Aampe

See, how Aampe personalizes messages and experiences? Very demure, very mindful! Aampe doesn't think in terms of segments, instead, it focuses on people, chooses the best alternates that our users find cutesy, and delivers the messages at the right time – very mindful, very demure! 😀

Aampe is hosting a mini-series of live masterclasses on Recommender Systems starting next week! The first session will have DJ Rich and Schaun Wheeler talk about what really matters when implementing a Recommender System. Arpit Choudhury will orchestrate 🥁 If you search for information on Recommender Systems, everything has been SEO-optimized to talk about general principles – collaborative filtering, content-based recommendations, etc – and delves into specifics like matrix factorization. None of this helps a business get a Recommender System up and running. This session has been designed specifically for product and growth practitioners to learn what really matters when implementing a Recommender System. When? August 14, 11:30 am ET (3:30 pm GMT) Register now: https://lu.ma/7i9sgzdz

Accenture recently published a fascinating and powerful research report on "Cutting through the noise in consumer experience". They reported that "74% of consumers abandoned their shopping baskets in the last three months simply because they felt bombarded by content, overwhelmed by choice and frustrated by the amount of effort they need to put in to making decisions." At Aampe we know that THE fundamental problem for every consumer business today is effectively and empathetically managing their customers' attention. We live in an attention-scarce economy and the businesses that take care of their customer's 'attention account' will be the ones that will sustainably access their customer's financial account. With Aampe's Content and Message Catalogue capabilities, your CRM and product teams can finally engage every single user in their own context, with the content that reduces their stress, and improves their buying journey. Reach out today for a demo and to find out more!

Many savvy CRM & Lifecycle leads are flocking to Aampe to improve outcomes & reduce operational load by using reinforcement learning at the individual-level. “Wait, what?” Okay, we get it. Terms like reinforcement learning don’t always drive the point home. So, we created this simple :60 second explainer video to provide more context on what our product does and why it’s a game changer. P.S. If this video piques your interest, send a DM to any of the following Aampe team members so we can continue the conversation: Daria Sidorova, Ryan Doyle Paul Meinshausen 

I really enjoyed SuperAI in Singapore earlier this month. It's so exciting to see people starting to realise how dramatically our consumer applications can improve, and how much value agentic AI will generate for consumer businesses and their customers. 🔑 Here was one of the key messages from my talk: When you put generic content in your messaging... when you organise your content/product using shallow taxonomies... you are losing the opportunity to learn something specific about your customers and audience. 👗👠👕💃🧥👜🥾👖 If you operate a fashion retail application and a customer clicks "Women -> Dresses", what have you learned about their fashion style? Not much. 🧐 Instead, if your application prompts the customer with an option like "Funky, Oversized, Streetwear", when they select that option you've learned a lot. 👨🎓 🚗🚙🚘🚐🚖 If you operate a ride-hailing application and you present a user with a choice of "UberX" and "UberXL" then their choice doesn't tell you anything more than their selection - it doesn't give you any insight into their underlying preferences or priorities.🤷 Instead, if the application presents the choices: ➡️"Ride efficiently with UberX" ➡️"Stretch out with UberXL" Then you can make an inference about their actual 'reason' for their selection. And your power to make these inferences is even greater when the application experimentally presented a similar user with 2 different choices, like: ➡️"Save money with UberX" ➡️"Get space for your party with UberXL" 💡💡💡 Aampe is the agentic AI platform that you need to help your consumer product become so, so much better, and your business become the best in your industry. 

We've been working on making improvements to our recommender system. (Aampe lets you enable as many recommender systems as you like - our agents learn individual user preferences for each system - but if you don't already have your own recommender system, you can just use ours.) I won't go into all of the details, but we've implemented something similar to Amazon (see link to their paper in the comments below), which involves a random permutation of all of the user-item associations. So imagine a big spreadsheet where first column is a user id and the second column is an item id and every time a user interacts with a new item, you add a row to the sheet. We needed to randomly shuffle the item id column of that sheet. However, we're not just talking about a siple random shuffle. We need two things to be true at the same time: 1. The number of users associated with an given item should be roughly what it was before we randomly shuffled anything. 2. No user should ever be associated with the same item twice. All of this is easy to do in Python, because it's easy to work iteratively in Python. We can shuffle all the item ids, check to see which users have duplicated ids in their new set, then shuffle again to replace those duplicates, and so on until we have the result we want. The problem is that we get well over a million unique user-item associations in a single week, and we need to ingest much more than one week to really explore that data enough to make robust recommendations. The Python route is simpler to code but it comes at a relatively large I/O cost from moving all that data out of and then back into databases. So I spent some time this last week figuring out a database-only solution. A quick very just uses the LAG or LEAD operator, with an ORDER BY RAND() statement in the window specification. However, LAG and LEAD lack a wrap-around functionality, so one value will always be left null. If that's unacceptable, then you have to ROW NUMBER your whole dataset, then create one or more alternate row numbers with ORDER BY RAND() in the window spec, and then self-join the table on different row number columns. To solve the duplicates issue, we self-joined on five different random columns, aggregated all of those together into a single array per user, and then selected the number of item ids we needed to meet that user's quota. At any rate, we now have the random permutation we need for our recommender system (and, in fact, we're looking into ways to expand this method to other parts of our agentic learning). And now we don't have the burden of moving lots of data back and forth between two storage systems.

Have you ever thought about value-based personas instead of demographics? Understanding the psychology of values unlocks fascinating insights into why we connect so deeply with certain people and brands. When individuals share similar values, they often experience a natural, effortless bond, leading to meaningful personal and professional relationships. Companies can harness this powerful dynamic by weaving their core values into every aspect of their brand. This attracts customers who resonate with these values and fosters loyalty and trust. At Aampe, we do exactly that. During the onboarding process, we identify the company values and why customers are using your app. What engages them? Once we know that, we build messages that speak to every user. Join me on Thursday for our webinar about value-based personas and how we identify them!

After a really engaging first day at CommerceNext here in NYC, I'm reminded of something that Peggy Anne Salz asked me about during our conversation on The Groove podcast - which we recorded while at MAU Vegas in April: Does progress - in either Retail/Commerce or Martech - come down to toolset 🛠👩💻 or mindset 🧠👩🎓 ? It's both. But what's really interesting right now is that technology - in the form of and through AI - is changing so fast, that it's revealing fascinating things about commerce and product/marketing leaders' mindsets. The leaders who are most effectively riding this wave 🏄♀️ of technological change have an incredible combination of Humility and Ambition: Humility in the sense of recognising that so many things that historically they could have said "they know" about what technology can do and about the best workflows and processes to follow - are in fact just not really the case anymore. But Ambition in the sense that they recognise that the work they've been doing can be done so much better; that they can generate really exciting improvements in the experiences they provide to their customers in order to drive ambitious business growth. Humility and Ambition. The leaders who demonstrate those 2 qualities are the leaders of tomorrow. And we're so excited to work with them at Aampe! 

Wanna talk apps? Join Aampe next week, June 10th, at Aldo Sohm Wine Bar in NYC as we host a laid-back meet-and-greet with the New York tech/AI scene. The apps and drinks may be on the house, but the conversations, connections, and relationships will be rich. 🍷 Also, if you’re in eCommerce, be sure to attend CommerceNext on June 11th-13th. Aampe will be there to discuss how Agentic AI is overhauling retailers' approach to customer engagement. (Reach out if you're interested! We still have a few VIP passes available. 🎫) We hope to see you there!

Mr D serves over 396 new user segments with Aampe! ✅ Indulgent sushi lovers who want to try some new varieties? — Check! ✅ Deal-focused health enthusiasts looking for a convenient way to get their fix? — Check! ✅ Practical grocery-getters who need food on the table in time for dinner? — Check! Mr D is finding and engaging with these groups and more using Aampe’s agentic CDI. The Mr D team is investing in creating unique and captivating content for all of their users—from health-conscious eaters and value-hunters to convenience-lovers and everything in between—harnessing Aampe’s agentic AI infrastructure to serve content according to each individual user’s timing, frequency, and content preferences. “Meeting the needs and the expectations of our customers is our North Star,” says Su-lise Tessendorf-Louw, CMO of Mr D, “As a technology business ourselves, it makes sense for us to employ tools that serve our users in the way that they want to be served.” Mr D plans on building on their existing momentum, using Aampe to: 🍔 Design and execute multi-channel campaigns 🍔 Run user-level discount optimization 🍔 Build unique product experiences for each user based on user profiles constructed with Aampe More exciting news to come!