Skip to content

Commit 4780b93

Browse files
committed
feat: add strategic tier to measurement, cost case + 2 failure scenarios to barriers, financial framing to decisions
- measurement.html: new Tier 1 (Strategic Intelligence) with 4 KPIs leading before operational tiers; existing tiers renumbered 02-05; intro copy rewritten to lead with decision quality - barriers.html: two new failure scenarios (D: LLM confidence inflation, E: governance bypass under time pressure); new cost case section (§9 from whitepaper) with 3 cards covering decision risk, silent intelligence loss, and AI ROI destruction; BCG 2024 added to citations - decisions.html: financial consequence framing added to all 5 cases; intro copy sharpened to name the cost of inaction - site.css: .case-fin style added for financial consequence blocks; .kpi-tier--strategic colour token added
1 parent 29e2330 commit 4780b93

4 files changed

Lines changed: 196 additions & 8 deletions

File tree

assets/css/site.css

Lines changed: 17 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -3263,6 +3263,19 @@ p:last-child {
32633263
line-height: 1.5;
32643264
}
32653265

3266+
.case-fin {
3267+
margin-top: 14px;
3268+
padding: 11px 14px;
3269+
background: rgba(0, 0, 0, 0.18);
3270+
border-radius: 8px;
3271+
font-size: 13px;
3272+
color: var(--text2);
3273+
border: 1px solid rgba(255, 255, 255, 0.05);
3274+
border-left: 2px solid rgba(255, 255, 255, 0.15);
3275+
line-height: 1.58;
3276+
font-style: italic;
3277+
}
3278+
32663279
/* KPI TIERS */
32673280
.kpi-tiers {
32683281
margin: 44px 0;
@@ -3285,6 +3298,10 @@ p:last-child {
32853298
box-shadow: 0 16px 32px rgba(0, 0, 0, 0.28);
32863299
}
32873300

3301+
.kpi-tier--strategic {
3302+
--tier-accent: #22b9dc;
3303+
}
3304+
32883305
.kpi-tier--ops {
32893306
--tier-accent: var(--success);
32903307
}

pages/barriers.html

Lines changed: 99 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -199,7 +199,7 @@ <h3>AI capability is necessary but not sufficient</h3>
199199
failures, not technology failures.</div>
200200
</div>
201201

202-
<h3 class="section-heading-tight-44">Three failure scenarios to actively govern</h3>
202+
<h3 class="section-heading-tight-44">Five failure scenarios to actively govern</h3>
203203
<div class="cases-grid cases-grid--failures fade-in">
204204
<div class="case-item case-item--risk">
205205
<div class="case-header">
@@ -243,6 +243,100 @@ <h3>Low contribution because value is not credited</h3>
243243
coverage and reuse impact.</div>
244244
</div>
245245
</div>
246+
<div class="case-item case-item--risk">
247+
<div class="case-header">
248+
<div class="cn">D</div>
249+
<div class="ctag">AI Output</div>
250+
<div class="case-icon"><i data-lucide="bot" class="lucide" aria-hidden="true"></i></div>
251+
</div>
252+
<div class="case-body">
253+
<h3>LLM confidence inflation masking low-quality sources</h3>
254+
<p>LLM synthesis outputs are fluent regardless of underlying knowledge quality. High-confidence-sounding
255+
language in a generated output does not indicate that the source assets warranted that confidence. Without
256+
explicit propagation of Knowledge Health Scores into generated outputs, users cannot distinguish
257+
well-grounded intelligence from well-phrased inference &#8212; and will act on it as if they are the same.</p>
258+
<div class="case-out">Mitigation owner: Platform and Knowledge Engineering leads. KHS scores must be
259+
surface-visible in every generated output. Confidence level is not optional metadata &#8212; it is part of
260+
the output.</div>
261+
</div>
262+
</div>
263+
<div class="case-item case-item--governance">
264+
<div class="case-header">
265+
<div class="cn">E</div>
266+
<div class="ctag">Adoption</div>
267+
<div class="case-icon"><i data-lucide="git-branch" class="lucide" aria-hidden="true"></i></div>
268+
</div>
269+
<div class="case-body">
270+
<h3>Governance bypass under time pressure</h3>
271+
<p>Decision-makers under time pressure will route around the governed intelligence layer entirely, using
272+
ungoverned search or direct AI tools. Governance thresholds only constrain what passes through the governed
273+
pipeline &#8212; they do not prevent parallel ungoverned use. If the governed channel is slower or less
274+
convenient, it will be abandoned in the moments that matter most.</p>
275+
<div class="case-out">Mitigation owner: Business unit leadership. Intelligence Cycle Time (Tier 1) must be
276+
competitive with ungoverned alternatives. A governed channel that is trusted but slow will lose to one that
277+
is fast but ungoverned.</div>
278+
</div>
279+
</div>
280+
</div>
281+
</div>
282+
283+
<!-- COST CASE -->
284+
<div class="sw" id="cost-case">
285+
<div class="section-tag">The Cost Case</div>
286+
<h2>The financial consequences of not building this capability.</h2>
287+
<p>The case for Knowledge Intelligence is usually made as a value argument. It is equally a cost argument &#8212;
288+
and the cost argument is, in most organisations, the larger of the two.</p>
289+
290+
<div class="cases-grid cases-grid--barriers fade-in">
291+
<div class="case-item case-item--risk">
292+
<div class="case-header">
293+
<div class="cn">01</div>
294+
<div class="ctag">Decision Risk</div>
295+
<div class="case-icon"><i data-lucide="triangle-alert" class="lucide" aria-hidden="true"></i></div>
296+
</div>
297+
<div class="case-body">
298+
<h3>The cost of one major decision made on low-confidence knowledge</h3>
299+
<p>Not a failed project &#8212; a decision made on knowledge that looked reliable but was not. Acquisitions
300+
that destroyed value. Market entries built on assumptions that had been superseded. Platform investments
301+
justified by internal consensus rather than evidence. The financial exposure from one such decision is, in
302+
most large organisations, larger than the entire cost of building a Knowledge Intelligence capability.</p>
303+
<div class="case-out">The question is not whether to invest in KI. It is whether the next high-stakes
304+
decision will be the one that proves the cost of not having invested.</div>
305+
</div>
306+
</div>
307+
<div class="case-item case-item--warning">
308+
<div class="case-header">
309+
<div class="cn">02</div>
310+
<div class="ctag">Silent Loss</div>
311+
<div class="case-icon"><i data-lucide="trending-down" class="lucide" aria-hidden="true"></i></div>
312+
</div>
313+
<div class="case-body">
314+
<h3>The accumulated cost of losing intelligence without knowing it</h3>
315+
<p>Senior departures where no transfer governance existed. Project closures where lessons were filed and
316+
forgotten. Strategic frameworks maintained past their useful life because no one measured their actual
317+
performance. These are not edge cases. They are the default operating condition of organisations without an
318+
intelligence layer. The cost is largely invisible &#8212; which is precisely why it is rarely challenged.</p>
319+
<div class="case-out">Invisible cost is not zero cost. It is cost that compounds without being measured,
320+
challenged, or stopped.</div>
321+
</div>
322+
</div>
323+
<div class="case-item case-item--risk">
324+
<div class="case-header">
325+
<div class="cn">03</div>
326+
<div class="ctag">AI ROI</div>
327+
<div class="case-icon"><i data-lucide="cpu" class="lucide" aria-hidden="true"></i></div>
328+
</div>
329+
<div class="case-body">
330+
<h3>What AI investment produces without a reliable knowledge foundation</h3>
331+
<p>74% of organisations report struggling to achieve and scale AI value despite widespread adoption.<span
332+
class="cite">BCG, 2024</span> The constraint is not the model &#8212; it is the quality of the knowledge
333+
beneath it. Every pound spent on AI capability without a corresponding investment in knowledge foundations
334+
is a pound spent on processing speed for an intelligence engine that cannot be trusted. More AI without
335+
better knowledge governance makes the problem larger, not smaller.</p>
336+
<div class="case-out">AI investment without knowledge governance is not a technology problem. It is a
337+
compounding cost problem dressed as a capability programme.</div>
338+
</div>
339+
</div>
246340
</div>
247341
</div>
248342

@@ -313,6 +407,10 @@ <h3>Sources and References</h3>
313407
<div class="citation-text"><em>McKinsey and Company</em> - The State of AI in 2025: Agents, Innovation, and
314408
Transformation. McKinsey Global Survey, 2025.</div>
315409
</div>
410+
<div class="citation-item">
411+
<div class="citation-num">9</div>
412+
<div class="citation-text"><em>Boston Consulting Group</em> - Why AI Transformations Fail. BCG, 2024.</div>
413+
</div>
316414
</div>
317415
</div>
318416
</main>

pages/decisions.html

Lines changed: 9 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -41,8 +41,10 @@
4141
<div class="sw" id="decisions">
4242
<div class="section-tag">Decision Cases</div>
4343
<h1>Five decisions that Knowledge Intelligence makes newly possible.</h1>
44-
<p>The value of Knowledge Intelligence becomes concrete when anchored to decisions that were previously unavailable,
45-
or available only through costly, one-off, and unreliable manual effort.</p>
44+
<p>The value of Knowledge Intelligence becomes concrete when anchored to specific decisions that organisations
45+
currently cannot make with confidence — and to the strategic and financial consequences of not being able to make
46+
them. Each case below names the question, the value of answering it earlier, and what it costs to leave it
47+
unanswered.</p>
4648

4749
<div class="cases-grid cases-grid--decisions fade-in">
4850

@@ -58,6 +60,7 @@ <h3>Detecting emerging themes before they are named</h3>
5860
visibility of persistent, converging signals across both the documented content estate and tacit knowledge
5961
networks, allowing leaders to see change forming before it becomes obvious - without being asked to react to
6062
noise.</p>
63+
<div class="case-fin">Organisations that act 6–12 months ahead of a strategic shift establish positions that are difficult to reverse. The intelligence value is timing: the difference between leading a change and responding to it.</div>
6164
<div class="case-out">Earlier, better-timed strategic awareness without overreaction.</div>
6265
</div>
6366
</div>
@@ -76,6 +79,7 @@ <h3>Measuring concentration risk in the knowledge estate</h3>
7679
with them when they leave.<span class="cite">KMWorld, 2024</span> Knowledge Intelligence makes this risk
7780
visible and measurable before it becomes a crisis - treating knowledge concentration the same way a finance
7881
team treats exposure in a portfolio.</p>
82+
<div class="case-fin">Replacing a senior knowledge holder costs between 1.5× and 2× their annual salary — and that figure does not account for the intelligence that cannot be recovered at any price. KI makes concentration risk visible before it becomes a loss event.</div>
7983
<div class="case-out">Knowledge risk becomes a manageable condition, not an invisible one.</div>
8084
</div>
8185
</div>
@@ -92,6 +96,7 @@ <h3>Evidence-based evolution of strategic concepts</h3>
9296
Intelligence provides visibility into how concepts perform across the full knowledge estate, in documents,
9397
in practice, and in expert judgement, enabling deliberate and evidenced evolution rather than conceptual
9498
drift.</p>
99+
<div class="case-fin">The cost is not just intellectual: it is the compounded investment in training, tooling, and consulting directed at frameworks whose performance has never been measured. KI provides evidence-based visibility into concept performance — enabling deliberate retirement rather than drift.</div>
95100
<div class="case-out">Concepts change based on evidence, not politics or fatigue.</div>
96101
</div>
97102
</div>
@@ -109,6 +114,7 @@ <h3>Governing confidence across the knowledge estate</h3>
109114
class="cite">Dataversity / Gartner, 2025</span> Knowledge Intelligence makes confidence a first-class
110115
governance dimension, ensuring the knowledge underpinning high-stakes decisions can be interrogated, not
111116
assumed.</p>
117+
<div class="case-fin">For high-stakes decisions — acquisitions, market entries, major platform investments — the financial consequence of acting on low-confidence knowledge can be measured in tens of millions. KI makes confidence a first-class intelligence dimension, so it can be interrogated before the decision is made.</div>
112118
<div class="case-out">Confidence becomes governable. Assumptions become visible.</div>
113119
</div>
114120
</div>
@@ -125,6 +131,7 @@ <h3>Deliberate alignment between internal knowledge and external context</h3>
125131
treats external context as weighted evidence, allowing internal understanding to be reinforced, challenged,
126132
or updated deliberately rather than by crisis. This applies to documented knowledge and to the assumptions
127133
embedded in expert practice.</p>
134+
<div class="case-fin">The intelligence output is not "this knowledge is outdated" — it is "the assumptions underlying this strategic position are no longer supported by current evidence." That is the difference between governance reporting and strategic intelligence.</div>
128135
<div class="case-out">Strategic choices grounded in real alignment, not assumption.</div>
129136
</div>
130137
</div>

pages/measurement.html

Lines changed: 71 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -46,12 +46,77 @@ <h1>How you know it's working.</h1>
4646
to recognise success. The organisations that have deployed this capability at scale share a common discipline:
4747
they defined what they were trying to move, established a baseline before they began, and tracked a small number
4848
of meaningful metrics rather than a large number of available ones.</p>
49-
<p>The framework below organises measurement into four tiers. Together they show quality, operational effect,
50-
financial efficiency, and estate health. Each metric is grounded in operational signals rather than abstract
51-
sentiment.</p>
49+
<p>Measurement must be anchored to what the capability exists to do: generate intelligence that improves the
50+
quality and confidence of strategic and financial decisions. Operational metrics matter but they are not the
51+
primary signal. A programme that improves search speed while failing to generate decision-grade intelligence is
52+
not delivering on the capability&#8217;s purpose.</p>
53+
<p>The framework organises measurement into five tiers, starting with strategic intelligence indicators. Together
54+
they show decision quality, knowledge quality, operational effect, financial efficiency, and estate health.</p>
5255

5356
<div class="kpi-tiers fade-in">
5457

58+
<div class="kpi-tier kpi-tier--strategic">
59+
<div class="kpi-tier-header">
60+
<div class="kpi-tier-num">01</div>
61+
<div>
62+
<div class="kpi-tier-name">Strategic Intelligence</div>
63+
<div class="kpi-tier-q">Is it generating intelligence that improves the quality of decisions?</div>
64+
</div>
65+
</div>
66+
<div class="kpi-grid">
67+
<div class="kpi-card">
68+
<div class="kpi-name">Decision Confidence Rate</div>
69+
<div class="kpi-desc">The proportion of material decisions made with confidence-weighted intelligence versus
70+
unverified assumption. The primary indicator that the capability is doing what it exists to do.</div>
71+
<div class="kpi-signal kpi-signal--good">
72+
<div class="kpi-signal-type">Good Signal</div>
73+
<div class="kpi-signal-bars"><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div></div><div class="kpi-signal-label">Financial significance</div>
74+
<div class="kpi-signal-val">Proxy for reduction in high-stakes decision risk. A single major strategic
75+
decision made on low-confidence knowledge typically costs more than the entire investment in building
76+
the capability.</div>
77+
</div>
78+
</div>
79+
<div class="kpi-card">
80+
<div class="kpi-name">Intelligence Cycle Time</div>
81+
<div class="kpi-desc">Elapsed time from a strategic question being asked to a confidence-weighted answer
82+
being accepted and acted upon. Tracks whether the governed channel is competitive with ungoverned
83+
alternatives.</div>
84+
<div class="kpi-signal kpi-signal--warn">
85+
<div class="kpi-signal-type">Warning</div>
86+
<div class="kpi-signal-bars"><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div></div><div class="kpi-signal-label">The adoption risk</div>
87+
<div class="kpi-signal-val">If the governed channel is slower than ungoverned search or direct AI use,
88+
decision-makers will route around it under time pressure. The governed channel must be faster, not just
89+
more trustworthy.</div>
90+
</div>
91+
</div>
92+
<div class="kpi-card">
93+
<div class="kpi-name">Concentration Risk Index</div>
94+
<div class="kpi-desc">The proportion of strategic intelligence domains where fewer than three people or
95+
repositories hold the relevant knowledge. Tracks fragility in the intelligence pipeline before it becomes
96+
a loss event.</div>
97+
<div class="kpi-signal kpi-signal--risk">
98+
<div class="kpi-signal-type">At Risk</div>
99+
<div class="kpi-signal-bars"><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div></div><div class="kpi-signal-label">Expected loss exposure</div>
100+
<div class="kpi-signal-val">Concentration risk is invisible until it crystallises &#8212; a departure, a
101+
reorg, a health event. Replacing a senior knowledge holder costs 1.5&#8211;2&times; annual salary, not
102+
counting unrecoverable intelligence.</div>
103+
</div>
104+
</div>
105+
<div class="kpi-card">
106+
<div class="kpi-name">Emerging Signal Lead Time</div>
107+
<div class="kpi-desc">Average time between KI detecting a convergent pattern across the knowledge estate
108+
and that pattern becoming mainstream-recognised. Tracks strategic positioning value.</div>
109+
<div class="kpi-signal kpi-signal--good">
110+
<div class="kpi-signal-type">Good Signal</div>
111+
<div class="kpi-signal-bars"><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div><div class="kpi-signal-bar"></div></div><div class="kpi-signal-label">Strategic positioning value</div>
112+
<div class="kpi-signal-val">Organisations acting 6&#8211;12 months ahead of a named shift establish
113+
positions that are difficult to reverse. Lead time is the measurable output of signal detection working
114+
as intended.</div>
115+
</div>
116+
</div>
117+
</div>
118+
</div>
119+
55120
<div class="kpi-tier kpi-tier--quality">
56121
<div class="kpi-tier-header">
57122
<div class="kpi-tier-num">01</div>
@@ -285,8 +350,9 @@ <h1>How you know it's working.</h1>
285350
<div class="kpi-principle-head">The measurement principle</div>
286351
<p>No single metric tells the full story. Retrieval success without freshness tracking means you are measuring
287352
how well people find information that may be wrong. Freshness SLA without reuse rate means you are maintaining
288-
a library no one is reading. The four tiers work together: knowledge quality sets the standard, operational
289-
impact proves workflow value, financial efficiency proves cost impact, and estate health protects long-term
353+
a library no one is reading. The five tiers work together: strategic intelligence indicators confirm the
354+
capability is changing what leaders decide; knowledge quality sets the standard for the estate; operational
355+
impact proves workflow value; financial efficiency proves cost impact; and estate health protects long-term
290356
reliability.</p>
291357
<p>Start with three or four metrics that cover all tiers, and establish baselines before deployment. The most
292358
common error is attempting to measure everything at once and measuring nothing with rigour.</p>

0 commit comments

Comments
 (0)