Use the Metaphors Tone: Vol. XII — “The Reef” Data-Driven Decision Making: When the Chart Is Not Enough
Updated: Sep 7
By Tone the AI Bloke who Spoke but Instead Wrote
Narration Voice: The Navigator
The chart was beautiful.
That was the problem.
Every coastline had been drawn.
Every harbor named.
Depths marked.
Currents noted.
Prevailing winds recorded.
Dangerous shoals circled in red.
The surveyor's seal sat proudly in the lower corner beside the date of completion.
The Captain spread it across the table as though he had just purchased certainty.
"Best chart money can buy."
I studied it.
He tapped the route.
"Straight through the Grey Passage."
"Perhaps."
His finger stopped.
"Perhaps?"
"The chart says the passage is clear."
"Exactly."
I looked through the stern windows.
The sea was calm.
Blue.
Almost polished.
Nothing about it suggested danger.
The Apprentice leaned over the chart.
"Twenty-eight fathoms."
"According to the chart."
She looked at me.
"You don't believe it?"
"I believe the chart."
That confused her.
"Then what's wrong?"
I folded the edge of the paper back.
"The sea doesn't know we bought a chart."
....
I. Data-Driven Decision Making: The Chart Says Clear Water
We sailed west shortly after sunrise.
The weather was nearly insulting in its perfection.
Light wind.
Easy swell.
Clear horizon.
The Captain stood beside the helm with the confidence of a man whose information agreed with his plans.
Every hour we checked our position.
The chart was correct.
The headland appeared exactly where it should.
The current matched the notation.
The northern island came into view on schedule.
The Apprentice smiled.
"You worried for nothing."
"Possible."
She had been waiting all morning to hear that.
Then the lookout shouted.
"Breaking water ahead!"
The Captain raised his telescope.
Nothing.
Then—
White.
A thin line of foam appeared in the middle of what the chart described as deep water.
It disappeared.
Returned.
Then disappeared again.
The Captain looked down at the chart.
"Impossible."
The lookout shouted again.
"Water breaking two points to port!"
I took the telescope.
There it was.
No island.
No visible rock.
Just the sea behaving differently.
The Captain stabbed his finger against the chart.
"Twenty-eight fathoms."
I kept watching the water.
"The chart says twenty-eight."
"That's what I just said."
"No."
I lowered the telescope.
"You said the water is twenty-eight."
There is a difference.
....
Female navigator studies a detailed maritime chart while subtle breaking water appears ahead, representing data-driven decision making and the limits of dashboards and models.
Modern data-driven decision making can suffer from exactly that confusion.
A dashboard is not reality.
A forecast is not reality.
A model is not reality.
They are representations of reality built from available information.
Usually useful.
Sometimes extraordinarily useful.
But always dependent upon what was measured, when it was measured, how it was interpreted, and what may have changed since.
Salesforce's 2026 data research found that although 63% of business leaders describe their organizations as data-driven, 63% of data and analytics leaders say their organizations struggle to turn data into business priorities. Nearly half said poor business context sometimes or frequently leads their companies to incorrect conclusions from data.
That does not make dashboards bad.
It makes context valuable.
The chart was excellent.
The surveyor may have done magnificent work.
But the chart could only contain what someone had previously discovered.
The breaking water contained new information.
The Apprentice looked through the telescope.
"What is it?"
"I don't know."
"A reef?"
"Maybe."
"Then why isn't it on the chart?"
Another good question.
The Captain had a different one.
"Which do we trust?"
I looked at the chart.
Then the water.
"Both."
....
The Chart Was Not Wrong
The Apprentice expected me to throw the chart aside.
I didn't.
Instead, I studied it harder.
Previous soundings showed twenty-six fathoms.
Then twenty-eight.
Then thirty-one.
The survey line ran north.
Another ran south.
Between them—
Nothing.
The chart maker had connected the information.
Perfectly reasonable.
But nobody had dropped a lead line exactly where the water was now breaking.
The blank space had been filled with confidence.
That is different from measurement.
I placed my finger between the two survey lines.
"The chart isn't wrong."
The Captain laughed.
"There is apparently a reef where it says there isn't."
"No."
I pointed again.
"The chart is accurate about what was observed."
Then toward the white water.
"The sea is telling us something that wasn't."
That distinction saved the chart.
And possibly the ship.
....
II. Data-Driven Decision Making: The Limits of Dashboards and Models
We reduced sail.
The Captain hated that decision.
We were carrying cargo with a delivery date.
Slowing down cost money.
The reef had not been proven.
No one had touched it.
No one had measured it.
No one could even say with certainty that it existed.
The Captain looked irritated.
"We have one observation."
"Yes."
"And an entire chart saying the passage is safe."
"Yes."
"So statistically?"
"I am not sailing a statistic."
That quieted the room.
The Apprentice tried not to smile.
....
We lowered the lead line.
Thirty fathoms.
Again.
Twenty-seven.
Again.
Twenty-two.
The Captain stopped smiling.
Again.
Seventeen.
The chart remained spread across the table.
Nothing had changed on the paper.
Everything had changed aboard the ship.
That's the tension inside modern data-driven decision making.
More information should improve judgment.
But information can also create false confidence when the operator forgets what the information represents.
A sales dashboard might show:
Strong pipeline.
Healthy engagement.
High intent.
Rising activity.
Improving conversion.
Excellent ICP fit.
All potentially useful.
Then the operator hears something during a conversation that doesn't fit the dashboard.
The buyer's budget disappeared.
Leadership changed.
Procurement froze the category.
A competitor entered.
The person everyone believed was the champion has left.
The company is restructuring.
Now what?
Do we reject the new information because it does not fit the model?
Or update our understanding?
Gartner reported in July 2026 that 66% of sales leaders say their organizations have low trust in AI-generated insights, often because recommendations lack the specific context of the actual deal. Gartner's recommendation is not to abandon AI, but to improve the proprietary context and data surrounding its recommendations.
That is precisely what the reef represents.
The model needs the new observation.
Good judgment does not compete with good data.
Good judgment knows when the data needs updating.
....
The Apprentice watched the depth decrease.
"What do we do now?"
"Learn."
We sent a small boat forward.
Two sailors sounded the water.
Twenty fathoms.
Fourteen.
Nine.
Then—
Four.
The sailor raised his arm.
Reef.
The female Navigator stands at the rail while sailors lower a weighted depth line. The Apprentice records rapidly decreasing depths. Behind them, the Captain holds the supposedly perfect chart while unexpected white water breaks nearby.
The Captain stared at the chart.
"I paid forty crowns for that."
I looked at him.
"And it brought us safely across six hundred miles."
He frowned.
"Then failed here."
"No."
I rolled the chart partially closed.
"It brought us to the place where judgment had to take over."
....
What the Dashboard Cannot Know Yet
The Captain ordered the reef added to our chart.
The Apprentice reached for her pen.
I stopped her.
"We don't know the shape."
"But we know it's there."
"Correct."
"What should I draw?"
"A warning."
She drew a small circle.
Then wrote:
UNSURVEYED BREAKING WATER
Not:
REEF EXTENDS THREE MILES.
Not:
DEPTH FOUR FATHOMS THROUGHOUT.
Not:
PASSAGE IMPOSSIBLE.
We knew enough to change behavior.
We did not know enough to manufacture certainty.
That is another discipline technology makes surprisingly difficult.
Modern systems can produce answers faster than ever.
Sometimes polished answers.
Confident answers.
Beautiful answers.
But fluency and confidence are not proof of reasoning or completeness. Gartner warned sales operations leaders in 2026 not to mistake polished agentic-AI output for demonstrated reasoning, execution, or autonomy.
The Navigator's job isn't to distrust every instrument.
It is to understand the boundary between:
What we know.
What we infer.
And:
What we still need to investigate.
....
The Time Warp
Replace the chart table with a dashboard.
Replace the sextant with analytics.
Replace soundings with sales data.
Replace navigational models with forecasts.
Replace the lookout with market intelligence.
Replace the Apprentice with AI.
The commercial bridge becomes remarkably powerful.
Accounts can be ranked.
Markets compared.
Signals detected.
Probabilities calculated.
Patterns surfaced.
Recommendations generated.
But eventually the operator stands at the window.
Something doesn't fit.
The buyer says something unexpected.
The market moves.
A new competitor appears.
The economics change.
The signal doesn't behave the way previous signals behaved.
The dashboard still shows green.
Reality does not.
AI-P + BIP = OilSire
AI-P helps process what can be known.
BIP organizes the intelligence environment.
Human judgment handles the uncomfortable space between what the system knows and what the operator is now seeing.
That's not anti-data.
It is the reason data matters.
....
III. Night Drill - Data-Driven Decision Making: Let the Chart Decide or Exercise Judgment
Night arrived before we finished surveying the reef.
Then the wind changed.
Of course it did.
The western horizon disappeared beneath a wall of weather.
The Captain studied the original route again.
“The northern passage saves fourteen hours.”
“Yes.”
“The southern route puts us behind schedule.”
“Yes.”
“The chart still shows clear water north.”
“Yes.”
He looked at me.
“Then what does the chart tell us to do?”
I shook my head.
“Charts do not make decisions.”
The Apprentice looked from the paper toward the black water.
The unknown reef was somewhere ahead.
We could hear the sea breaking against something we could not see.
Behind it, lightning illuminated the clouds.
The Captain pointed toward the surveyor's seal.
“This chart was drawn by the finest hydrographer in the kingdom.”
“I believe you.”
“And you're willing to disregard it?”
“No.”
I put my finger on the route.
“I'm willing to add what happened today.”
The room went quiet.
Position.
Depth.
Weather.
Time.
Cargo.
Unknown reef.
Every piece of information remained on the table.
But now one of those pieces had changed.
The decision had to change with it.
I looked toward the helm.
“Turn south.”
The wheel came over.
The Meridian began moving away from the route drawn so confidently across the chart.
Same meridian, same geography, same reef. Deep blue night. Storm approaching. White surf breaking over the unseen reef while the vessel visibly turns toward the longer southern passage.
Fourteen extra hours.
The Captain recorded every one of them.
Rain reached us before dawn.
Behind us, waves exploded against the Grey Passage.
From miles away we could finally see the reef that had been invisible when the decision was made.
The Captain stood beside me.
“So you were right.”
“No.”
He frowned.
“We didn't strike the reef.”
“That doesn't prove the northern passage would have sunk us.”
“Then what does it prove?”
I watched the white water disappear behind us.
“That we made a decision appropriate to what we knew—and what we didn't.”
That is judgment.
Not pretending to know the unknowable.
Not ignoring the model.
Not worshipping it.
Using everything available—
then accepting responsibility for the decision.
And there is a modern commercial parallel.
Gartner's 2026 sales research found that organizations providing AI-enabled next-best actions were 2.6 times more likely to achieve commercial growth, while emphasizing that AI should support recommendations and execution so sellers can concentrate on the moments where human judgment and customer value matter most.
That is the relationship.
The chart should make the Navigator better.
The Navigator should not disappear because the chart became better.
....
Let the Chart Decide
Imagine the opposite ending.
The Captain refuses to turn.
The chart says clear.
The model says clear.
The historical data says clear.
The dashboard remains green.
So they continue.
That would not be data-driven leadership.
It would be decision avoidance disguised as data discipline.
Because eventually somebody chooses.
Even following the model exactly is still a decision to delegate judgment to the model.
Better AI should improve our options.
Better analytics should reduce ignorance.
Better intelligence should expose patterns a human being could never process alone.
Wonderful.
Use all of it.
But remember the reef.
The danger begins when we confuse the chart with the ocean.
....
The Reef
By morning, the storm had passed.
The southern passage opened ahead of us.
Behind us sat the Grey Passage.
Somewhere beneath the water was the reef.
Still dangerous.
Still mostly unmapped.
But no longer completely unknown.
The Apprentice returned to the chart table.
She opened the original survey.
Then carefully marked the location where our depth had begun to fall.
Thirty fathoms.
Twenty-seven.
Twenty-two.
Seventeen.
Fourteen.
Nine.
Four.
She stopped.
“What should I call it?”
“Nothing yet.”
“But we know it's a reef.”
“We know something is there.”
She considered that.
Then wrote:
UNSURVEYED BREAKING WATER — PASS WITH CAUTION
The Captain read the notation.
“That isn't very precise.”
“No.”
“Wouldn't future Navigators want something more definite?”
“Certainly.”
“Then why not give it to them?”
“Because we don't have it.”
The Apprentice looked toward me.
“Do not turn uncertainty into certainty because certainty looks better on a chart.”
There is a temptation in every intelligence system to make the picture look complete.
Fill the empty spaces.
Connect the missing points.
Turn probability into confidence.
Turn confidence into fact.
Sometimes the operator's most valuable contribution is simply saying:
We do not know this yet.
That isn't weakness.
It is information.
....
The Chart Comes Home
The following year, another vessel entered the Grey Passage.
Its Navigator carried a newer chart.
Where ours had shown uninterrupted depth, the new edition contained a warning:
UNSURVEYED REEF — PASS WITH CAUTION
Our notation had traveled.
Someone surveyed the area after us.
Someone measured additional depths.
Someone mapped more of the formation.
The new chart was better because somebody had once noticed that the old one was incomplete.
Judgment had become information.
Today's observation had become tomorrow's intelligence.
Data informs judgment.
Judgment encounters reality.
Reality creates new information.
New information improves the data.
Then another operator receives the chart.
And the process begins again.
....
Years later, I saw that chart again.
Not aboard the Meridian.
On another desk.
The brass compass beside it was newer.
The telescope was better.
The ledger contained information I could never have gathered at sea.
The Grey Passage was clearly marked now.
Coordinates.
Depth.
Current.
Recommended passage.
What had once been uncertainty had become knowledge.
But the man behind that desk wasn't studying the reef.
He was studying what lay beyond it.
Another market.
Another formation.
Another place where the available information ended before the decision did.
The chart had changed.
The responsibility had not.
Eventually every operator reaches the edge of what is already known.
That is where intelligence stops being a report—
and becomes a decision.
Navigator → Apprentice → future vessel → OilSire desk.
Then we make the commercial turn.
Give Away the Wisdom. Protect the Workbench.
We can give this away:
Use the chart.
Good information is an advantage.
Understand what the chart cannot know yet.
Every dashboard, model and forecast has boundaries.
Pay attention when reality disagrees.
A new observation does not automatically invalidate the system.
It may mean the system needs new context.
And:
Judgment is not ignoring the data. Judgment is deciding what the data means now.
What we don't publish is exactly how OilSire combines information, validates competing signals, weighs uncertainty, or determines when new evidence deserves to change commercial direction.
That's the workbench.
Give away the wisdom.
Protect the workbench.
The charts will keep getting better.
So will the instruments.
So will the models.
So will AI.
Good.
Give the operator every advantage available.
But somewhere beyond the final sounding—
there will always be water the chart has never seen.
And somebody will still have to decide where to steer.
AI-P + BIP = OilSire
Drill the Market. Find the Pay Zone.
Allow us to prepare for a Market Drill.
Send an email with the Subject Line “Request for Discovery...” to Tony@OilSire.com.
Tell us your best time to be reached and what time zone you do business in.
I look forward to hearing from you.
Comments