Short answer
Beyond a few dozen people, a company loses direct knowledge of what its teams actually do, who holds critical expertise, and why performance varies across units. Traditional methods rely on manager filters and static forms, which delay insight and flatten nuance. A working alternative talks directly to every employee through personalized interviews, captures real work as it happens, and makes the organization queryable: you can ask why one team outperforms another, identify who already knows how to solve a problem, and transmit proven practices to those who need them.
Why do companies lose knowledge of themselves as they grow?
A distribution company told us in June that interviewing colleagues is not consistent across levels. Directors, pillars, and frontline managers do not conduct interviews the same way in France as they do in Italy. The quality of the conversation depends on the manager, creating what one executive called a lottery. As a result, field information arrives filtered, delayed, and incomplete. The company wanted to talk directly with its teams.
This is not unique. In a retail organization with tens of thousands of employees, more than sixty per cent of the workforce is under thirty. Heterogeneity in how managers handle reviews is high. Some are excellent, others are not, and the company had no way to standardize the conversation without imposing rigid forms that miss context. The problem is structural: once you pass a certain size, the organization becomes opaque to itself. As IBM notes, organizational development is a planned, systematic process of changing strategies, procedures and culture to improve performance, but this requires knowing what is actually happening on the ground.
What does it mean to make an organization queryable?
An international organization we spoke to in June built a system that conducts personalized interviews with every employee. Each conversation is singular: the system knows who you are, your role, your business unit, your manager, and what you said in previous interviews. If a store assistant mentioned in February that learning English would help serve international customers, the system reminds them during the next review and offers to arrange training.
Every interview is analyzed against the company's pillars: customer focus, execution excellence, innovation, and others. The system identifies explicit intentions, implicit aspirations, weak signals, and areas of improvement. Most importantly, the organization becomes queryable. A head of people development can ask: why does the US team perform better on customer focus than Australia? The system compares both, identifies what the US does well, finds champions who can help, and suggests concrete actions to close the gap.
How do you capture real work without adding burden?
The system we saw operates through weekly or monthly interviews, depending on the role. Employees receive an email or a notification in their global app. They can speak live or use a reflective mode where they talk, read what they said, and continue. The majority prefer the reflective mode. The conversation is not a questionnaire. If a manager rates someone poorly, the system asks for specific examples and refuses vague judgments. One executive tried to write that a colleague was stupid; the system redirected him to describe behaviors and outcomes instead.
KPIs are injected into the conversation. A manager preparing a review can ask to see all the KPIs at once, so they have the full picture before commenting. The system does not replace the human conversation; it prepares both sides so they arrive with a clear, structured document. This is especially valuable for young employees who may not know how to articulate their development needs.
What patterns emerge when you talk to everyone?
A campaign covering several hundred interviews revealed that retail and commercial orientation was an area for improvement across departments. The system showed distribution by business unit, allowing the head of HR to drill down into specific teams and individuals. One employee explicitly mentioned stress and working at twice capacity. The system flagged this as a potential concern. The HR team could see the context, translated into their language, and take action immediately.
This is about early intervention. In large organizations, information passes through the manager filter, creating a temporal gap. By the time a problem surfaces, it is often too late. Direct, regular conversations with every employee compress that gap. From a legal standpoint, it also demonstrates that the company has put measures in place for mental and physical health.
How do you transmit what works from one unit to another?
The same organization asked: how can we help Australia become more customer-focused, based on what the US and another subsidiary do well? The system analyzed what Australia lacked, what the US excelled at, and proposed three concrete shifts:
- Move from reactive to proactive support
- Treat internal operations as key customers
- Formalize feedback loops
It also identified champions from the US team who could help. This is the core of organizational development at scale: taking the genius of the company and using it to help other parts of the company. Generic training does not do this. A chief people officer told us in June that she needed to train teams but wanted internal training, not generic modules. The system now generates training content based on real interviews, in the format employees prefer: text, podcast, conversation, comic strip, or short video. Cultural adaptation is automatic; humor and tone differ between France, Portugal, and the Middle East.
According to Maryville University's guide, organizational development involves changes in strategy, structure, and processes focusing on an entire organizational system. The approach described here operationalizes that definition by making the system's actual workings visible and actionable.
What does a manager see when preparing a review?
A manager preparing a review selects a team member and rates them on eight pillars: organizational knowledge, skill set, quality of work, and others. The system then conducts a live or reflective interview, asking the manager to justify each rating with concrete examples. If the manager skips a question, the system moves on but insists on specificity. The goal is not to judge but to gather objective observations.
After the interview, the manager sees a detailed breakdown of the employee's strengths and development areas, explicit and implicit aspirations, career projections, resources to activate, obstacles identified, and the management style that would help them most. The manager can compare the employee to others in the same payroll band, department, or business unit. This is not a static report; it is a dynamic view that improves with each interview.
How do you ensure the system is auditable?
In Europe, the AI Act requires auditability for systems that make decisions about people. Every rating the system assigns is backed by quotes from the interview. You can see why the system rated someone on a given dimension, and you can always view the full transcript. This is not optional; it is a legal requirement and a design principle. The system does not make decisions; it gives every manager and HR professional the tools to make informed decisions based on fact.
One executive told us that the system is hosted internally, on their own instance, reviewed by IT, and fully integrated into their Microsoft Azure environment. This is part of the infrastructure, with the same security and compliance standards as any other system.
What happens when job descriptions do not match reality?
In large organizations, most developers were listed as architects. The reason was salary: when someone wanted a raise, their title was changed. The system detects misalignment between what an employee says they do and what their job description says. It can generate an amended job description or flag the discrepancy for manual review. The goal is to have a dynamic, accurate view of the organization, not a static document that no one updates.
This matters because organizational development depends on knowing what people actually do, not what a form says they do. A chief digital officer told us that his company had no centralized place for job descriptions. They were not in Workday, not in Day Force, nowhere. The system became that place, but the ideal is to integrate it with the source of truth in the existing HR system.
How do you identify problematic management patterns before it is too late?
The system calculates a retention risk indicator for each manager based on their team's interviews. If every member of a manager's team shows declining engagement over time, the system flags it. This does not mean the manager is at fault; it means HR needs to investigate. The same applies to excessive workload, job misalignment, and other weak signals. The system does not make accusations; it surfaces patterns that would otherwise remain invisible until someone quits.
A head of people development told us that the system allows them to be very reactive. Sometimes it is just sending an email to ask how someone is doing. But from a legal standpoint, it shows that the company is paying attention and taking action. In a large workforce, this is the difference between knowing your organization and guessing.
What is the method for turning interviews into training?
The process has four steps:
- Conduct regular interviews with every employee, capturing what they do, how they do it, and what they need.
- Analyze interviews against the company's pillars, identifying strengths, gaps, and champions.
- Compare units to find what works and what does not, then extract the practices that explain the difference.
- Generate training content in the format employees prefer, using the company's own language, processes, and examples.
The system suggests training topics based on the interviews. A people officer can select a topic, generate a full training plan, edit it, and produce modules in text, audio, video, or print. The content is not generic; it is built from what the best performers in the company actually do. This is how you scale expertise without losing it in translation.
How do you compare performance across countries and cultures?
Organizations operating in dozens of countries need the system to adapt not just the language but the cultural context. A training module on customer service is not translated; it is rewritten with local humor, tone, and examples. The system knows that what works in France does not work in the Middle East, and it adjusts accordingly.
This extends to compliance. Different countries have different rules about what you can ask employees, how you store data, and who can see it. The system manages consent and validation by country and subsidiary. This is a working system that handles the complexity of a global workforce.
What does this mean for the future of organizational development?
A chief digital officer told us that his company is at the beginning of what it can do with data. They have sensors on ships, generating data constantly, but they are only starting to exploit it. The same is true for employee data. Most companies have performance reviews, engagement surveys, and exit interviews, but they do not have a continuous, structured conversation with every employee. They do not know what their organization knows, who knows it, or how to move it from one place to another.
Organizational development at scale is not about forms or surveys. It is about making the organization legible to itself, in real time, so that leaders can act on what is actually happening rather than what they hope is happening. This requires talking to everyone, capturing real work, and building a system that can answer questions like: why does this team win? Who already knows how to solve this? What should we teach next? The companies that do this well will have an advantage that no generic training program can replicate.
If you want to see how this approach would work for your organization, we can walk through your specific case in a demonstration tailored to your context.