Today’s data landscape is dominated by artificial intelligence (AI), and there are four factors shaping this landscape, two familiar, two relatively new. The “speed of change” and the “volume of data” have been on the radar for years, but today’s business leaders also need to consider “data dimensions” and “data expiration.”
Even if you go back only 15 years or so, data was finite, easy to understand and implementable into business functions and decision making. Data could be handled by humans—volumes were manageable, and the speed of change was very limited. We were mainly dealing with a single dimension of data, and the data had a long shelf life. A skilled and trained human could digest and crunch the data and make business decisions. In this case, your business was being run by pilots.
The first disruption was that the speed of change started to increase. Humans initially struggled with this, but a wave of new tools came into play to help with not only analytics but also communications, customer relations, project management, enterprise performance and more. Software-as-a-service emerged as an option, with businesses now overseen by humans and software, working together as co-pilots of the business.
There was a holding pattern for a few years before the volume of data coming into an enterprise started to increase. The enterprise software was generating inputs, and the consumer was starting to share social signals, geo-locations, web (and smartphone) activity and more. Humans and software couldn’t digest the amount of data that was now coming through, so automations started to be written into the software.
Or, to be accurate, the software got better and allowed rules to be written into its programming. These rules allowed enterprises to effectively digitize their knowledge, and by giving the software defined parameters in which to operate based on these rules, automation became a primary use case for data and fundamental to successful businesses.
Businesses were now relying on autopilots, making decisions based on the rules and driving the business forward more effectively than during the pilot and co-pilot stages.
But there’s been an altitude shift over the past few years, and today, my opinion is that businesses need to have an AI pilot at the controls because of the two new data concepts. Data dimensions have increased significantly, not only because there are more sources of data, but also, the expectation is that data can be analyzed multi-dimensionally and holistically, not just in a linear way using single variables as was the case before. The change in dimensions allows deeper insights from increasing volumes of data.
And the expiration of the data is now a lot shorter in terms of how long enterprises can legally hold data and how long a single data point remains relevant for the specific field it is being applied to. Prioritizing real-time decision making, the idea of the right product for the right customer at the right time, is behind this new data trend, where signals from weeks ago, if not hours ago, can be out-of-date.
Rethinking legacy data
Having said that, there remain many use cases where there is value in historical legacy data. This value can be realized by using AI to convert batch data—collected and stored in databases before being analyzed and acted upon—into streaming data—collected, analyzed and acted upon in near-real-time.
This means that, across the entire digital ecosystem, today’s tools are starting to understand and learn from the multi-dimensionality of the data being ingested, becoming aware of the timeliness of data. In turn, they can create their own business rules, on demand and aligned with the service being requested to carry out.
Enterprises need to understand that if your business is controlled by pilots or co-pilots, you’re in trouble. Businesses that have their data ready for the autopilots and pilot phase will grow their competitive advantage, but there are ways to catch up.
Making up lost ground
In my opinion, everything starts with the business leaders. Executives need to acknowledge where they are, and where they want to get to to understand the complexity of the transformation to AI pilots.
Any business whose day-to-day operations are not fully digitized needs to address this immediately. There are tools to ease this transition, and once operations have been digitized, the next phases fall easily into place—the door to the benefit of data is opened, allowing software to digitize the knowledge base held within the data through automation.
Legacy enterprises that have caught up with the chasing pack can assess their options and make sure their move into the autopilot and AI pilot stages is proactive and strategic. The gap between the leading pack appears to be widening, but again, there are ways to catch up, specifically around data structure and partnerships.
Structural engineers
There is no consensus in the data world on whether business needs a new data infrastructure. In my opinion, leaders need to focus on the right things in terms of crunching their existing legacy data. Creating new data streams and structures will take time and incur costs without speeding the transition.
AI is solving its own problem—it can understand how to convert unstructured data into formats it can understand, interpret and act on. As noted, insights from legacy data can be converted by converting it into streaming data.
There’s a good example from recent tech history that backs my belief that creating new data silos should not be on the agenda of businesses playing catch-up. During the first decade of the data economy, tens if not hundreds of billions were spent creating data warehouses to clean, structure and integrate data to support decision making.
Today, AI can bypass many established data engineering practices with a cost-effective and reliable alternative. Converting legacy data into streaming data so that it can send signals to autonomous agents—those operating outside traditional data warehousing and engineering structures—in a near-real-time way would be my priority as a leader.
Partnerships as standard
The travel industry has suffered from a lack of enterprise-scale, long-term data partnerships. In the current data landscape, my opinion is that all businesses need to start looking for new types of data partnerships to have a smooth and successful transformation into the AI pilot stage. These partnerships should focus on multi-dimensional data, as mentioned earlier.
As an example, destination marketing organizations (DMOs) and convention and visitor bureaus have traditionally had a single, linear aim—to create demand for a destination—and their data partnerships reflected this. But this limited data set can be enriched by adding new dimensions, such as expenditure and conversion, which must come from a different data partner. If the DMO is at the AI pilot stage, real-time connectivity of the data coming from different dimensions into an agentic system can automate decision-making.
If we accept that enterprises need new data partnerships to fully transition toward AI pilots, then there is a strong argument, which I subscribe to, that we also need to start the conversion around data standards. Consistency around how data is used, stored and analyzed in the future could help the travel industry and its tech ecosystem address issues around trust.
Trust is going to be pivotal moving forward. In a B2C context, there is a virtuous circle where guests are willing to share data with a trusted AI agent. The more data the agent ingests, the better the experience is for the consumer, in turn strengthening the bonds of trust, which then result in a greater willingness to share data.
Standards are also needed in B2B. Data partnerships will clearly have a commercial angle, but these need to be framed within a trusted and consistent framework that gives businesses the tools to deliver positive outcomes while maintaining the trust of consumers.
Next steps
Some businesses are already being reshaped through the adoption of autopilots and AI pilots. For the C-suite overseeing businesses still in the pilot and co-pilot stage, catching up is possible. AI itself can help you get your business in the right shape, provided you have a clear understanding of the specific use cases and incremental value that can be unlocked.
The question is simple: Will you still be flying with co-pilots while your competitors move into the signal-driven AI era?
About the author…