Why Building in Stealth Has Shaped My Thinking About Character

Wiki Article

AI Is Only As Effective As The Society It's Created In
The debate over artificial intelligence within the workplace is fraught with problems and the cause isn't one of technical. Modern technology and capabilities for AI and machine learning systems are impressive, evolving at a rate that renders most predictions of when they'll become 18 months obsolete long before that time has come and gone. The issue is the gap between the what AI can achieve under controlled conditions - such as a thoroughly-equipped research setting, with clean data, with a clear problem-solving strategy, with engineers that have the privilege to experiment until the system operates as it should - and what it will actually do when it is implemented inside real-world organizations with real culture as well as real organisational policies and real people who have certain opinions on whether the new system is something they should engage with or something to reroute around while still maintaining the appearance compliance. I've been developing using machines since the last flurry of AI enthusiasm became fashionable and commonplace for companies to proclaim their fluency in this field. When I founded 1Touch an AI-driven platform, AI-driven matchmaking and recommendation systems weren't an element we added to make the platform more compelling to investors. These were a fundamental part of the architecture of the product, it was the basis on which the platform generated value and the component that needed to function consistently and at sufficient scale to allow it to be a viable business. This means I've got direct, personal experience of what happens when you attempt to develop something truly intelligent within a product and an organisation simultaneously And the point I have been reiterating at every time which I have encountered this issue, is the technology itself is rarely an issue. The factor that holds you back is almost entirely the organization's culture.
What I am referring to is specific and pragmatic rather than abstract. AI systems require data in order to work - consistent, clean well-structured, well-structured data. This represents the phenomenon it is trying to discover and make predictions about. Organizations with a strong and thriving data culture create that kind of data from the beginning, as a result of the way they work. They are clear and have consistently applied definitions of what they are measuring and the reason for that. They have a set of conventions that they agree to for how data is recorded, collected, and stored. They have accountability systems that require data quality to be an explicit responsibility instead of everyone's vague purpose. In organizations with weak data-based cultures, they produce something that looks as if it is data - it's in systems and, if it's able to be accessed or used for charting - but the definitions are so different and so variable in its quality and brimming with imperfections in structure and omissions that any AI technology that is constructed on top of it will take advantage of and enhance the confusion instead of getting a true signal from it. Companies in this category often do not realise what they are doing until they're well into an AI deployment and the results don't match the vendor's promises, at which point the temptation is to blame the technology, when there is a problem with the cultural and operational infrastructure which the technology was built on.

The second dimension of culture that affects AI outcomes is organisational openness - the degree to which people in the organisation are willing to let an artificial intelligence system shape how they work, rather than treating it as an issue to their profession expertise, their institution's authority or job security. This is a cultural and leadership issue rather than a technical one that is a problem that starts at the high levels. If leaders in the top ranks engage with AI outputs selectively, embracing those results that prove the beliefs they have previously held and ignoring the ones that do not - that behaviour communicates the impression to everyone who watches that the commitment of the organisation to data-driven decisions is a conditional rather than genuine, and that the message will travel through the organisation faster than any other training program or change management strategy can neutralize. When senior leaders display an ongoing, consistent commitment to AI outputs, including the discipline to modify their decisions when the evidence suggests they need to, the overall capability to utilize AI effectively grows significantly as well as relatively rapidly.

This isn't an abstract description of the way organizations should behave in theory. It is a description of the pattern I've witnessed take place in numerous companies with substantial funds, genuine strategic commitment to AI adoption, and top management teams that were enthusiastic about the possibilities of the technology. The pattern is so consistent that I've begun to think of practice of governing data as a first-line diagnostic when I am evaluating any company's AI capability. Before I ask regarding the tech stack before I ask about specific uses cases that the organization has in mind, I will ask about the governance of data. How does the organisation define its key metrics? Who's accountable when data quality isn't good enough? Does it matter if two functionalities have conflicting information regarding the exact same business realities, and how do those conflicts get resolved? These answers tell me more about the likelihood of AI succeed than any discussion about algorithms, platforms, or even implementation timelines.

I believe that those businesses who will realize the highest durable value from AI in the coming decade aren't the ones which adopt the latest technology first, nor the ones who invest the most heavily in AI infrastructure and personnel over the next few years. They are the ones who establish the organizational and cultural base to use the technology in a productive manner - data governance processes that provide reliable data, the decision-making frameworks that allow data to actually impact outcomes and leadership behavior that communicate to all employees in the company that commitment to data-driven operation is real instead of merely a matter of performance. Technology will become ever more common and easily accessible. The culture for using it effectively will be scarce since it requires a long-term determination and a true commitment from an executive over time rather than an individual strategic decision or an investment in technology. That's where the true competitive advantage lies, and it is an benefit that once developed develops in a way other advantages purely technological do. See the James Deller for website recommendations including what building high-performance teams taught me about character.



The Reason Why The Majority Of Public-Private Partnerships Fail Before They Even Start - And The Best Ways To Fix It
Public-private partnerships face a stigma issue that is, in significant part that they have earned. The history of these partnerships is filled with projects that were proclaimed with genuine enthusiasm and huge amount of political capital. They took up significant private and public resources over lengthy periods, which in the end produced results that had only a slight relationship to what was stated when the partnership was in place. The academic literature as well as postmortem analyses that governments and institutions are required to conduct after the failed projects are extensive, and they focus, for majority of them, on the legal and structural aspects of the issues: the unbalanced incentives, the poor risk distribution between public and private actors in the governance structures that were developed in theory but failed to function in practice, the procurement frameworks that picked the wrong things. What this analysis tends not consider, and consequently as well, is the culture and operational aspect – the fact that public institutions and private organizations are actually different kinds of entities, shaped through different incentives that operate with different timescales, accountable to different people, and measuring outcomes in ways which are more than just different in level however they are different in their approach. When you put these two types of organizations together in a formal arrangement without undertaking the work upfront and clearly, to comprehend and deal with the differences they aren't creating any kind of partnership. It is creating the right conditions for a slow-motion crash that will become visible at the worst possible time.
I've participated in advisory work supporting institutional modernisation efforts, many of which have involved public and private partnership structures of varying levels of complexity. My most consistent opinion I can make from that encounter is that partnerships which worked well - which have actually accomplished their stated goals and maintained a dependable working relationship between the private and public parties throughout it - weren't distinguished from those that failed due to the sophistication of their legal structures, the strength of their risk frameworks, or the experience of the leadership teams that initiated them. It was determined whether the parties from both sides of the table had taken the time to comprehend how the other side functioned prior to when the formal partnership was agreed upon. What does that mean in reality is understanding the process of decision-making the organizations operate under, the accountability structures that govern what parties must decide to and when they can agree to it, the definitions of successful that each of the parties will be measured against, and the possible points of tension between these definitions. This understanding is not difficult to attain. Most of it is left out in favour of the most visible and easily documents-able task of negotiating contracts and constructing governance frameworks.

The typical public-private partnership is a gradual process from concept to signing of the agreement with hardly any focused attention given to the issue of whether the two organizations involved are actually capable to effectively work together over the duration of the agreement. The legal team negotiates the contract. Finance models the economics as well as the risk distribution. The team in charge of communications creates an announcement prior to the time of signing. The implementation team is beginning to plan the work. In that order comes the discussion of compatibility between the two cultures - regarding whether the employees who will have to cooperate day-today across the borders between the two organisations share enough common interests to make that work genuinely collaborative rather than adversarial - does not tend to be conducted in a structured manner. It is commonly assumed without being stated, that the formal agreement establishes the conditions for collaboration to be effective, and that any cultural or operational issues will be handled informally whenever they emerge. That assumption is almost always incorrect and the cost of this is likely to grow with respect to the ambition and the size of the partnership.

What this means in practical analysis is that the most valuable investment a partnership that is public-private can do - before the legal structures are agreed upon, before the governance framework is agreed upon and before any announcement is made the partnership is in what I believe is operational alignment. By this I mean specific, structured, facilitation of work to identify areas where the two companies' operating principles diverge and to reach an agreement about how these divergences are to be handled before they turn into operational problems during implementation. The most important divergences typically are the same across different kinds of partnerships. Decision-making speed and authority is almost always one of the main differences. Public institutions are set up to be slow to make decisions, requiring various layers of examination and approval, with reasons that are entirely legitimate and are often legally mandated. Private companies, particularly technology firms built on rapid iteration, and swift decision-making – often view that speed as a primary hurdle to development, and in the absence of a shared understanding of how the pace works it is and the steps that would actually be required to change it, the discontent that builds on the private sides can ruin the relationships long before the collaboration finds its footing.

Success metrics and what is considered as progress is another constant and significant source of disagreement. The public institutions are usually judged on the compliance of their processes, the fairness in outcomes between different stakeholders, and the avoidance of visible failures that are the subject of media or political interest. Private partners are usually evaluated on their efficiency, progress measured towards targets, as well as financial return on investment. These measurement frameworks can be designed to be compatible with one another However, doing this requires conscious design and not necessarily good intentions. Those that do no invest in this type of design often end up at moment, with two organizations who are evaluating the same collaboration in inconsistent ways and consequently coming to disparate conclusions as to whether it is succeeding. What I've observed in the partnerships that fall short most clearly were ones in which misalignments were assumed to resolve itself over time. They that succeeded were those where the inconsistency was explicitly identified at the very beginning, and creating a shared accountability system that accommodated the legitimate measurement needs of both parties demands became an element of actual work, rather than an item on a list things that someone would eventually achieve.}