The End of One-Size-Fits-All Organizations

For most of the digital era, organizations have lived with an awkward bargain.

They understood their own work better than anyone else. They knew their people, their habits, their exceptions, their history, their pressures, their language, and the thousand small decisions that make the organization actually function. But they usually lacked the ability to build systems around that knowledge.

Software companies, meanwhile, could build the systems. But they usually had to build them for broad markets. A tool for “church management” had to serve thousands of churches. A learning platform had to serve many kinds of schools. A project management system had to work for agencies, nonprofits, contractors, startups, and corporate teams. A customer relationship platform had to be flexible enough for almost any sales process.

That arrangement made sense. It gave organizations access to powerful tools they could never have built on their own. But it also created a quiet reversal: instead of systems being shaped around the work, the work was often reshaped around the systems.

That is one of the reasons organizational life can feel so strangely fragmented.

The church has a database, a giving platform, a website, a scheduling tool, a file system, a livestream archive, a planning tool, a messaging system, and several informal spreadsheets that everyone depends on but no one wants to admit are essential. The school has a learning management system, a student information system, curriculum platforms, enrollment forms, parent communication tools, compliance documents, accreditation evidence, staff notes, classroom workflows, and a few key people who know how all of it really fits together. The small business has invoices, customer records, inboxes, project boards, templates, payment systems, shared folders, onboarding documents, and an employee who has become the living map of the whole operation.

In each case, the organization may have plenty of software. What it often lacks is coherence.

The pieces may work individually, but the organization itself is still held together by memory, habit, improvisation, and a few people who know where everything is. The knowledge required to operate the organization already exists. It lives in people, documents, workflows, history, culture, policies, philosophy, communication patterns, and accumulated experience. But it is rarely structured as a working asset. It is usually scattered, duplicated, buried, assumed, or trapped inside individuals.

That is the problem worth paying attention to.

Not because every organization needs a new app. Not because artificial intelligence should suddenly become the center of every office, school, church, or business. And not because the future can be reduced to chatbots answering questions over old PDFs.

The larger possibility is more interesting than that.

The possibility is that organizations may finally be able to build more directly around what they already know.

The Limit of Generic Software

Generic software has been necessary, and much of it has been genuinely helpful. It has made small organizations capable of doing things that once required far more staff, infrastructure, or money. Most leaders would not want to go backward.

But generic software has a built-in limitation: it is designed around patterns, not particularity.

It can support common needs. It can offer fields, templates, automations, dashboards, calendars, records, permissions, integrations, and reports. But it cannot fully understand why a certain church handles member care the way it does, why a school’s parent partnership model requires unusual communication, why a nonprofit’s donor language has to preserve a certain moral tone, or why a small business treats one part of customer onboarding as more important than the rest.

So the organization adapts. It builds workarounds. It creates side processes. It invents naming conventions. It trains staff to remember the exceptions. It keeps extra notes outside the official system because the official system has no natural place for them.

Some of this is harmless. Some of it is even wise. No tool can carry the whole life of an organization. But when the gap between the real work and the official system gets too large, the organization starts paying a hidden tax.

People spend time looking for information that should be findable. They repeat explanations that should have been captured. They depend on employees who remember what the system does not. They maintain multiple versions of the same truth. They make decisions without the history that should have informed them. They communicate inconsistently because the organization’s language lives in scattered places.

The numbers behind this are not small. McKinsey Global Institute estimated that improved communication, collaboration, and knowledge sharing through social technologies could raise the productivity of high-skill knowledge workers by 20 to 25 percent.[1] In the same report, McKinsey found that interaction workers spent a significant portion of the workweek communicating, collaborating, and searching for information needed to do their jobs.[2] The precise percentages will vary from one organization to another, and a church or school should not measure its health only in productivity gains. Still, the underlying point is familiar: scattered knowledge costs real time.

A separate experiment on professional writing tasks found that access to ChatGPT reduced task completion time and improved average output quality among college-educated professionals performing writing assignments.[3] A controlled study of GitHub Copilot found that developers using the tool completed a programming task 55.8 percent faster than those who did not.[4] These studies do not prove that every use of generative AI will be valuable, and they certainly do not prove that AI automatically improves organizational judgment. But they do show that, in certain kinds of knowledge work, the technology can meaningfully reduce the distance between intention and execution.

That matters because the historic bottleneck has often been execution.

People inside the organization know what would help. They know the parent guide is confusing. They know the volunteer training does not match the actual ministry. They know the onboarding process depends too much on a particular staff member. They know the same donor explanation gets rewritten every quarter. They know the school model is compelling in conversation but still too hard to understand on paper. They know the business has a process that works, but only when one experienced employee handles it.

Until recently, turning that knowledge into better systems was expensive, slow, or simply unrealistic for smaller organizations.

That is beginning to change.

The Shift From Tools to Infrastructure

It is tempting to describe the change by saying that organizations will “use AI.” That is true as far as it goes, but it is too thin.

Most organizations are already using AI in some form, even if unevenly. The Stanford AI Index, drawing from McKinsey’s survey work, reported that organizational AI adoption reached 55 percent in 2023, up from 20 percent in 2017.[5] Since then, generative AI has moved into search, writing, coding, design, customer service, office suites, and countless specialized platforms. Deloitte’s enterprise research has likewise tracked a surge of generative AI experimentation, along with growing pressure on leaders to produce value while managing risks around governance, talent, and trust.[6]

But “using AI” is not the same thing as becoming more intelligent as an organization.

An organization can have a chatbot and still have fragmented knowledge. It can automate reminders and still have a broken workflow. It can generate more content and still communicate less clearly. It can summarize documents and still fail to preserve the judgment behind them. It can adopt the newest platform and still be dependent on the same three people who know how everything really works.

The more important shift is not from no AI to AI. It is from disconnected tools to institutional intelligence.

That means treating an organization’s knowledge, memory, processes, language, and judgment as part of its working infrastructure. Not as clutter. Not as archives. Not as “content” sitting around the real work. As infrastructure.

A policy is not just a document. It is preserved judgment. A curriculum guide is not just a file. It carries educational philosophy. A volunteer manual is not just instructions. It teaches people how to care, serve, and notice what matters. A customer response template is not just convenience. It preserves the kind of promise the business wants to keep. A sermon archive, board decision, parent letter, staff workflow, training video, or strategic plan may contain knowledge that should continue serving the organization long after its original moment has passed.

The question is whether that knowledge can be found, trusted, adapted, and put to work.

This is where artificial intelligence could become more than a productivity layer. Used carefully, it can help organizations retrieve knowledge, connect documents, draft from established language, build internal tools, support workflow design, summarize patterns, expose inconsistencies, and make institutional memory more usable. Used carelessly, it can produce confident nonsense, flatten the organization’s voice, leak sensitive information, automate unclear processes, and create more noise.

So the issue is not enthusiasm versus caution. Both are needed.

The real issue is whether organizations will build the interpretive layer that makes the technology useful.

The Knowledge Is Already Inside

Michael Polanyi’s famous statement that “we can know more than we can tell” remains one of the clearest ways to understand this problem.[7] Organizations are full of knowledge that has never been fully articulated. People know how to handle a delicate situation. They know which phrase will land badly with parents. They know why a donor cares about one part of the mission more than another. They know which process looks clean on paper but fails every August. They know when an exception is wise and when it undermines the whole system.

This kind of knowledge is valuable precisely because it includes judgment. It cannot all be reduced to a checklist, and it should not be. But if it remains entirely tacit, the organization becomes fragile. The knowledge is available only when the right person is available. When that person leaves, burns out, changes roles, or simply cannot be present for every decision, the organization loses more than labor. It loses memory.

Ikujiro Nonaka and Hirotaka Takeuchi’s work on organizational knowledge helps explain what has to happen. Healthy organizations learn by moving between tacit knowledge and explicit knowledge: between the experience people carry and the forms of knowledge that can be shared, examined, practiced, and passed on.[8] This is not the same as documenting everything. It is the discipline of translating what matters into forms that can serve others.

That translation can take many forms. It might be a better onboarding process, a searchable knowledge base, a staff guide, a decision framework, a parent communication system, a donor language bank, a volunteer training pathway, a curriculum map, a workflow dashboard, or an internal assistant grounded in the organization’s own material.

The form matters less than the movement.

The organization is taking what it already knows and making it usable.

That movement is especially important for organizations whose work is deeply human. Churches, schools, nonprofits, ministries, and small businesses often depend on trust, continuity, formation, care, and judgment. Their knowledge is rarely just technical. It is relational and interpretive. A church does not merely need to know who is scheduled to serve; it needs to know how its philosophy of ministry shapes volunteer care. A school does not merely need to record assignments; it needs to communicate what kind of partnership parents are entering. A nonprofit does not merely need donor records; it needs language that carries the moral weight of the mission without becoming manipulative or generic. A business does not merely need process documentation; it needs to preserve the customer judgment that made the business trustworthy in the first place.

This is where the opportunity becomes broader than software.

It becomes a new kind of organizational work: discovering what the organization knows, structuring it, connecting it to workflows, and building practical systems around it.

The Return of the Particular

One of the most hopeful possibilities in this moment is that smaller organizations may become less dependent on one-size-fits-all systems.

That does not mean they will stop using established platforms. They will not, and they should not. There will still be tools for accounting, email, scheduling, payments, records, publishing, compliance, and a hundred other needs. But the layer above and between those tools may become more adaptable.

AI-assisted software development is already changing what can be built and who can participate in building it. The GitHub Copilot experiment is only one example, but it points toward a larger shift: the ability to create working software is becoming less sealed off from the people who understand the work.[4] At the same time, agentic systems and workflow orchestration are moving from speculative language into enterprise roadmaps. Accenture’s Technology Vision described a movement from AI performing individual tasks toward AI agents that, with oversight, can work together and act on behalf of people and organizations.[9]

This does not mean every organization is ready to run complex agent systems, and it does not mean the risks are small. NIST’s AI Risk Management Framework is useful precisely because it refuses to treat AI as magic. It emphasizes governance, mapping, measuring, and managing risk throughout the lifecycle of AI systems.[10] That kind of caution is not a brake on innovation. It is part of what makes responsible innovation possible.

Still, the direction is significant. As the cost of building internal tools falls, and as retrieval, workflow, and automation systems become more accessible, the bottleneck shifts. The hardest question is no longer only, “Can this be coded?” Increasingly, the harder question is, “Do we understand the work well enough to build the right thing?”

That is a very different kind of challenge.

It requires careful listening. It requires organizational diagnosis. It requires writers who can clarify language, operators who understand process, leaders who know the mission, technologists who can build responsibly, and experienced staff who can explain what really happens. It requires attention to the particular.

The particular church. The particular school. The particular nonprofit. The particular business.

Not “users” in the abstract. Real people doing real work inside an organization with a history, a mission, a culture, and constraints.

This is why the opportunity should not be reduced to automation. Automation is part of it, but not the whole. The deeper opportunity is fit.

A school could build systems that reflect its educational philosophy rather than forcing parents through generic communication. A church could make decades of teaching and ministry practice useful for discipleship, training, care, and continuity. A nonprofit could preserve its voice across donor communication, volunteer onboarding, and program operations. A small business could make its best customer judgment teachable instead of leaving it to personality and memory.

None of that requires grand promises. It requires practical imagination.

The Risk of Shallow Adoption

There is also a way to do all of this badly.

Organizations can bolt AI onto broken processes and call it innovation. They can create chatbots over messy knowledge bases and wonder why the answers are uneven. They can generate more communication without becoming more coherent. They can automate workarounds that should have been reconsidered. They can hand sensitive institutional knowledge to tools they do not understand. They can let generic language replace the words that once made the organization trustworthy.

The result would not be intelligence. It would be faster fragmentation.

This is not a theoretical concern. Deloitte’s research has repeatedly noted that enterprise leaders feel pressure to realize generative AI value while managing concerns around governance, talent, risk, and organizational readiness.[6] NIST’s framework exists because AI systems raise real questions about validity, reliability, safety, security, accountability, transparency, privacy, and bias.[10] The more AI becomes woven into organizational operations, the more these questions matter.

A small organization may be tempted to think governance is only for large corporations. That would be a mistake. Governance does not have to mean bureaucracy. At its best, it means the organization knows what it is doing, why it is doing it, who is responsible, what data is involved, what risks are acceptable, and how the system will be corrected when it fails.

That kind of discipline will matter for churches and schools as much as for companies.

Maybe more.

Because the work often involves people’s children, spiritual care, donor trust, sensitive family information, educational records, internal conflict, financial stewardship, or private communication. A careless system can do more than waste time. It can weaken trust.

That is why the future worth pursuing is not simply AI-enabled. It is judgment-enabled.

Technology should extend good judgment, not replace it. It should make the organization’s best knowledge more available, not bury people under more output. It should help staff and volunteers act with greater clarity, not force them to trust a system no one has shaped or examined.

What Might Become Possible

The possibility, then, is not one product or one model.

It is a category of work that many organizations are going to need.

Some will need help gathering and structuring institutional knowledge. Some will need help turning scattered documents into a usable internal knowledge system. Some will need help translating a philosophy into parent-facing or customer-facing language. Some will need help redesigning workflows that grew by accident. Some will need internal tools that connect existing systems in ways their SaaS platforms do not. Some will need communication ecosystems that preserve voice across websites, email, social content, onboarding, and long-term public explanation. Some will need AI-supported assistants that are narrow, governed, and grounded in trusted materials.

And some will first need someone to sit with them long enough to understand how the work actually happens.

That may be the most important part. The new technology makes certain things possible, but it does not remove the need for wise interpretation. Someone still has to ask what knowledge matters. Someone has to notice where the workflow breaks. Someone has to distinguish between a process that needs automation and a process that needs simplification. Someone has to protect the organization’s voice from becoming generic. Someone has to understand that a church’s doctrine, a school’s philosophy, a nonprofit’s mission, or a business’s promise cannot be treated as interchangeable content.

The organizations that thrive in this next stage will probably not be the ones that adopt the most tools the fastest.

They will be the ones that learn how to make their own intelligence operational.

That means knowing what they know. Preserving what should not be lost. Translating experience into usable form. Building systems that fit the work. Communicating with coherence. Using technology where it serves the mission, and refusing it where it only adds noise.

The opportunity is large, but it should be approached without hype. Much of the work will look ordinary from the outside: sorting documents, interviewing staff, rewriting explanations, mapping workflows, building small tools, testing retrieval systems, cleaning up communication, training people, and revisiting assumptions.

But ordinary work can become foundational when it gives an organization a stronger way to carry what it has been entrusted with.

The future will not belong only to organizations with better software.

It may belong to organizations that finally learn how to build around what they already know.

Notes and Sources

[1] McKinsey Global Institute, The Social Economy: Unlocking Value and Productivity Through Social Technologies (2012). McKinsey estimated that improved communication, collaboration, and knowledge sharing through social technologies could raise productivity among high-skill knowledge workers by 20 to 25 percent.

[2] McKinsey Global Institute, The Social Economy full report (2012). The report discusses the significant amount of time knowledge workers spend communicating, collaborating, and searching for information needed to complete their work.

[3] Shakked Noy and Whitney Zhang, “Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence,” SSRN / MIT-related research (2023). In a preregistered experiment involving professional writing tasks, ChatGPT reduced completion time and improved average output quality.

[4] Sida Peng, Eirini Kalliamvakou, Peter Cihon, and Mert Demirer, “The Impact of AI on Developer Productivity: Evidence from GitHub Copilot” (2023). In a controlled experiment, developers with access to GitHub Copilot completed a programming task 55.8 percent faster than the control group.

[5] Stanford Institute for Human-Centered AI, 2024 AI Index Report. Stanford reports that, according to McKinsey survey data, 55 percent of organizations used AI in at least one business unit or function in 2023, compared with 20 percent in 2017.

[6] Deloitte AI Institute, The State of Generative AI in the Enterprise (2024). Deloitte’s enterprise research tracks adoption, expectations, and concerns around governance, talent, value realization, and risk as organizations move from experimentation toward broader generative AI use.

[7] Michael Polanyi, The Tacit Dimension (1966). Polanyi’s claim that “we can know more than we can tell” remains a foundational idea for understanding tacit knowledge.

[8] Ikujiro Nonaka and Hirotaka Takeuchi, “The Knowledge-Creating Company,” Harvard Business Review. Their work describes the relationship between tacit and explicit knowledge in organizational learning and innovation.

[9] Accenture, Technology Vision 2024. Accenture describes a movement from AI tools performing singular tasks toward agent ecosystems that, with oversight, can work together and act for people and organizations.

[10] National Institute of Standards and Technology, AI Risk Management Framework (AI RMF 1.0). NIST frames trustworthy AI around governance and risk-management functions, including mapping, measuring, and managing AI risks.