The discussion around artificial Intelligence for business has a glitch and the root of the issue is not technical. The capabilities of modern AI and machine learning systems are truly astonishing, and growing at a pace that makes the majority of forecasts of the place they'll be in eighteen months obsolete well before the time has passed. The issue lies in the gap between what AI can achieve under controlled conditions - such as a well-resourced research environment, with clear data, and a clear definition of the issue, and engineers who are capable of tweaking the system until it runs as planned - and what it will actually do when implemented in real organizations with real cultures as well as real organisational policies and real people who have distinct opinions about the quality of a system. something to engage with genuinely or a thing to take care of while still maintaining the appearance compliance. I've been building products using AI since prior to when the flurry of AI enthusiasm was a reason for business professionals to claim to be fluent in the field. When I co-founded 1Touch AI-driven matching and recommendation systems weren't a distinctive feature we added to make our product more compelling to investors. These were a fundamental part of the architecture of the product, the method by which the platform generated value and needed to be functional and reliable at an appropriate scale in order for the business to succeed. Thus, I've direct personal experience of what happens when you attempt to integrate an intelligent firm and a service simultaneously, and the lesson I keep coming back to regardless of the context in the past I've faced this problem, is that the technology will never be the factor that limits your success. What is the most important factor is all the time its culture.
What I consider to be specific and concrete, not abstract. AI systems require data in order to work - consistent, clean properly-structured data which describes the situation the system is attempting to analyze and make predictions about. Organizations with a strong and thriving data culture produce that type of information naturally, as a byproduct of their current operations. They have clearly defined and consistently implemented definitions of what they're doing and what they are measuring. They have a set of conventions that they agree to for the way data is collected, recorded and stored. They have accountability arrangements that allow data quality to be a distinct responsibility rather than everyone's vague purpose. Companies that lack strong data cultures produce something that technically looks like data - it's in systems which can be searched, it can be used to produce charts - but has a definition that is wildly inconsistent and in terms of quality and full of defects in structure, and non-mapped anomalies that any AI device built on the top of it will enhance and reflect the mess rather than extracting genuine signal from it. Companies in this category tend to not realize they exist until they're deep into an AI implementation and their outputs do not match the vendor's promises. At this point, it is tempting to blame the technology. But what is really at issue is the organizational and cultural foundation which the technology was built on.
The second dimension of culture which determines AI outcomes is openness within the organisation or the extent to which members of the organisation are willing to let the system influence or alter how they work, rather than treating it as an obstacle to their professional expertise, their authority in institutions and their job security. This is a moral and leadership issue but not one that can be solved by technology that is a problem that begins at the highest level. If senior leaders respond to AI outputs in a way that is selective - accepting the results that confirm their beliefs and refusing to accept those that do and do not, this behaviour sends the impression to everyone who watches that the organisation's stated commitment to data-driven decision-making is conditional instead of genuine, which will then spread throughout the organisation much faster than any training program or change management project can be able to counter. If senior leaders exhibit real and consistent engagement with AI outputs, and demonstrate the ability to make changes to their decision-making when evidence suggests that they should, the organisation's collective capacity to make use of AI effectively will improve dramatically as well as relatively rapidly.
This is not a speculative observation of the behavior of organizations in theory. It is a description of what I have seen take place in numerous companies which had significant financial resources, genuine strategic dedication to AI implementation, and leadership teams that were truly enthusiastic about the possibilities of AI technology. The pattern is similar enough that I consider data governance practices as a crucial diagnostic tool in assessing any organization's AI ability. Before I inquire whether the company's technology stack has been established, and before I ask what are the most relevant application scenarios the organization is working on, I ask about the governance of data. How does the organisation define its most important metrics? Who's in charge when quality of the data isn't high enough? What happens when two functionalities have conflicting information regarding the same facts about business, and how are those conflicts resolved? The answers to these questions give me more insight into the likelyhood of AI succeed as opposed to the endless debate about platforms, algorithms, or timeframes for implementation.
I believe that businesses who will realize the highest durable value from AI over the next decade are not the ones who embrace the most sophisticated technology first, nor those that invest the most significantly in AI infrastructure and talent in the near future. They are the ones who establish the organizational and cultural infrastructure to utilize that technology to its fullest extent - the data governance methods that produce solid inputs, the decision making structures that allow evidence to influence outcomes and leadership behavior that tell everyone within the company that the commitment to a data-driven approach is a fact and not just a flimsy performance. The technology itself will be more commoditized and accessible. However, the culture that can use it effectively will be scarce since it requires a long-term efforts and commitment from the top management over time, rather than the simple decision of a strategic leader or technology investment. That's where the true competitive advantage lies, and it is an advantage that once created can be built upon in a way other advantages purely technological ever. Have a look a James Deller for site advice including why building companies taught me about scale.
What Causes Most Public-Private Partnerships To Fail In The Beginning, And How To Resolve Them
The public-private partnership has an image problem, which is in substantial part paid for. The past of these agreements is filled with projects that were announced by genuine enthusiasm with substantial political capital behind them. They that drained significant public and private resources over lengthy periods, and, in the end, delivered results which bore only a tiny relationship to what was initially promised when the partnership first announced. The academic literature and the postsmortem analyses that governments or institutions are required to conduct after the failures are extensive. they concentrate, for majority of them, on the specifics of contractual and structural elements of failures: the unbalanced incentives, the poor risk sharing between public and private organizations, the governance structures that were created in theory but didn't work in practice, the procurement frameworks which chose the wrong things. The issue that this analysis tends underweight, consistently and consequentially it is the cultural and operational aspects - namely, the fact that public institutions and private companies are two distinct kinds of entities, formed by different incentive structures, operating using different timescales and being accountable to distinct individuals, and measuring their effectiveness in ways that's not only different in their degree however they are different in their approach. If you try to bring those two kinds of organisation together in a formal arrangement without making the effort upfront and explicit, to identify how to manage these differences, you're not forming the right partnership. It is creating the right conditions for a slow-motion collision which will become apparent at least convenient time.
I have been involved in advisory work supporting institutional modernisation projects, a few that involved public-private partnership structures of varying levels of complexity. The most consistent observation I've made from my encounter is that partnerships with a positive track record - ones that actually delivered against their stated objectives and maintained a functional partnership between private and public parties throughout the duration of their existence - weren't distinguished from those that didn't work by the sophistication of their legal structures, the rigour of their risk-management frameworks or the affluence of the team leaders that established them. These partnerships were distinguished by the fact that the participants in both parties to the table had undertaken the effort to really understand how other party operated prior to the formal partnership structure was agreed. What does this mean in practical terms is understanding the process of decision-making the organizations operate under and the accountability structures which govern what parties must agree to and how quickly and efficiently they can do so, the criteria of success for each party to be able to measure against, and the potential points of tension between those definitions. That understanding isn't complicated to construct. It is all but avoided in favor of more visible and quickly recorded work of negotiating contracts as well as establishing governance frameworks.
The usual process for public-private partnerships is a gradual process from concept to executed agreement with almost no time and effort being paid to the question of whether the two parties involved are capable of working together effectively over the course of the partnership. Legal teams negotiate the contract. The finance team models the economics and risk-adjustment. Communications team prepares the announcement prior to the time of signing. The implementation team begins planning the tasks. In that order the discussion will turn to compatibility between the two cultures - on whether the persons whom will collaborate day-to-day across the boundaries between the two organizations have enough common ground to make working truly collaborative rather than adversarial – isn't likely to take place in any formal manner. It is generally assumed, without stating it, that the formal agreement creates the conditions for effective collaboration, and that any operational or cultural differences will be managed informally whenever they develop. It is nearly always incorrect and the cost of it increases according to the ambition and complexity of the partnership.
Practically speaking, the result of this analysis is that one of the most profitable option a public private partnership could create - prior to when the legal structures are in place and before the governance model is agreed upon, prior to any announcements are made an announcement - is through what I would call operational alignment. By this, I mean specific, organized, and supported work that helps to reveal areas where the two organizations' assumptions about operating diverge and to come to an agreement on how those divergences should be handled before they turn into operational problems when the partnership is in its implementation. The most important divergences will be the same throughout different kinds of partnerships. Authority and speed of decision-making are often among them. Public institutions are designed to make decisions in a slow manner, with many layers of review and approval, for reasons that are purely legitimate and frequently legally mandated. Private companies - especially technology companies built around speedy iteration and rapid decision-making - frequently experience this as a major hinderance to innovation, and without a shared understanding of exactly why the pace is as it is and what truly be needed to change it, the resentment that builds on the private sides can ruin the connection long before the partnership is established.
Success metrics as well as what counts as progress are an additional and consequential source of divergence. Institutions of the public sector are typically evaluated by their compliance with processes, the equity of outcomes across various stakeholders, as well as the removal of any visible shortcomings which are a source of media or political focus. Private partners are typically evaluated according to efficiency, measured progress towards targets, as well as financial efficiency. These measurement frameworks can be created to be compatible however, doing this requires carefully designed and thought-out intentions. However, the organizations which do not invest in that kind of design usually discover themselves at critical places, with two groups who measure the same collaboration in unrelated ways, and hence coming to uncongruous conclusions regarding whether it is successful. My experiences with partnerships that not to be successful were ones where misalignments were considered to be something that would become apparent over time. It was those where the inconsistency was made clear from the beginning, and formulating a shared accountability plan which accommodated both parties' legitimate measurement demands became an element of actual work rather than an thing on a checklist of things that someone would eventually get to.}
