This post is part of the AI for Business series
Shedding light on AI from a non-technical business/product/design perspective.
This is post #4 in our AI for Business series.
As we discussed in previous blogs in this series, to harness the power of AI you first need to clarify the problem you want to solve and your existing business processes. Only then should you consider where to introduce AI. In the previous blog we looked at introducing AI powered automation. Now we will look at another application for AI; cognitive insight (CI). This is a big one!
What is CI?
Imagine you’re running a business, and you’ve got more data than you know what to do with. Sales numbers, customer interactions, market trends; an overwhelming ocean of information. You know there are insights buried in there, but finding them? That’s the challenge and it is where cognitive insight can help.
Cognitive insight refers to the ability of AI systems to analyse vast amounts of data, recognise patterns, and generate insights. It can spot patterns that would be invisible to the human eye, connect dots that seem unrelated, and even predict what might happen next. It’s not magic, but it feels pretty close. Cognitive insight uses advanced techniques such as machine learning, deep learning, and natural language processing (NLP) to discover these hidden relationships and to provide predictive analytics.
In short, cognitive insight is useful because it:
- Recognises hidden patterns
- Predicts future outcomes based on historical data.
- Recommends actions; not just showing data, but suggesting what to do with it.
- Gets smarter over time, learning from new information to refine its insights.
The advantages of this technology for strategic decision making are huge. Let’s take a look at some examples.
Cognitive Insight helps thoughtburgers deliver delicious food 🍔
Our delectable side-business, thoughtburgers (home to the tastiest and most thought-provoking burgers!), has been a big hit and multiple branches have been opened in London. 🥳 We want to deliver our delicious burgers to hungry customers but, during rush hour, deliveries can take up to 50 minutes due to limited kitchen capacity and heavy traffic. 😢 That is too long to wait for a thought-provoking burger!
Our team wants to reduce this rush hour wait by 30% (our measurement of success) so we look at our end to end process:
And then we identify where Cognitive Insights might help:
Steps 2 & 4:
Using data like order volume and location, we could train a model to predict demand and optimise driver schedules better than a human.
The model could also be trained to cluster and identify patterns in driver illness data which are hard to identify with the naked eye.
A more efficient diver schedule could reduce delivery time during peak hours.
Step 6:
Using data about order volumes, locations, kitchen capacity, driver locations etc., the triaging of orders to specific restaurant branches could be optimised using CI.
While the default behaviour might be to assign a delivery to the nearest branch geographically, CI might spot patterns that allow us to consider alternative approaches.
For example, shipping a burger from a branch that is slightly further away from the customer geographically, but that has the kitchen capacity, necessary stock and drivers available, could result in the burger reaching the hands of our customers more quickly than if it was shipped from the nearest geographical branch. Being able to predict demand geographically could help our branches become even more prepared, further reducing wait times.
The outcome?
These data-driven decisions could improve operating efficiency of all branches, reduce order rejections and, ultimately, contribute to a reduction in delivery wait times during rush hour. 🤤
Other use cases
Some other use cases for Cognitive Insight that might apply to your business include:
Sentiment analysis
Cognitive insight takes sentiment analysis to the next level by not just identifying whether feedback is positive or negative, but also understanding context, nuance, and intent. Traditional sentiment analysis might flag a review like “The wait time was ridiculously long, but the service was fantastic!” as negative due to words like “ridiculously long.” Cognitive AI, however, recognises the mixed sentiment and weighs it accordingly. It can also detect sarcasm, emerging trends, and shifts in customer opinion over time giving businesses deeper insights into how people truly feel.
Combatting churn
Cognitive insight helps prevent customer churn by identifying at-risk customers before they leave. By analysing purchase history, engagement patterns and support interactions, AI can detect early warning signs like declining usage or negative feedback. More importantly, it doesn’t just flag the risk; it predicts why a customer might leave and suggests personalised actions to re-engage them like a custom special offer or support outreach. This allows businesses to be proactive rather than reactive.
Predictive maintenance
Cognitive insight can transform predictive maintenance by detecting subtle patterns in sensor data, equipment logs, and historical failures that humans might overlook. Instead of relying on fixed schedules, AI continuously analyses real-time performance metrics to predict when a machine is likely to fail before it does. This means fewer unexpected breakdowns, lower maintenance costs, and improved operational efficiency. Whether it’s a factory conveyor belt or a fleet of delivery trucks, cognitive AI helps businesses move from reactive fixes to proactive, data-driven maintenance strategies.
Spotting trends
Cognitive insight helps fashion companies spot trends across different demographics and regions before they go mainstream by analysing data from social media, online searches, influencer activity, and even real-time sales patterns. Not only does this reduce overstocking of items that won’t sell, it also means fashion designers can actually create completely new designs and products to cater to future demands rather than the demands of today or yesterday.
Supply chain management
Companies like Amazon rely on CI to analyse vast amounts of data, from weather patterns and shipping delays to customer demand and supplier performance, to predict disruptions before they happen. Cognitive insights can help optimise inventory levels in real time, ensuring products are stocked where and when they’re needed while minimising waste. This can even result in anticipatory shipping, where products are on their way to a warehouse near you before you’ve even bought them! 📦
Challenges
While cognitive insight can provide businesses with a real competitive advantage, it is important to note that there are challenges and limitations with it. Total reliance on CI systems would be ill advised.
Unexpected events
If we revisit our thoughtburgers example, restaurants may still need to manually override the system and to reject orders if a customer or member of staff had a medical emergency, there was a fire in the kitchen, the water supply was cut off etc. Unforeseen circumstances that the model won’t have accounted for must be considered.
Transparency & accountability
AI systems often operate like a “black box,” offering limited insight into how they work and how they arrived at certain decisions. Transparency is vital to ascertain how decisions are made. It is also essential to understanding who bears responsibility when AI systems make errors or cause harm, ensuring appropriate corrective actions can be taken.
For example, if thoughtburgers delivery triaging system forced a particular branch to close because an error resulted in the algorithm not sending orders there anymore, who is culpable for this?
Bias & data quality
AI systems are trained on massive amounts of data, and embedded in that data are societal biases. These biases can become ingrained in AI algorithms, perpetuating unfair or discriminatory outcomes.
For example, one of thoughtburgers’ delivery drivers may have religious commitments that make them unavailable to work on certain days. Penalising the number of shifts they are offered on religious grounds is a serious breach of ethics and is illegal in many countries, but an unsupervised algorithm may start doing so nonetheless.
Privacy & security
The effectiveness of AI often hinges on the availability of large volumes of personal data. As AI usage expands, concerns arise regarding how this information is collected, stored and utilised, especially in relation to GDPR regulations and compliance.
Big companies win big
Cognitive insight is most helpful to companies with access to large amounts of data. Usually, this means larger corporations. This can make the already competitive landscape even tougher for small and medium sized companies to stay afloat in their market.
Combatting challenges:
To combat these challenges, it is critical to regularly ensure that the models are operating efficiently and as expected. Algorithms are not a “one-and-done” like a delicious burger; they require constant maintenance, testing and supervision after they have been built to ensure harmful biases have not crept in and that logical decisions are still being made.
Incorporating “explainable AI” to help characterise a model’s fairness, accuracy, and potential bias is extremely beneficial and makes decisions more transparent. Sharing the model’s training data is another good move.
In case decisions lead to damaging results, accountability guidelines should also be considered.
Customers should be able to opt-out of having their data used to train the models. Robust safeguards against data breaches and unauthorised access to sensitive information is also crucial and may result in higher cybersecurity costs for businesses looking to utilise CI.
Support programmes for SMEs and open source projects could contribute to a more competitive landscape where small businesses have a better chance to survive against big giants with lots of valuable data.
Bottom line?
Cognitive insight is an incredibly powerful and beneficial application of AI but it should be used to aid strategic decision making, not to replace it.
💡 If you’re ready to start using AI to transform your business, thoughtbot would love to work with you. Let’s talk about making your AI initiative a success!
This blog post is part of a series based on the course Artificial Intelligence (AI) for Business run by University College Dublin (UCD). I took this course and found it so helpful that I’m recapping my top insights. thoughbot has no affiliation with UCD.
