The deep-dive behind Capability Intelligence. How we measure, preserve and build the three things capability is made of, from evidence rather than titles, CVs or memory.
Firms have CV databases and skill tags, yet still fall back on interviews, keyword matching and memory when forming a squad.
Clients rarely express the real outcomes, responsibilities, context and critical capabilities the work demands.
A profile omits relevant, transferable evidence, and uses different language to the brief.
Matching people one by one doesn't prove the squad covers the work or actually works together.
Capability data detaches from delivery, so it ages and stops reflecting what people can really do.
Skills is more than CV analysis. It is framing the problem around the capabilities needed to solve it, matching people to that brief on evidence, and growing them into the squad member the work needs.
Guide a client beyond the job title: what outcome must this achieve? What decisions and risk will the squad own? Which capabilities are essential, and what evidence would prove them?
Output: an evidence-ready capability brief.
Guide a person beyond the CV: what situations have they handled? What did they decide? What changed because of their work? Which capabilities transfer?
Output: an evidence-backed capability profile.
Match people to the brief where demonstrated capability meets the requirement, surface coverage, gaps and key-person risk before the squad is committed, and set a growth path.
Output: a defensible squad, with visible strengths and gaps.
An illustrative walkthrough on synthetic data. SFIA codes work invisibly; you see plain-English capabilities. Steps noted on the roadmap are not yet built.
The original request
"We need a team to modernise our regulated payments journey: cloud-native, compliant, in about six months."
Clarified outcomes & critical capabilities
Solution architecture, software development, systems integration, information security, risk management and data engineering, each with the outcomes and evidence that would prove it.
Evidence discovered beyond CVs
For example, "Led the architecture for a cloud payments migration, 8-person team" maps to Solution architecture with the evidence attached, validated against the framework. Invented or out-of-range skills are flagged, not stored.
Coverage, gaps & key-person risk
Candidates matched where demonstrated capability meets the requirement, with thinly-covered capabilities and single points of dependency surfaced before you commit the team.
Collaboration-optimised composition & growth plan (on the roadmap)
Preferring people with a track record of delivering together, and a growth plan that closes gaps, are on the roadmap. Today the prototype surfaces matches, gaps and collaborator links for a human to compose and decide.
SFIA, the global Skills Framework for the Information Age, provides a recognised structure for skills and levels of responsibility, so capability is explicit and portable rather than personality-dependent. You see clear human language; the codes work invisibly. DTA's use is properly licensed, and Sailesh Panchal is a contributor to the SFIA framework. SFIA® is a registered trademark of the SFIA Foundation.
Two people with the same skills perform very differently if one understands the domain, the policy constraints, the architecture and the delivery history, and the other does not. Knowledge is the context that makes a skill usable.
Product and domain understanding, regulatory and control context, prior decisions and delivery history, architectural dependencies, and the institutional memory worth preserving through change.
Weak practice leaves firms dependent on inboxes, shared drives and long-tenured staff. Stronger practice makes concepts, rules, events, evidence and relationships governed and usable, so people and AI reason over trusted context, not fragments.
Capability should not disappear when people move roles or a programme ends. Preserving and structuring institutional context is how a squad acts with shared understanding, and how a bank keeps what it knows through transformation.
Human-in-the-loop only works when the human can do more than approve. They must understand what the system did, hold the authority to intervene, and have the judgment to act well when the brief is incomplete.
Which decisions can be automated, and which still need a person: the trade-offs, the edge cases, and the moments an agent should hand back to a human.
The evidence that shows a person can exercise judgment well, drawn from the decisions they have actually made, never a score inferred without proof.
Judgment develops through work, coaching, escalation and experience. We help define where it matters and how to grow it deliberately, not reduce it to a rating.
Three offers, in sequence, without forcing you to buy all three at once.
Translate a proposition, role or live brief into required outcomes, the human and AI-assisted capabilities, evidence criteria, current gaps and delivery risks, and squad-formation recommendations. A work-product within the six-week Proposition Blueprint.
Discover evidence beyond CVs, find complementary strengths, expose gaps and key-person dependencies, and assemble a defensible squad with a client-ready rationale.
Grow the squad through evidence from real delivery, targeted assignments, coaching, reassessment and internal mobility. Evidence-led, not a ratings system or a course catalogue.
AI assists the work, with human oversight, consent, policy controls and an evidence trail. People remain accountable, and sensitive information can be processed on device or within a controlled environment.
Worth bringing: the decision you're trying to make, your current workaround, where it fails, and whether the problem is framing the brief, finding the evidence, forming the squad or growing it.
Discuss a Capability Blueprint