In 2016, millions of people watched AlphaGo play Go against Lee Sedol. The question on our minds was the machine's next move. When ChatGPT became available in 2022, this time we were the ones typing the questions. For us, the practical change was that we could open a tool we had used for years and ask, "Which part of this do we actually need? Would we build it ourselves today?"
At eBrandValue, we use AI for two things: improving how we work, and growing through the product. On the engineering side, we replaced external dependencies and add-on services billed per user or per event with software we built in-house, designed around our own needs and, above all, fully under our control. We now rely much less on paid services and licenses. On the product side, we built a platform that keeps pace with social media and its users' habits, both of which change quickly. Today, when a brand appears in a video, is named in a conversation or is compared with a competitor by a consumer, we can detect it and use it in our analysis. For a product that forecasts sales, these distinctions matter a great deal.
The team that builds the product owns it
eBrandValue doesn't have a separate product manager. We want the engineering team that builds a feature to own it fully and to have a say in which customer decision it will improve. These days we spend less time discussing how to write the code and more time discussing how to solve the problem, and we hand a good part of the coding process to AI.
Say a brand asks us for a new report to evaluate its campaigns. We have a few options. The quickest is to build the report they asked for. But if the team first asks, "What decision will you make once you have this report?", the direction of the work can change. If the brand is trying to understand what to do differently in its next campaign, the real need may be to look at three things together: what consumers said during the campaign, the questions that reached the customer team, and the sales at the end of the period. At that point, the decision in front of us is bigger than which chart to add to the report. Are we meeting one customer's need, or investing in a feature other brands will use too?
As CTO, this is the judgment I expect from our engineers. If a shared feature will take time to build, we should be able to discuss why that investment is worth more than the next item in line. Sometimes the right call is to meet the customer's need with the existing product and leave the roadmap as it is. Sometimes a single customer's question reveals a capability the whole product is missing. For us, being a product owner means knowing the customer well enough to make that choice, and following the decision through to its outcome.
Social media never sleeps
Social media is active around the clock and unpredictable. A pipeline built for it has to cope with sudden spikes, and also with high volume that doesn't let up. A topic that comes up overnight should already be analyzed and waiting for brand managers and marketing teams in the morning.

A single piece of content a customer sees on the platform has passed through a multi-stage pipeline behind the scenes. When the number of posts around a topic rises quickly, we use Neo4j to examine the network of interactions between authors and understand how the topic is spreading. We use object detection models such as YOLO to identify products and services that appear in video frames. Whisper turns speech into text, and we then look at which products, services and brands come up in that text. An accurate transcript is only the start. The next step is understanding the context in which a brand is mentioned. With eBrandVision AI, which we built in-house, we can detect anomalies in the data and turn the results of different enrichment steps into insights a brand team can act on.
It's not just people anymore: LLMs talk about your brand too
When a consumer asks an AI assistant which product or service to buy, the assistant picks a few brands, ranks them and explains its choice. For a brand, visibility here has several separate dimensions. Is your name in the answer? Are you the first recommendation? Which of your features does the assistant highlight, and which sources is that based on? User-generated content on social media is no longer the only layer that matters. LLMs are a new one.
For a marketing team, the most useful finding is sometimes in the gap between two data sets. Consumers may talk about one of your products a lot on social media, while AI assistants regularly recommend a competitor when someone asks a shopping question that describes the same need. With AI Mindshare™, we can track that gap down to a specific category, consumer profile and question. That moves the conversation from a broad question like "How well does AI know us?" to a concrete one a team can work on: "For which needs does the assistant recommend our competitor, what reason does it give, and which sources does it cite?"

We cut costs by writing our own tools
When we reviewed the paid internal tools we depended on, and the hidden costs that came with them, we asked a similar question: which functions do we really use every day, and what would it cost us to build them ourselves? With AI, we've replaced some subscriptions with our own tools and cut a large share of what we pay per user, per quota and per event. On the other side of the decision are development time, infrastructure and maintenance. As CTO, what changed the math for me was time. Our team can now set aside time to build enough of its own internal tools for owning them to pay off.
Checking results against real sales data
Sales forecasting is at the core of what eBrandValue does. So when I look at a new analysis capability, the question I ask is: how does this information help us understand the consumer's purchase decision better? How often people see a product, how they compare it with competitors and what they criticize after buying it each tell us about a different stage of interest in a brand.
In eBrandValue, brands can track these behaviors, compare themselves with competitors and find what should guide their next campaign decisions. With Connect CXM, they can manage customer relationships at the point of contact and run new campaigns. At the end of the period, we come back to eBrandValue and compare our forecast with actual sales.
For me, this is where AI's contribution to the company shows most clearly. An engineer who understands why a bill grows with usage can now do something about it with a tool they built. The same team uses AI to ask better questions. It looks at how a new generation of customers' expectations of the product are changing, the way a product owner would, and builds the answer into the product. The time it saves on coding and testing goes into judgment and better decisions. That is the value of keeping product responsibility close to engineering. The team that chooses a technology sees what that choice leads to, and makes the next decision with that in mind.
To learn more about eBrandValue Platform, and how such metrics influence brand value, contact us using the form below.









