
Yum! Brands, the parent company of Taco Bell, KFC and Pizza Hut, has embraced AI decisioning to personalize retention and upselling campaigns. Yum Brands Chief Digital and Technology Officer Joe Park told The Wall Street Journal the company optimized its messaging for use cases like upselling, retention and more.
"Instead of waiting weeks for test results, we're seeing real-time results that we're continuously fine-tuning," Park said. "As we collect more data, we see AI playing a role in personalizing the menu board that you see or the kiosk that you're at, to know what you would more likely purchase at that moment, what kind of promos attract you."
The results have been transformational: double-digit increases in customer engagement
and reduced churn that highlight how AI decisioning can redefine marketing effectiveness.
Wit mhaattitseAr?I decisioning and why does
Personalization isn't an end unto itself--marketers personalize to drive business goals. "Most business growth happens because you create new value for customers, and you give them an easier way to get something done, be faster, discover new things and have fun," said David Edelman, senior lecturer at Harvard Business School and coauthor of "Personalized: Customer Strategy in the Age of AI." "They look at it from the customer's perspective and figure out how to add value for them."
70%
Consumers who say they value a personalized experience in which the brand knows who they are and their history with the company (past purchases, buying patterns, support calls and more)
Source: ACA Study "The State of Customer Service and CX"
COVER AND LEFT: GETTY IMAGES
2 The Business of Brands
Ad AgeDecember 2024
How AI decisioning will bring marketing into a new era
AI has introduced various tools into marketing, each designed to address specific challenges. Generative AI enables businesses to create text and visuals rapidly, while predictive modeling provides insights like churn propensity or purchase likelihood. While these technologies are valuable, they fall short of enabling true one-to-one personalization.
Beyond the hype of generative AI and large language models (LLMs), there is a growing trajectory of using AI to automate parts of the customer experience, which helps manage the scale and complexity large brands face today, said Juan Mendoza, CEO of The Martech Weekly.
"Imagine a world where you have AI doing the work of 1,000 A/B testing practitioners, continuously, 24/7 without any office politics, human bias and inefficiency," Mendoza said. "Just pure data-driven optimization. It might seem like a distant future, but for many brands, it's already a reality with AI decisioning technology."
SAunpeewrvfirsoendtivesr. reinforcement learning:
To understand the transformative power of AI decisioning, it's essential to grasp the difference between supervised learning and reinforcement learning. Predictive models built on supervised learning rely on labeled data to make predictions. For example, they might estimate that a customer has an 80% likelihood of churning in the next 90 days. While useful, this information alone doesn't guide marketers on what action to take.
"Predictive models make predictions," said Khachatryan. "They tell you, `This person will churn with a high probability.' But they don't tell you what action to take. Marketers often divide people into segments, run A/B tests for each segment and encode those findings as rules. This process is time-consuming, tedious and hard to maintain."
Brinks tripled profits on its contract renewal offers.
Case study
Brinks Home boosts retention with AI decisioning
The challenge: Brinks Home needed to improve the profitability of its contract renewal offers.
The solution: The company used AI decisioning to optimize its customer interactions.
The results: "By using OfferFit, we achieved a 200% improvement in the profit of our contract renewal offers in just weeks," said Veronica Moturi, senior VP of customer experience at Brinks. "If we were trying to get there through A/B testing, it would have probably taken us another 18 months."
"Predictive models make predictions. They tell you, `This person will churn with a high probability.' But they don't tell you what action to take."
George Khachatryan, CEO, OfferFit
BRINKS
3 The Business of Brands
Ad AgeDecember 2024
How AI decisioning will bring marketing into a new era
Marketers have benefited from predictive learning models, even with the difficulties involved, because it is helpful for identifying trends. However, it's not enough to have information about consumer behavior without an idea about the next step. Predictive models leave marketers with the burden of determining how to act. Turning predictions into decisions often involves manual processes like segmentation and A/B testing, which are time-intensive, prone to error and difficult to scale.
Reinforcement learning changes the game with AI agents that learn through experience. These agents continuously adapt, testing different approaches to see what works best. Instead of relying on segments or static rules,
reinforcement learning agents operate at the individual level, delivering truly personalized recommendations.
Khachatryan elaborated, "In traditional models, if two people have similar churn scores but for different reasons--say one moved cities and the other had a bad service experience-- they might be marketed to the same way. But with AI decisioning, the agent uses all available data to personalize the response for each individual."
For example, a leading credit card company found that predictive models couldn't account for critical nuances in customer behavior. By switching to AI decisioning, the company achieved a 92% uplift in referral conversions,
Why predictive models fall short
While predictive models have been a staple in marketing for years, they come with inherent limitations:
1. Manual segmentation Predictive models group customers into segments based on their propensity scores. However, customers with similar scores might have vastly different reasons for their behavior, leading to generic and ineffective messaging.
2. Static rules Predictive models require marketers to manually encode rules based on test results, making the system rigid and hard to adapt.
3. Slow iteration Traditional A/B testing can take weeks or months to yield results, delaying time-to-impact.
The benefits of AI decisioning
AI decisioning offers several key advantages over traditional approaches:
1. True personalization AI decisioning operates at the individual level, delivering customized actions based on each customer's unique characteristics.
2. Efficiency By automating segmentation, testing and rule-setting, AI decisioning significantly reduces the time and effort required to optimize marketing campaigns.
3. Continuous learning AI agents adapt in real time, refining their decisions as new data becomes available.
4. Scalability Unlike manual processes, AI decisioning can handle largescale operations seamlessly.
ILLUSTRATIONS: ADOBE STOCK
4 The Business of Brands
Ad AgeDecember 2024
How AI decisioning will bring marketing into a new era
adding $16 million in annual revenue. "For AI decisioning to work well, it requires
significant customization to align with each business' goals," Khachatryan said. "If you don't offer that level of flexibility, businesses can't use the product. It's not a nice-to-have--it's essential."
BAuI diledcyisoiuorntienagm to prepare for
Whether you're a CMO or a general marketing leader, you'll want to ensure you have experts in both martech and data science available to your team, said Edelman. "You need someone who is really thinking about your marketing technology as a product that's providing value," he said. "They are constantly looking for ways that you can take the product you already are paying for and get more out of it." The right person can help the rest of the team understand new ideas and capabilities, as well as new ways to automate so you can save money and reinvest that in other things that can drive growth.
Most enterprise brands are using data
science to help them gain insights and make strategic decisions.
By aligning data science with marketing, and having access to an expert who can help extract and utilize the right information, you'll be better able to connect with your consumers and move them to action.
On top of that, Khachatryan recommends having these four people on board part-time for the first few months while your team gets set up: "You need a project manager for the AI decisioning install, a data engineer, a marketer to do additional creative assets and a part-time IT colleague to do the integration with the marketing automation tool," he said.
The reason AI decisioning needs this focused support, Khachatryan explains, is that reinforcement learning is notoriously finicky and requires highly specialized expertise. "As an AI decisioning company, we bundle our product with expert services to ensure customers succeed. Our machine learning engineers configure and maintain the tool, continuously tuning it to deliver optimal performance."
"For AI decisioning to work well, it requires significant customization to align with each business' goals. It's not a nice-to-have--it's essential."
George Khachatryan, CEO, OfferFit
Case study
Quick wins with AI decisioning
The challenge: CleanChoice Energy wanted to optimize pricing recommendations for website visitors.
The results: A 30% increase in customer value and a 10% improvement in conversion efficiency.
The takeaway: "AI will optimize for the target metric you set, so make sure it aligns with your business goals," said Patrick Nyffeler, CleanChoice's VP of growth and revenue operations.
CleanChoice customer value soared 30%.
CLEANCHOICE ENERGY
5 The Business of Brands
Ad AgeDecember 2024
How AI decisioning will bring marketing into a new era
Checklist: How to get started with AI decisioning
Implementing AI decisioning may seem complex, but with the right approach,
businesses can achieve quick wins while building toward larger goals. Here's how:
Assess your data readiness
Ensure your customer data is unified and accessible. A data warehouse or customer data platform (CDP) is ideal for this purpose. ("It doesn't need to be perfect," Khachatryan said. "If you've made meaningful progress, you're ready to go.")
Start small
Choose a high-value, manageable use case to demonstrate early success. For example, a restaurant chain improved retention by targeting lapsed diners with personalized offers.
Build the big vision
Map out all potential applications of AI decisioning to excite stakeholders and secure resources. Focus on the long-term impact while achieving short-term wins.
Vet your vendor
Avoid "out-of-the-box" solutions. Work with a partner who offers both robust technology and expert services to navigate the complexities of AI decisioning.
Looking ahead: The future of AI decisioning
AI decisioning is not a fleeting trend--it's the future of marketing. Over the next five years, it will become table stakes for businesses aiming to deliver one-to-one personalization at scale.
"We're seeing very rapid adoption that's evident in the growth rate of OfferFit," Khachatryan said. "As marketers keep adopting this, we'll see dramatic reductions in bottlenecks like data integration, experimentation and creative asset generation."
Marketers who act now will be well positioned to lead in tomorrow's increasingly personalized landscape. The question is no longer whether AI decisioning is valuable, but whether your business is ready to embrace it.
6 The Business of Brands
Ad AgeDecember 2024
Ad Age Studio 30 Ad Age Studio 30 helps your brand connect with an influential audience actively seeking new partners, solutions and products. Through original custom articles, thoughtleadership content, events, research, webcasts, white papers, infographics and more, our end-to-end solutions help your content reach and resonate. Studio30@adage.com
Staff Writer Lisa Ferber Design Director Jennifer Chiu Senior Designer Natalie Skopelja Copy Editor Brian Moran
Contact us James Palma Senior VP, Sales and Client Partnerships jpalma@adage.com
John Dioso Editor, Studio 30 jdioso@adage.com
Deana M. Lykos Director, Activations deana.lykos@adage.com
Cassandra Hom Activations Specialist cassandra.hom@adage.com
About OfferFit OfferFit's AI Decisioning Platform autonomously experiments and empirically discovers the optimal actions 1:1 for each customer. OfferFit's ML experts configure AI Decisioning agents, a type of ML also called reinforcement learning, which personalize communication to existing, identified customers. OfferFit works with top brands in telecom, energy, retail, travel, streaming video and financial services, among others. Current customers include Fubo TV, MetLife, Wyndham and Yelp.
Adobe PDF Library 16.0 Adobe InDesign 16.4 (Macintosh)