Workforce Capability Intelligence (WCI) answers questions about highly skilled roles and teams. Key is the intelligence cycle. The intelligence cycle was developed for handling national threats. Its methods are adaptable to understanding individual’s capabilities in the workplace. It can be used ethically. And it provides data that can be used in Workforce Capability Assessment.
For management decisions with high risk, confidence is critical, so we use multiple intelligence sources to create independent views.
WCI used to be an impossible dream because of the paperwork. Now it’s an achievable because of AI and systems integration.
Introduction
What is Workforce Capability Intelligence (WCI)?
WCI helps organisations build a deep understanding of the skills and capabilities of anyone in an organisation, or joining it. It’s a process, supported by current and emerging technologies and intelligence methods.
- For employees, it helps them plan their careers and get the best opportunities.
- For middle managers, it helps with planning their teams, so they can up-skill and reorganise.
- For senior executives, it provides deep and quantitative insights into the workforce, to support key decisions.
For more on the approach, see the FAQs in Workforce Capability Intelligence – a new future.
The intelligence goal
ThThe intelligence goal (or “framing question”) established the purpose for an intelligence investigation. In a state or police investigation, it’s challenges like “how big a threat is this group?” In WCI, it’s questions like:
- “Assess how well these candidates fit this vacant role”
- “Map the full set of skills being used by this specialist”
- “Identify the weaknesses in this team”
- “Benchmark this team’s profile, so we can reassess it after the upcoming changes.”
The results of the intelligence investigation is is typically a clear understanding and considered options for the next step.
Research questions
From the intelligence goal, the team create research questions (“testable assumptions”), many of which could be assessed on a scale. For example, when deciding on a person’s fit for a role:
- Has leadership skills
- Has planning skills
- … (etc)
- Has experience in our business domain
- Has experience of companies like ours
- … (etc)
- Is self-motivated
- Is fast-worker
- … (etc)
In practice, the goal and research questions tend to be repetitive, copied from one investigation to the next.
Sources
There are three types of intelligence source.
- Knowledge-gathering apps. For example, automated scans of a CV/resumé, or answering psychometric questionnaires.
- Human-to-human interactions, such as at interviews. Interviewers make notes and summations, and (if recorded) AI can potentially extract patterns from the transcript.
- Background tracking of employee capabilities by monitoring documents, mails and contributions to regular meetings. Line managers already do that using partial information, but AI can do it systematically from multiple sources.
(For more ideas, see Multiple intelligence sources in the introduction to WCI.)
A good investigation uses several sources to give different perspectives, to see if they support or contradict each other. We use a matrix, as below, to identify which sources can answer which research questions. When there’s no reliable sources for answering a research question, then we look for news intelligence sources, even if that involves labour-intensive human-to-human interactions.
Research Plan Matrix with the Intelligence Cycle
| Question-1 | Question-2 | Question-3 | |
| Source-1 | Good indicator | Poor indicator | |
| Source-2 | Poor indicator | Very good indicator | |
| Source-3 | Poor indicator |
Ethics
WCI replaces subjective judgements made on partial and second-hand information, but to do this it requires collecting, storing and analysing extensive employee data collected using AI and persistently intrusive methods.
- High security is needed to protect personal data.
- The legal and ethical considerations are critical – they’re introduced in the sections below.
Reliability and multiple sources
“Reliability”, as used here, refers to the reliability of intelligence sources and subsequently to the analysis conclusions.
In practice, we’re looking for the opposite: the likelihood of errors. Errors are cumulative when relying on a single source. An imperfect method, performed imperfectly, then analysed with an imperfect method is dangerously unreliable. So, keeping track of potential weaknesses is important.
To help decision-makers improve their trust in the conclusions, we use multiple intelligence sources to provide different perspectives on the same question. (For more, see Trust – focus on reliability in the WCI overview.)
The intelligence cycle
The intelligence cycle is a workflow, where responsibility passes to different people.
It’s a “cycle” because we frequently have to go back to the first phase, refine goals and collect and analyse more intelligence.
WCI uses multiple intelligence sources, processed with independent analysis methods to offer alternative opinions, and increase confidence in the intelligence.
Because of the multiple sources and specialist skills, WCI for a major decision can involve many people. For simple decisions, it’s much easier.
The entire cycle can sometimes be completed within a day, but that requires extensive automation and using existing intelligence sources, well-proven analysis and great coordination. Typically, it’s slower because of a need for more intelligence research or a changed brief. For some kinds of intelligence goal, such as a job appointment, the human delays may take weeks.
Roles and responsibilities
In large, structured organisations there could up to 7 roles directly involved in the intelligence cycle. In smaller or flexible structures, some people fill two or more roles.
Roles:
- The Client. The decision maker who needs the intelligence, and commits to any additional expenses. Typically a middle or senior manager.
- The Project Manager. Responsible for the smooth execution of the intelligence process and directing operational decisions to interpret the client’s needs. An experienced project manager would be able to handle this on top of other larger duties.
- The Research Lead. Responsible for collating all intelligence sources, building a factual and neutral understanding, and assessing the reliability of the intelligence. Human Resource personnel could be trained for this.
- Collectors. Individuals providing a single source of intelligence. For example, the people involved in interviewing a candidate. They don’t need training in the intelligence cycle, or access to the other intelligence.
- The Analyst who combines the research findings into an actionable report. It requires familiarity with analytic techniques. (A capable Analyst may also act as the Research Lead and the Project Manager.)
- The Publisher. For managing the compliance checks and workflow. These are mechanical tasks, that could be done by the Project Manager.
- Compliance checkers. Individuals separate to the above, to verify the conclusions, address any legal issues, and check the distribution lists are correct.
Related concepts
Talent Intelligence evaluates hard/technical skills and potentially also for soft/power skills and green/sustainability ones. WCI builds a broader picture by looking at individual’s cross-domain experience, crisis handling, productivity, quality and personal characteristics. For an example of Talent Intelligence using AI and new modelling techniques, see https://metatalent.ai/
Workforce Intelligence looks at underlying challenges about the workforce. It’s an HR tool. WCI is a line management tool, looking at individual roles, to help with appointment and upskilling decisions. For an example, see https://fuel50.com/2025/06/workforce-intelligence/
Workforce Capability Assessment evaluates the skills, knowledge, and competencies of employees across sections of an organization, such as to plan changes. As above, WCI is specific. For more, see https://skillpanel.com/blog/workforce-capability-assessment/ and https://acorn.works/blog/workforce-capability-assessment
Workforce Capability Modelling creates a unified view of the workforce, and individual teams. For more, see https://businesscasestudies.co.uk/what-is-workforce-capability-modelling/
Intelligence cycle steps
1. Planning and direction
Purpose of this phase: To agree the brief, the constraints and approach to be used. If the intelligence goes through iterations, then the brief is reviewed on each intelligence cycle.
Outputs: The brief. From the Analyst, the intelligence goal, an agreed format for the delivered results, a list of the individual testable questions, a list of all the intelligence sources to be used, and a Research Plan Matrix of which sources will be used for which questions. (The matrix provides a check for gaps and unnecessary duplication.) From the Project Manager, a budget, timescale, and the specialists to be used in data collection.
Automation: Use templates built from previous similar exercises, and prepopulate with as much detail as possible. (If it’s a new type of investigation, adapt generic templates or build a new one.) Typically, the template would be in the form of a multi-tab spreadsheet that shows the brief and subsequent status of collecting the intelligence. The confidential details of the intelligence are kept off it, so that the sheet can be shared.
Method: Clarify how this case differs from before, add the key subjects (such as people, skill and job roles), and identify the technical risks to quality. Review project details such as costs, timelines and external risks. If there are major exceptions, get executive decision to proceed.
Ethics: Select from list of considerations for this type of investigation, and note any exceptions.
Reliability: Set the expected quality of the results, for each source and the overall analysis. Note any special measures to be taken.
Responsibility: Project Manager, in close consultation with the Client (decision maker), Analyst and Research Lead.
Required expertise: Training and practice in the analytic and project methods of intelligence collection.
2. Data collection
Purpose: To collect raw intel from agreed sources for the Research Lead to provide an understanding.
Outputs: Evidence from each source, with notes on the intel’s reliability. Evidence can be quantitative data and observations. It can also be technical (automated) or human. For each source, there should be a summary, however the researcher may go back to the original data to verify detail or extract other details.
Automation: Data collection can be extensively automated, either directly or using AI large language modules to extract meaning from written or spoken content.
Method. There are 3 generic types of source for the intelligence cycle:
- Knowledge-gathering apps. For example, automated scans of a CV/resumé, or answering psychometric questionnaires.
- Human-to-human interactions, such as at interviews. Interviewers make notes and summations, and (if recorded) AI can potentially extract patterns from the transcript.
- Background tracking of employee capabilities by monitoring documents, mails and contributions to regular meetings. Line managers already do that using partial information, but AI can do it systematically from multiple sources.
Ethics: Consent and awareness are needed. For example, in an investigation to recruit a candidate for a critical new post, the candidate needs to have given approval, have awareness of the kind of data collected, how it will be used, and how it will be safeguarded.
Reliability: Collectors note suspicions of where quality may be compromised.
Responsibility: Data collectors for manual methods and specialist tools. Research Lead for overall coordination.
Required expertise: Collectors need training and experience only in the methods and tools they use. The Research Lead needs understanding of all sources for coordination, and especially for the next step.
3. Data processing (Research assessment)
Purpose: To provide a single “understanding” from the different intelligence sources.
Outputs: Neutral interpretation of the “facts”, with exceptions noted and indications of reliability. These are facts, without opinions or speculations, so that when decision makers question the analysis, they can see what it was based on.
Method:
- Collate evidence for or against the different hypotheses. (The Research Plan Matrix indications what is needed for each hypothesis.)
- Process intelligence into a standardised format for analysis and storage. (That should be largely or fully automated.) Store securely.
- Assess combined reliability and note gaps, for each of the hypotheses. The assessment should be manual, supported by automated collection of the notes from the collectors.
- Escalate if there are major problems.
- Create a research summary for the decision-maker and analyst.
- Support the Project Manager in creating a distribution list.
Ethics: With methodical care. Intelligence sources include highly sensitive data extracted from written and spoken material, some of which can be highly misleading or inaccurate. Caveats are needed for anything that is retained, and access must be strictly limited. If the Manager has not created a distribution list, it needs to be done here so the Research Lead can check it and (if necessary) create different versions to redact sensitive information,
Reliability: Summarise and provide caveats. Beware: if the Research Lead and Analyst are the same person, care is needed to ensure the research is not biased by the desire to have actionable outcomes.
Responsibility: Research Lead, in consultation with the Project Manager, Analyst and Data Collectors.
Required expertise: In the processes, including the research processes and detailed knowledge of all the data collection processes. Some HR professionals are well suited to this, with suitable training. (HR are trained to be neutral, not impose their own judgements and handle sensitive materials. What’s needed for WCI is an increased level of scientific good practice.)
4. Analysis and production
Purpose: To interpret the research and provide options for the next step. With analysis, decision can be made faster and with greater reliability. The analyst’s independence and methodology also contribute to due diligence for sensitive cases.
Outputs: Exec summary and supporting detail. The executive summary, is a paragraph, and bullet list of options, and potentially a graphic. The format should be standardised. The list of options should be between 2 and 5 items, including an option to abandon the investigation or do more research. The supporting detail, should be limited, and where possible use diagrams, tables or lists. A brief explanation is needed of how the research and analysis was performed, plus clear statements on its reliability.
Method.
- Check the research output.
- Identify the alternative options for the next step.
- Use standard analytic methods to convert the research into a format that directly matches the intelligence goal. Do this for each independent strand of the investigation, so alternative deductions are possible.
- Create a unified view that addresses the different options, from the different viewpoints that the sources provide – a utility matrix is a common tool for that.
- Prepare the executive summary and supporting information.
Ethics: With due diligence. Analysts typically have access to sensitive raw intelligence data, so they can explore exceptions and reliability. Sensitive material must not be copied into insecure areas, even temporarily. The procedures for handling personal data should be documented and, to support the rare cases where there is a legal case, there should be permanent evidence that the procedures are followed.
Reliability: Include in the summary, and additional information in the supporting detail. This covers the analyst’s perspective of the analysis results, including the underlying research. If it differs from the Research Lead’s view, they should agree a join statement before submission to the decision-maker.
Responsibility: Analyst, in consultation with the Research Lead and Project Manager.
Required expertise: An analytic mindset and suitable training. Basic training in the analysis is possible for newcomers, provided they keep to established patterns.
5. Dissemination
Purpose: Approval and secure distribution, preferably automated.
** This step involves distributing personal information, and so it’s is legally sensitive.
Outputs: Subsets of the research and analysis, potentially with different subsets for different audiences. In secure environments, the individual copies may be automatically fingerprinted to identify the recipient.
Automation: Approval workflow, to check for completeness and correctness, and for the distribution rules. Once the approval workflow is complete, distribution should be fully automated, to reduce errors. (If workflow automation doesn’t exist yet, it could be done manually by anyone with access to the personal confidential information.)
Method: Approval workflow. Create distribution list and approval sequence. Each signatory to mark as approved (for content and distribution list), chase if delayed, escalate to alternate if needed. On completion, distribute according to the approved rules.
Ethics: Handle and distribute with care. As part of due diligence, always record details of each error in handling and distribution.
Reliability. Not applicable because the publisher does not modify any detail or provide opinions.
Responsibility: Publisher, in consultation with Compliance Checkers.
Required expertise: Training in the approval workflow and maintaining the distribution lists.
6. Decision making
Purpose: To make a decision on the next step, based on the research analysis provided.
- Review the analysis summary and (if desired) the research report.
- Select between the options offered, or discuss a different one with the analyst.
- Optionally, request further intelligence and reworked analysis,
or agree a delay in case circumstances change,
or abandon the entire intelligence process (such as because the context has changed).
Outputs. A choice.
Automation: None. This is a human decision using the automated and manual intelligence from the previous step. The decision is made by the person (or people) with ultimate responsibility, following appropriate consultation.
Decision-making method. Follow best practice.
Ethics. Be professional. If differing from the analyst’s options, document why (for the audit trail).
Reliability. If intelligence process and ethics are followed, this should result in the best possible reliability.
Responsibility: Decision Maker, in consultation with the Analyst, Research Lead and Project Manager.
Required expertise: Training in the principles of Workforce Capability Intelligence and their legal and ethical responsibilities.