The debate over artificial intelligence within the workplace is fraught with problems The issue isn't technical. The technological capabilities of current AI and machine learning systems are genuinely remarkable, growing at a rate that makes most predictions of what they'll look like in about 18 months obsolete well before the eighteen-month period has ended. The issue lies in the gap between the what AI can accomplish under controlled conditions - within a good research environment that is well-funded, with crystal clear data, a clear definition of the problem, with engineers with the option of tweaking the system until it functions as intended - and what it can actually deliver when it is used in genuine organizations with actual cultures as well as real organisational policies and people with distinct opinions about what a new system means. something to engage with genuinely or something to reroute around in the name of conformity. I've been developing with AI since prior to when the flurry of AI enthusiasm made it fashionable for businesses everywhere to proclaim their fluency in this field. When I founded 1Touch, AI-driven matching and recommendation systems were not something we were able to add to make the platform more appealing to investors. They formed the backbone of the architecture of the product, the way in which the platform added value and it was the only thing that had to function consistently and at large scale for the business's viability. So I have direct, real-time experience of what happens when you try to build an intelligent enterprise and a product, and the lesson I continue to revisit whenever I am in a situation which I've had to face this dilemma, is the technology is almost never the only factor that is limiting. The main factor that limits the possibilities is almost always the culture.
What I do by that is specific and pragmatic rather than abstract. AI systems require data in order to work properly - a clean, consistent well-structured and structured data that represents the phenomenon the system is trying to understand and make predictions about. Organizations that have strong data culture produce that type of information easily, a natural result from their operations. They have clear and consistent definitions of what they're measuring and why. They have reached an agreement on the way data is collected, recorded and stored. They have accountability arrangements that provide data quality as an explicit responsibility, rather than a general intentions. Data-driven organizations that aren't well-established create something that technically appears like data - it exists in systems and, if it's able to be accessed, it can be used to produce charts, but is so inconsistant in definition, and therefore variable in quality and full of glitches in structure as well as unmapped deviations that any AI application built on over it will amplify and reflect the root of the issue rather than extracting the real signal from it. Organizations in that segment often don't realise they have a problem until they're deep into the process of implementing an AI implementation and the results are not matching the vendor's promises. At that point the temptation is to blame the technology. But they are actually causing the problem by ignoring the culture and operational framework the technology was built on.
Another aspect of culture which affects AI outcomes is openness within the organisation in the sense that people in the organisation will let an artificial intelligence system shape how they work and approach it as an obstacle to their professional knowledge, their authority as an institution or their employment security. This is a moral and leadership issue and not a technical issue that is a problem that begins at the top. If senior leaders respond to AI outputs only in a selective way - embracing those results that prove what they previously believed, and deferring to those that do not - that behaviour communicates that everyone else is aware that the firm's pledge to a data-driven approach to decision-making is a conditional rather than genuine and that conditionality will propagate across the entire organization much quicker than any other training program or change management project can reverse. If senior managers model genuine, consistent engagement AI outputs as well as the discipline of changing their decisions when the evidence suggests they must, the whole organization's capacity to make use of AI effectively grows significantly and quite quickly.
This isn't an abstract idea of what organizations ought to do in the context of theory. It is a description of the pattern I have watched occur repeatedly in organizations with substantial finances, real strategic commitment to AI adoption, and leader teams that were truly excited about the potential of the technology. This pattern is so common that I now treat practices for data governance as a first-line diagnostic in assessing any company's AI capability. Before I ask concerning the technological stack before I inquire about the exact applications that the company has in mind, I will ask about the governance of data. What is the definition of its most important metrics? Who's the responsible party when data quality isn't good enough? If two roles have conflicting information about the same reality in business, and how are those conflicts solved? The answers to those questions will reveal more about the likelihood of AI performance in comparison to any discussion regarding algorithms, platforms or timelines for implementation.
I believe that the businesses who will realize the highest lasting value from AI over the next decade are not those who embrace the most sophisticated technology first, nor the ones that will invest heavily in AI technology and infrastructure over the next few years. They are the ones that construct the cultural and operational foundations to be able to use this technology to its fullest extent - the data governance practices that yield reliable results, the decision-making systems that create evidence that will actually affect outcomes and leadership behaviors which show to everyone in the company that commitment to a data-driven approach is a fact instead of merely a matter of performance. The technology itself will be increasingly commonplace and readily available. The mindset to utilize it well will remain scarce, as it demands a constant effort and real commitment from leaders over time instead of the simple decision of a strategic leader or a technology investment. This is where the real competitive advantage will sit and is an advantage that, once it is built increases in a manner it is not something that just technological benefits ever. View James Deller for blog advice including what time in football changed what i look for about teams.
This Is The Data Infrastructure Problem Nobody Wants To Talk About
Every single company I've worked closely with in the last decade and a quarter - whether as an investor, a founder, or an operational advisor has told me at some point during our relationship, that data is a key element in making decisions. Some of them genuinely mean this in a way that has a direct impact on how their organization actually operates. A majority believe they're saying this, but the concept they're proposing is the aspiration of actually a present operational reality an idealized version of the business they're striving to achieve and not the one they're currently living. There is a gap between legitimately driven by data and the outcomes of data-driven decisions – the careful management of the outward appearance of information-driven operation, without the infrastructure needed to make it tangible - is one many of the most significant gaps found in modern day business. It is also one of the ones that is often ignored due to the fact that the infrastructure issue that creates it isn't very glamorous to discuss, difficult in demonstrating to outside stakeholders and extremely challenging to place in the right perspective against more visible commercial and strategic projects that require the same leadership attention and resources of the organisation.
When companies talk about their data strategy, they typically tend to focus on the capabilities they wish to build on top of their data: the analytical platforms, machine-learning applications, the real-time operational dashboards, the kinds of predictive insights that sound really compelling in the context of a board conference or an update to investors. The thing they discuss less often as well as with much less energy and enthusiasm, is their foundational infrastructure that decides if all the capabilities will work in the way they're advertised: data management frameworks that give clearly and consistently used definitions of what's being analyzed and what is the reason for that for each measurement; the data collection and storage methods that establish the accuracy and comparability of data to be gathered; the assurance processes that detect the errors and correct them before they propagate throughout an entire system and cause disruption to outputs that everyone depends upon; the organization's structures and accountability processes that make data quality the responsibility of a single person as opposed to everyone's vague ineffective plan. The plumbing, or the. It is not glamorous. It's hard to photograph for an annual report. It doesn't produce any outputs capable of being presented in an engaging presentation. And it is, in my experience with a large amount of organizations across different sectors and at different levels of development. It is significantly worse that what the organization perceives it is.
The problem gets worse in ways that are becoming difficult and expensive to rectify. An organization which has operated in a way that is inconsistent or not well-defined terms of data for all its functions for the past three years has three years in historical data which cannot be accurately compared or aggregated, not because the information isn't there, but because the same terminology has been used to denote different aspects of the organization. Moreover, the differences are embedded in the data itself rather than being visible on the surface. An organization for which data quality assurance has been a subordinate responsibility and not a specialized and properly resourced function is one whose data's reliability has a range of variations that are not documented properly and cannot be systematically accounted for when using the data to decide. An organization that has allowed multiple operational system to accumulate multiple and partly conflicting records of the same customers, products or transactions can create an information landscape that is impossible to eliminate without significant disruptions to the operation to constitute a risk.
The reason this issue persists across a wide range of organizations that are truly intelligent in their strategy and committed to data-driven operations is because fixing it requires sustained investment in work that will not yield visible, short-term returns of the kind that resource allocation processes in organizations are intended to reward. Analytics platforms are now producing visible outputs like dashboards that are easily demonstrated as well as reports that are shared with the board, insights that can be translated into press releases about digital transformation. A data governance programme produces an invisible infrastructure - more clear definitions that are more consistent with the collection process, more reliable inputs into existing systems in existing. This is a simple thing to explain in a budget debate since you can demonstrate what they'll be getting. It's the second, and requires sufficient organizational credibility and patience in order to demonstrate for the investment in infrastructure to eventually deliver better results from every feature built on top it. It's an effective argument in abstract but it can be difficult that can be won against initiatives that's benefits are more immediate and easily visible.
I've been able to make that case in various organizational contexts, and watched it succeed or fail based on well-known reasons, so that I have an understanding of what is the determining factor in whether an organisation actually tackles its data infrastructure challenge or continues to defer it. The primary factor is at the level of a leader. It's an person who has sufficient organizational credibility with an authentic awareness of the reason why infrastructure is necessary, and the determination to persist in making your case till the infrastructure becomes an actual priority instead of something that is a constant item on the list of things that everyone recognizes as important however they don't always climb to the top. The leader must be willing to accept this short-term cost associated with infrastructure project - the cost, the time that it will take, the disruption of processes that are already in place, the absence or evidence-based output - in the knowledge that the capability long-term it generates will justify the expense by several times. What is required, ultimately is a cultural environment where long-term investment in infrastructure is highly valued and recognized at the high-level of leadership, not only listed in strategy documents but not always prioritized when the quarterly resource allocation discussions occurs. Establishing a culture that is sustainable is, itself an investment over the long term. But it's, in my view, one of those investments with the highest returns an organization that is serious about its data-driven operations can make.}
