Communication & Influence

From Scarcity to Filtering

The problem flipped

For most of professional history, producing communication was the hard part. Writing a solid memo took hours. Building a deck took days. Scarcity of output was the constraint, so effort was a reliable signal: if someone sent you a polished document, someone had thought hard about it.

AI removed the production constraint. Anyone can now generate a polished-looking memo, summary, or deck in minutes. The volume of professional communication is rising while the effort behind any given piece of it has become unknowable. Polish no longer proves thinking.

That flip changes the core skill. When production was scarce, the valuable skill was producing. Now that production is abundant, the valuable skill is filtering: deciding what deserves attention, and making your own messages worth the attention you are asking for.

How attention actually gets allocated

Before the evidence, a short model of how readers behave, because the framework in Module 2 is built on it.

Professional readers run a fast triage on every incoming message, usually in under twenty seconds. The triage answers three questions in order: Is this for me? Does it require anything of me? Is there a cost to ignoring it? A message that answers all three in its first lines gets read. A message that answers none gets deferred, and deferred usually means never.

This behavior predates AI. What AI changed is the base rate. When most messages were written by a person who spent time on them, readers extended the benefit of the doubt. When a growing share of messages can be produced without thought, readers withdraw it, and the triage becomes stricter. Your messages now compete against a higher volume of polished text for a reader who has learned that polish is uninformative.

The evidence: noise has a price tag

This is measurable, not a vibe.

The sender side: workslop. Research from BetterUp Labs and the Stanford Social Media Lab, based on a survey of 1,150 full-time U.S. desk workers in September 2025, coined a name for the incoming problem: workslop, AI-generated work that looks polished but lacks the substance to move a task forward. The findings:

• About 40 percent of workers reported receiving it within the past month.
• Those who received it estimated that roughly 15 percent of everything crossing their desk qualified.
• Each instance took close to two hours to sort out. The researchers priced that at about $186 per employee per month, or over $9 million a year for a 10,000-person organization.
• It flows in every direction: mostly between peers, but a meaningful share travels upward to managers and downward from leaders to teams.
• The social cost ran deeper than the hours. Over half of recipients reported being annoyed, and about a fifth reported being offended. Roughly half viewed the senders as less capable and less reliable, and 42 percent said they trusted those colleagues less.

Read the last point carefully. Sending noise does not just waste the reader's time. It spends the sender's credibility, and credibility is the account every future message draws on.

The receiver side: tuning out. The 2026 State of Workplace Communication Survey from Korbyt and Reworked, based on 1,175 full-time U.S. employees at organizations with 50 or more staff, looked at the same problem from the reader's chair:

• 81 percent of workers say they can usually tell when a message was created with AI.
• About 45 percent question a message's accuracy when they suspect AI was involved.
• Roughly half say the volume of communication they receive is "about right," yet 44 percent still report feeling overwhelmed and tuning out. The researchers read this as passive disengagement: people are no longer absorbing more, they are quietly filtering.
• Only a small share are confident they are not missing important updates; most are only moderately confident.
• What makes people pay attention: 56 percent named a clear action, and 59 percent said their confidence in a message depends on how clear and specific it is.
• What makes them tune out: 46 percent cited messages repeated across channels.
• When asked what they want AI to do, 92 percent agreed it should reduce information overload rather than add to it.

Read those numbers together and the picture is clear. Your audience is already filtering, already detecting generic output, already discounting it, and already telling researchers exactly what earns their attention: clarity, specificity, and a reason to act. The question is which side of their filter your messages land on.

The trust ledger

A useful way to hold both studies at once: every message you send makes a deposit or a withdrawal in a trust ledger the reader keeps about you.

• A message that was clearly thought through, that told them something they needed, and that respected their time: deposit.
• A message that cost them time and gave them nothing, regardless of how polished it looked: withdrawal.
• A message they later discovered was AI-generated without disclosure, on a topic where it mattered: large withdrawal, because it retroactively discounts earlier deposits.

The ledger explains a pattern many leaders notice without naming: their messages start getting slower responses and vaguer replies over months, with no single incident to point to. The balance ran down one withdrawal at a time.

Why this raises the value of your voice

Here is the counterintuitive part: an AI-saturated environment makes distinctly human communication more valuable, because it is scarcer relative to everything else in the inbox.

A message with a clear point of view, specific facts, and a recognizable voice now stands out the way a handwritten note stands out in a stack of form letters. The skill this course builds is deliberate scarcity: putting the thinking into your messages that the generic flood conspicuously lacks, and using AI in ways that amplify your thinking rather than replace it.

A note on where AI belongs

None of this is an argument against AI in communication. The same research points at a clear division of labor. Workers want AI to summarize, prioritize, shorten, and filter. They penalize AI used to multiply output and to substitute for the sender's own thinking. Module 4 builds a workflow on exactly that line.

Check your understanding

1. Why did effort stop being a reliable signal of quality?
2. What are the two costs of sending noise, per the workslop research, and which one compounds?
3. What did readers in the Korbyt and Reworked survey say earns their attention, and what makes them tune out?
4. Why does communication abundance increase the value of a distinct voice rather than drowning it?

The Signal Test

A framework that works in both directions

You need one framework that evaluates communication whether you wrote it, an AI wrote it, or a colleague sent it. The Signal Test has four parts. A message that passes all four is worth someone's time. Each failed part tells you exactly what to fix.

The test draws on ideas that predate AI by decades. Military and government writers have long used "bottom line up front." Consulting has the pyramid principle: answer first, supporting arguments below. Philosophers of science gave us falsifiability as the mark of a claim that carries information. The Signal Test packages those ideas into four checks you can run in twenty seconds.

Part 1: Stakes

The question: if the reader skips this, does anything change for them?

Signal names its consequence. It tells the reader what this affects, what they need to decide, or what happens next. Noise circulates because circulating is easy: FYI messages with no F worth Y-ing, updates that update nothing, summaries of things nobody asked about.

Noise version:

"Sharing an update on the vendor evaluation process. We continue to make progress across multiple workstreams and will keep the group posted as things develop."

Signal version:

"Vendor decision is due Friday. We are down to two options, and the open question is whether we accept a 6-week implementation delay for a 15 percent lower cost. I need your call on that trade-off by Thursday."

Nothing changes for the reader of the first message. Everything in the second exists because the reader matters to the outcome.

Sub-checks for stakes:

• Is there a decision, action, or consequence in the first four lines?
• Is it clear who it applies to? A message to twelve people with no named owner has stakes for no one.
• Is there a date? Stakes without a deadline are aspirations.
• If the reader did nothing, would anything happen? If the honest answer is no, the message may not need sending.

The edge case: some messages are legitimately informational, with no action required. They still have stakes if they change what the reader knows in a way that matters later ("no action needed, but this changes the Q2 forecast you are building"). Say so explicitly. "No action needed; this affects X" is a complete stakes statement.

Part 2: Specificity

The question: could someone verify or disagree with this?

Signal contains claims concrete enough to be checked and, in principle, wrong. Names, numbers, dates, positions. Noise is unfalsifiable: it hedges every claim into mist, and it would read the same in any company, any quarter, any industry. Unfalsifiable content is one of the most reliable markers of low-effort AI output, because generating safely vague text is what an unbriefed model does best.

Noise version:

"The initiative is well positioned to drive meaningful improvements in operational efficiency across the organization."

Signal version:

"The new intake form cut duplicate data entry for admissions staff from 4 systems to 1. Early result: about 20 minutes saved per admission, based on the first 30 admissions at two communities."

You can argue with the second version. That is the point. Arguable claims are the ones that carry information.

Sub-checks for specificity:

• Count the checkable facts. A message of two hundred words with zero numbers, names, or dates is almost certainly noise.
• Look for the sample size. "Early result" with "based on the first 30" is honest specificity. "Results show" with no denominator is a claim wearing a lab coat.
• Watch for importance words standing in for evidence: significant, meaningful, robust, substantial, impactful. Each one should be followed by a number or an example, or deleted.
• Check whether the specifics are relevant. Precision about the wrong thing (the exact date a meeting was held, rather than what it decided) is a form of noise.

The edge case: confidentiality sometimes prevents full specificity. The honest move is to say so and give the reader what you can: "I cannot share the vendor's number, but the difference is in the range of 15 to 20 percent, and I can walk you through it in person."

Part 3: Stance

The question: do I know what the sender thinks and wants?

Signal takes a position. It recommends, decides, or asks for something specific, and the sender absorbs the risk of being wrong. Noise is symmetrical: it lists considerations on the one hand and the other hand, ends at "it depends," and transfers all the thinking to the reader. Balance reads as thoroughness but often functions as hiding.

Noise version:

"There are several factors to weigh. Option A offers cost advantages while Option B provides greater flexibility. Both approaches have merit depending on organizational priorities."

Signal version:

"Recommend Option A. The cost difference funds the training budget we cut in March, and the flexibility Option B offers solves a problem we do not currently have. The main risk is vendor lock-in, and I think the 3-year contract term makes it acceptable."

A reader can act on the second version in one reply. The first version schedules a meeting.

Sub-checks for stance:

• Is there a sentence starting with "I recommend," "I decided," "I need," or an equivalent?
• Is the reason attached? A recommendation without a reason is an instruction, and readers push back on instructions.
• Are the risks named and answered, rather than listed and abandoned? "The main risk is X, and here is why I accept it" is stance. "There are risks to consider" is not.
• If the sender genuinely cannot recommend yet, do they say what would resolve it? "I cannot recommend until I see the Q3 numbers on Friday; I will send my call by Monday" is a stance about the process.

The edge case: some messages legitimately present options for someone else's decision. Even then, stance is available: your assessment of each option, your default if they do not respond, and what you would do in their position.

Part 4: Signature

The question: could only this person, in this situation, have sent this?

Signal sounds like a specific person who knows the reader, shares history with them, and writes the way they talk. Noise is generic in register: it could have come from anyone, addressed to anyone. Since 81 percent of your audience believes they can detect AI-generated messages, generic register does not read as neutral. It reads as "the sender did not consider this worth their own time," and the message gets discounted before its content is even weighed.

Signature lives in small things: referencing the conversation you actually had last week, using the shorthand your team uses, acknowledging the objection this specific reader raised last time, sounding the way you sound.

Noise version:

"Thank you for your continued partnership on this initiative. I look forward to collaborating further as we move into the next phase."

Signal version:

"You flagged the training dates as the risk on 8/20 and you were right. Here is what I want to do about it."

Sub-checks for signature:

• Is there at least one reference to shared history, a prior conversation, or something only this reader would recognize?
• Does the register match the relationship? A formal message to a close colleague, or a casual one to a board committee, both read as misjudged.
• Would the sender say these sentences aloud to this person? If not, the sentences are register, not voice.
• Could you swap the recipient's name for anyone else's and send it unchanged? If yes, signature is missing.

The edge case: broadcast messages to large groups cannot be individually signed. They can still carry signature at the group level: the sender's recognizable voice, a reference to what this group has been through, an acknowledgment of the specific cost to them.

Scoring and using the test

Score any message 0 to 4, one point per part. The test works three ways:

• As a reader: a message scoring 0 or 1 gets minimal attention, whatever its polish.
• As a writer: any part you fail is your revision instruction. No stakes? State the consequence. No stance? Commit to a recommendation.
• As an editor of AI drafts: the test is your quality gate, which is Module 4's whole subject.

A note on what the test does not measure: length, formatting, or how impressive the vocabulary is. A two-line message can score 4. A ten-page deck can score 0. Most workslop scores 0 to 1 while looking like a 4.

Calibration: six scored messages

Scoring is a skill, and it needs calibration. Score each of the following before reading the verdict.

If your scores were within one point of the verdicts, you are calibrated. If you scored C or E above 1, look again at what "specific" and "stakes" require: the presence of a topic is not the presence of a claim.

The message that looks like a 4 and scores a 2

The most dangerous category for a busy reader is the message that has the shape of signal without the substance. It has a subject line with a date. It has bullet points. It has a "recommendation" section. And when you look for the checkable claim or the actual position, it is not there. Training yourself to score these correctly is the single most useful skill in the course, because these are what AI produces by default when briefed poorly, and they are what your readers are learning to skim.

Apply the Filter, as Reader and Writer

As reader: the triage protocol

Your attention is the resource everyone else's noise is spending. Protect it with a standing protocol for incoming communication:

Step 1: Score before you study. A 20-second scan against the Signal Test comes before any deep read. Stakes and stance are usually visible in the first four lines. If they are missing there, they are usually missing everywhere.

Step 2: Route by score.

• 3 to 4: Read fully. This person did the thinking.
• 2: Read for the signal part and skim the rest.
• 0 to 1: Do not absorb the cost silently. Reply with the question that forces the missing part: "What do you need from me on this?" (missing stakes) or "What is your recommendation?" (missing stance). One sentence from you transfers the thinking work back to where it belongs.
• Suspected workslop: Same move, sharpened: "Before I dig in, can you give me the three-line version of what this concludes and what you want me to do with it?" Anyone who did real thinking answers easily. Anyone who did not will have to do the thinking now, which is the outcome you want.

Step 3: Train your senders. The workslop research found the cost lands on recipients because recipients absorb it quietly. Consistent, polite score-forcing questions change what people send you within weeks. You are not being difficult. You are declining to complete other people's unfinished work.

Step 4: Reward signal visibly. When someone sends you a 4, say so, briefly and specifically: "This was easy to act on. The ask in line one helped." Senders repeat what gets noticed. This step is the one most people skip and the one that changes behavior fastest.

The tone of a score-forcing reply

The reply that forces a missing part should be short, neutral, and free of judgment. Compare:

Too sharp: "I do not have time to read all this. What is the point?"

Right: "Can you give me the one-line ask and your recommendation? I will move fast once I have them."

The second version does the same work without a withdrawal from the trust ledger.

Use AI on the incoming side

Where volume is the problem, AI is a legitimate filter, and the research says it is exactly where readers want it used. Paste a long document and ask:

Read this and answer three questions only:
1. What does it conclude?
2. What does it want from me, specifically, and by when?
3. What verifiable claims does it make (numbers, dates,
   names, commitments)?
If any answer is missing from the document, say
"not stated" rather than inferring one.

If the AI cannot find answers, the document likely does not contain them, and you just saved the two hours the research says each workslop instance costs.

Two cautions. First, an AI summary of a long thread can miss a contradiction between participants; ask it to flag disagreements explicitly. Second, never act on an AI-extracted commitment ("Sam agreed to X") without checking the source line, because summarizers promote suggestions to agreements.

As writer: the pre-send pass

Now flip the filter outward. Before sending anything with stakes, run this pass. With practice it takes three minutes.

1. State the stakes in line one. The reader should know within four lines why this message exists and what it affects. If your point currently arrives in paragraph three, move it.
2. Replace one abstraction with one specific. Find your vaguest claim and make it checkable: a number, a name, a date, an example.
3. Commit to a stance. If the message contains options, it also contains your recommendation and your reason. If you genuinely cannot recommend, say precisely what information would change that, which is itself a stance.
4. Add one line only you could write. A reference to the actual prior conversation, the specific concern this reader raised, the shorthand you share. One authentic line shifts the register of an entire message.
5. Cut a fifth. Then read it aloud once. Anything you would not say to the person's face gets rewritten in words you would.

Worked example 1: bad news to leadership

The scenario: you must tell a leadership group that a system rollout is slipping six weeks.

Before the pass (scores 1 of 4):

Subject: Rollout Update

Hi all, sharing a status update on the rollout. The team has been working through several implementation challenges, and while significant progress has been made, we are adjusting the timeline to ensure a successful outcome. We remain committed to delivering a high-quality result and will continue to keep everyone informed as the plan evolves. Please reach out with any questions.

After the pass (scores 4 of 4):

Subject: Rollout slipping 6 weeks. New go-live May 12. One decision needed.

The rollout is moving from March 31 to May 12. The cause: integration testing surfaced a data-mapping error affecting about 15 percent of records, and fixing it right costs six weeks.

I recommend we take the delay rather than launch and correct live. A wrong-data launch costs us the frontline trust we spent last quarter building, and that is harder to schedule a fix for.

One decision needed from this group by Friday: hold the staff training dates and re-run them in May, or push training too. I recommend holding them; the refresher cost is small and it keeps momentum.

Same news. Same author. The second version is shorter, harder to write, and the only one of the two that will be remembered accurately in a week.

Worked example 2: a request for resources

Before (scores 1 of 4):

Subject: Analytics support

Hi Jordan, I wanted to touch base regarding analytics capacity for the coming quarter. As you know, our reporting needs have been growing, and we are finding it increasingly challenging to keep up with demand. It would be great to explore options for additional support. Happy to discuss whenever convenient.

After (scores 4 of 4):

Subject: Request: 0.5 FTE analyst for Q4, decision by 9/26

I am asking for half an analyst for Q4, about $22K. Without it, the monthly board pack goes back to a 10-day turnaround (it was 4 days in August) and the no-show dashboard you asked for in July does not ship until January.

I recommend borrowing from the finance analytics pool rather than hiring; Sam has said he can spare half of Lena through December.

Your call by 9/26 so I can confirm with Sam before his planning cycle closes.

Worked example 3: a change announcement to a wide group

Before (scores 1 of 4):

Subject: Exciting update to our tools

Team, we are thrilled to announce the launch of our new scheduling platform, which will streamline workflows and enhance collaboration across the organization. Training resources will be made available. We appreciate your flexibility during this transition.

After (scores 4 of 4):

Subject: Scheduling moves to the new platform Oct 6. What changes for you.

On Monday, Oct 6, scheduling moves from the old system to the new platform. What changes for you: you will book from your phone, and swaps no longer need manager approval under 8 hours.

What it costs you: a 45-minute training between Sept 22 and Oct 3, and about two weeks of slower booking while everyone learns it.

I know this is the second system change this year, and the last one cost several of you weekends. This one has a two-week overlap with the old system so nothing is lost if it goes wrong. Questions to me directly; I will answer within a day.

Adapting the pass by channel

The five steps hold everywhere. The proportions shift.

The most common cross-channel failure is the deck whose slide titles are topics ("Market Overview," "Risks") instead of claims ("The market is growing 6 percent, below our plan's 25"). A deck of topic titles has no stance, and readers can tell from the table of contents.

Pressure-Testing AI Drafts

The right relationship with the tool

The research from Module 1 points at a clear division of labor. AI is a strong editor, summarizer, challenger, and drafting accelerant. It is a poor replacement for your judgment, your stance, and your knowledge of the reader, and audiences penalize output where the replacement shows.

The failure pattern behind most workslop is sequence: generate first, think later, and later never comes. The fix is to reverse the sequence. Thinking first, generation second, pressure-testing third, ownership last.

Why unbriefed generation produces the generic register

Understanding the mechanism helps you avoid it. A language model asked to "write an email about the project delay" has no information about the delay, the reader, or you. It fills every gap with the statistically safest content: a courteous opener, balanced framing, importance words without numbers, and a closing that commits to nothing. That output is optimized to be inoffensive to any reader, which is the same as being useful to none.

The generic register is a symptom of missing input. Every red flag in this module maps to a gap in the brief.

Step 1: Brief before you generate

Never open with "write an email about X." Brief the model the way you would brief a capable new chief of staff:

Draft [message type] to [audience].

The stakes: [why this matters to them, what it affects,
what happens if they do nothing]
My position: [what I recommend or have decided, and why]
Must include: [the specific facts: numbers, dates, names]
This reader: [what they care about, what they pushed
back on before, how formal we are with each other]
My voice: [how you write, described concretely, or
paste your identity file if you built one]

Put the stakes and the ask in the first four lines.
Do not add facts I have not given you. Where you need
a fact I have not supplied, insert [NEEDED: description]
instead of inventing one.

Notice what the brief forces: you cannot fill it in without having done the thinking. The brief is the work. The generation is the typing.

If a field is hard to fill, that is diagnostic. Cannot state your position? You are not ready to send. Cannot describe this reader? You do not know your audience well enough for this message to land. The brief surfaces the gap before the reader does.

Step 2: Pressure-test the draft

Run the draft through these prompts before you touch it yourself.

The signal check:

Score this draft against four criteria: Does it state
consequences for the reader? Does it make specific,
verifiable claims? Does it take a clear position? Does
it sound like a specific person rather than anyone?
Identify the weakest of the four and show me exactly
which sentences fail it.

The skeptic:

You are the most skeptical person in the room reading
this. What would you push back on, what would you not
believe, and what question would you ask that this
draft cannot answer?

The compressor:

Cut this by a third without losing any specific claim,
any part of the recommendation, or the reasoning.
Flag every sentence you cut because it carried no
information.

The compressor's flag list deserves attention. Sentences that cut without loss are the generic filler your readers have learned to skim past. Seeing your own filler listed is the fastest way to stop producing it.

The recipient simulation:

You are [role/name], who cares most about [their
priority] and previously raised [their objection].
You have 20 seconds. React to this message: what do
you take away, what do you skip, and what do you reply?

Run this one for any message where the reader's likely reaction matters more than the message's correctness, which is most messages to people senior to you.

Step 3: The fact gate

Before anything leaves your hands, verify every name, number, date, and claim about what other people said or decided. AI inserts plausible specifics with complete confidence, and a wrong specific costs more credibility than a vague one. The polish is the model's. The accuracy is yours.

List every specific claim in this draft: numbers, dates,
names, quotes, and statements about what other people
said or decided. Present them as a checklist for me to
verify. Mark any that you generated or inferred rather
than took from my input.

Anything marked as generated gets replaced with the real fact or deleted. There is no third option.

Step 4: The ownership pass

Last step, no AI involved. Read the draft aloud and change every phrase you would not say. Add the one line only you could write. Confirm the recommendation is one you will defend in the meeting when someone pushes back, because someone will, and "the AI suggested it" is not a stance.

A useful ownership standard: send nothing you could not summarize from memory, without the document, in three sentences. If you cannot, you have not finished thinking, and your reader will be the one to discover that.

Disclosure: when to say AI helped

The Korbyt and Reworked findings are direct on this: workers trust messages based on sender identity and consistency, and discovering after the fact that AI wrote something substantive reads as a trust violation. A working norm:

• No disclosure needed: AI used to fix grammar, shorten, reformat, or summarize your own material. The thinking and the claims are yours.
• Disclose briefly: AI drafted a substantial portion of something with consequences (a policy, a performance-related message, an external commitment). One line: "Drafted with AI assistance and reviewed by me."
• Do not send: anything AI wrote that you have not read fully, verified, and would defend as your own.

The test underneath all three: would the reader feel deceived if they knew? If yes, disclose or rewrite.

The red-flag list

Keep this list nearby when reviewing any draft, AI-generated or otherwise. Each item is a marker of the generic register, and each maps to a missing part of the Signal Test.

One instance is style. A cluster is a message nobody thought about.

Worked example: the full cycle

The message: telling a peer that their team's data feed is causing errors in your reports, and asking them to fix it.

Brief:

Draft a chat message to Alex (peer, runs the data team; we are informal; she pushed back last quarter when I raised a similar issue without evidence). Stakes: the September board pack has 3 wrong numbers traced to her feed; if not fixed by 9/24 I have to caveat the pack. Position: I need the feed fixed or a manual override by 9/24; I would rather fix the feed. Must include: 3 errors, 9/24 deadline, the two specific fields (census count, occupancy percent). Voice: direct, no preamble, friendly.

Generated draft:

Hi Alex! Hope you're doing well. I wanted to flag a potential issue with the data feed that seems to be impacting some of our reporting. Specifically, we've noticed a few discrepancies in the census count and occupancy percent fields. It would be great if your team could take a look when you have a chance, ideally before 9/24 when the board pack is due. Thanks so much for your help with this!

Signal check output: Weakest part is stance. "It would be great if" and "when you have a chance" hide the fact that this is a requirement with a deadline.

Skeptic output: "Three errors" is not in the draft; "a few discrepancies" is vague. Alex will ask which three rows.

Compressor output: Cut "Hope you're doing well," "I wanted to flag," "seems to be," "Thanks so much." Down 40 percent, no loss.

Fact gate: The draft did not invent anything, but it dropped the count of errors. Restore it.

Ownership pass, final:

Alex, three numbers in the September board pack trace back to the feed: two census counts and one occupancy percent. I can send the rows. I need either the feed fixed or a manual override by 9/24, or I have to caveat the pack. I would rather fix the feed. Last quarter I raised this without the rows and that was on me; this time they are attached.

Shorter, specific, committed, and it acknowledges the history Alex would remember.

What Is Landing and What Is Getting Tuned Out

The pattern behind the pattern

Across the research and across organizations, what breaks through and what gets filtered is remarkably consistent. Every winning pattern below is a Signal Test part doing its work in the wild.

Landing: short messages with the decision up front. The strongest trend in leadership communication is the decision-first format: what was decided or what is needed in the subject line and first sentence, reasoning below for those who want it. It wins because it respects the 20-second triage everyone now runs, and it matches what 56 percent of surveyed workers named as the thing that makes them pay attention: a clear action. Getting tuned out: the long update where the point is buried in paragraph four, which readers correctly treat as optional.

Landing: AI used to reduce, disclosed when it matters. 92 percent of workers want AI used to cut information volume, and workers trust messages based on the sender's identity and consistency. Organizations getting this right use AI to synthesize and shorten, and disclose AI assistance on substantive communications rather than letting people discover it. Getting tuned out: AI used to multiply output, and the broadcast update sent to everyone because sending to everyone became free. Recipients read untargeted distribution as a signal the sender did not think about who needs this, and they respond by tuning out the channel, not just the message.

Landing: one channel, one message. 46 percent of workers tune out when the same message arrives across multiple channels. The email, the chat post, the intranet banner, and the all-hands mention of the same announcement do not reinforce each other. They train readers to skip all four. Getting tuned out: the multi-channel blast, which reads as the sender covering themselves rather than informing the reader.

Landing: named trade-offs and owned misses. Messages that state what was given up ("we chose speed over completeness, here is what that costs us") and updates that lead with the miss and the correction consistently outperform success-only communication on the measure that matters: whether people believe the next message. Specificity plus stance compounds into credibility. Getting tuned out: the all-wins update, which readers discount as marketing regardless of accuracy.

Landing: the human line. In change communication especially, one sentence of genuine, specific acknowledgment ("I know this is the second system change this year, and the last one cost you two weekends") does more than paragraphs of framing. It is the signature part of the test, and it cannot be faked by register alone, because it requires actually knowing what the reader has been through.

By audience

The patterns hold everywhere. The emphasis shifts with who is reading.

Leadership and boards. They are running the strictest triage and increasingly using AI to test your claims (the third course in this series covers that in depth). What lands: the decision or ask in line one, the numbers checkable, the risks named with your response. What gets tuned out: narrative build-up, slides that summarize activity rather than state conclusions, and anything they can tell was written to impress rather than inform.

Peers and cross-functional partners. They have their own priorities and no obligation to yours. What lands: the ask stated in terms of their cost and their benefit, the history acknowledged, and a deadline that respects their planning cycle. What gets tuned out: requests framed as urgent without saying why, and messages that assume your priority is theirs.

Direct reports and teams. They read for what changes for them and whether you are being straight with them. What lands: the decision, what it means for them concretely, what it costs them, and the human line. What gets tuned out: enthusiasm they did not ask for, "exciting" announcements of things that cost them effort, and anything that reads as generated.

External stakeholders (clients, partners, regulators). They read for commitments and reliability. What lands: exact statements of what you will do and by when, plainly worded. What gets tuned out: relationship language without substance, and anything that creates a commitment you have not thought through (the commitment scan in Appendix B exists for this).

Two short cases

Case 1: the status update that got read. A program lead replaced a weekly two-page status email with a five-line version: status in one word, what changed, what is at risk, what she needs, and one line of context. Response time from her steering group dropped from days to hours, and she was asked to share the format with other program leads. Nothing about the underlying work changed. The reader's triage cost dropped to near zero.

Case 2: the announcement that got ignored. A department sent a system-change announcement by email, chat, intranet, and a mention at the all-hands, using AI-drafted text that opened with "We are excited to announce." Training attendance was below half. When the department resent one message, from the manager, opening with the date, what changed for staff, and an acknowledgment of the previous change's cost, attendance for the remaining sessions rose sharply. Same information. Different sender, channel discipline, and signature.

The meta-lesson

None of the winning patterns require writing talent. They require the thing AI made scarce: evidence of thought. Stakes, specifics, a stance, and a signature are all just thinking made visible. That is why the same four-part test predicts what lands across functions and industries, and why the habit in the wrap-up is a thinking habit, not a writing trick.

From Signal to Influence

Signal is necessary, not sufficient

A message can score 4 and still fail to move a decision, because the decision depends on people and timing the message does not control. This module covers the three habits that turn a clear message into an influential one. Each is short, and each connects the Signal Test to the meeting or decision the message serves.

Habit 1: Sequence the readers

A recommendation sent to a group of six lands with six different readers at once, each of whom reacts partly to how the others react. That is a poor environment for a decision.

Before any consequential message to a group, map the readers:

• The decider: who actually makes the call.
• The influencers: whose reaction the decider watches.
• The affected: who bears the cost and will speak to it.
• The skeptic: who will find the flaw if there is one.

Then sequence: the skeptic and the key influencer see the message, or a version of it, before the group does. This is sometimes called pre-wiring, and it is standard practice in organizations that make decisions well. It is not manipulation. It is giving the people most likely to have objections the time to raise them when they can still be addressed.

Habit 2: Pre-wire the objection

The skeptic's objection is more useful before the meeting than during it. A pre-wire message is a Signal Test message with a specific stance: "I am going to recommend X to the group on Thursday. You raised Y last time. Here is how I have addressed it. What am I missing?"

Three things happen. The skeptic feels heard, which changes how they show up. You learn the objection you did not anticipate while there is time to fix it. And in the meeting, the objection either does not come or comes with the skeptic already partly on your side.

Use AI here for one thing: generating the objections you have not thought of. The skeptic prompt from Module 4, aimed at your recommendation rather than your draft, produces a list to pre-wire against.

Habit 3: Make the follow-through visible

Influence compounds across messages, and the mechanism is follow-through. Every commitment in a message is a promise, and the reader tracks whether it was kept.

• State commitments precisely ("you will have the breakdown by Thursday"), and keep them on the day.
• Close loops explicitly ("here is the breakdown I promised Tuesday"), so the reader registers the kept promise.
• When you miss, say so first, with the new date, before the reader has to ask.

The commitment scan in Appendix B lists every promise a draft makes on your behalf. Run it on any message with more than one commitment, because a forgotten promise costs more than the message earned.

The connection to the next course

Messages that carry recommendations to boards, investors, and executive teams face a different kind of pressure: readers who use AI to check the numbers, generate the hardest questions, and surface the alternatives you did not present. The third course in this series, on pressure-testing strategy, picks up exactly there. The Signal Test still applies; it becomes the format for a document whose every claim has already been attacked.

Your Signal Habit

The habit, precisely defined

Courses end with general takeaways. This one ends with a habit with a name, a trigger, and a duration.

The habit: the three-minute signal pass. The trigger: any message where being misunderstood carries a cost. The routine:

1. Stakes in line one.
2. One abstraction replaced with one specific.
3. A stance the message commits to.
4. One line only you could write.
5. Cut a fifth, read it aloud, send.

For AI-assisted drafts, the pass extends backward and forward: brief first, pressure-test after, fact gate always, ownership last.

For consequential recommendations, add the influence layer: sequence the readers, pre-wire the skeptic, track the commitments.

Install it now: add the five steps as a saved note, a text snippet, or a checklist wherever you write. If you built a second brain, the steps go in preferences.md. The habit survives if the checklist is one keystroke away and dies if it is not.

What you built in this course

• A noise audit of your own inbox with a triage protocol to manage it
• The Signal Test: Stakes, Specificity, Stance, Signature, calibrated against scored examples
• A scored, revised, ready-to-send version of a real message
• A briefing template and pressure-test prompts for every future AI draft
• A disclosure norm for AI-assisted communication
• The red-flag list for detecting the generic register
• A reader map and pre-wire message for one real decision
• The three-minute signal pass, saved where you write

The next 30 days

Days 1 to 7: Install the pass. Run it on every message that matters. Slow at first, automatic by Friday. Send the message you built in this course.

Days 8 to 14: Fix your gap. Module 2 revealed the Signal Test part you fail most. Every message this week gets checked against that one part before sending.

Days 15 to 21: Train your inbox. Use the triage protocol's score-forcing questions on incoming noise. Politely, consistently. Reward the 4s you receive. Watch what people start sending you change.

Days 22 to 30: Adopt your pattern and pre-wire one decision. Implement the landing pattern you chose in Module 5. Pre-wire the skeptic before your next consequential recommendation. Then review: pull up three messages you sent this month and three from before the course. Score all six. The difference is your evidence.

The one-month test

One question at day 30: do the people you communicate with respond faster, push back more usefully, and ask fewer clarifying questions?

Those three outcomes are what "cutting through" actually measures. If they are moving, keep going. If not, the diagnosis is almost always the same: the pass is being skipped under time pressure, which is precisely when it pays most.

Final principle

Everything in this course reduces to one sentence: in a world where anyone can generate polished text, the scarce and valuable thing is visible evidence that a specific person thought about a specific reader.

AI will keep getting better at producing the generic version of everything. That is the backdrop that makes the genuine version unmistakable.

Appendix A: Quick Reference

The Signal Test

Stakes: does anything change for the reader? Specificity: could someone verify or disagree? Stance: do I know what the sender thinks and wants? Signature: could only this person have sent it? Score 0 to 4.

The triage protocol

Score before you study. Route by score. Force missing parts with one question. Train your senders. Reward signal visibly.

The three-minute signal pass

Stakes in line one. One specific. A stance. One line only you could write. Cut a fifth, read aloud, send.

The AI draft workflow

Brief first (stakes, position, specifics, reader, voice). Pressure-test (signal check, skeptic, compressor, recipient simulation). Fact gate every specific. Ownership pass aloud. Disclose when it matters.

The disclosure norm

Edits to your own material: no disclosure. Substantial AI drafting with consequences: one line. Unread or unverified AI output: do not send.

The red flags

Filler openers. Importance words with no content. Symmetrical hedging. Suspiciously tidy structure. Praise inflation. Conclusions that conclude nothing. Passive ownership. Universal applicability.

The influence layer

Map the decider, influencer, affected, skeptic. Pre-wire the skeptic. Track and close every commitment.

Key evidence

BetterUp Labs and Stanford Social Media Lab (2025), 1,150 U.S. desk workers: 40 percent received workslop in the past month; about 15 percent of incoming work qualified for those who received it; roughly two hours lost per instance; about $186 per employee per month; roughly half viewed senders as less capable, 42 percent trusted them less. Korbyt and Reworked, 2026 State of Workplace Communication Survey, 1,175 U.S. employees: 81 percent can usually detect AI-written messages; about 45 percent question accuracy when AI is suspected; 44 percent tune out; 56 percent pay attention to a clear action; 59 percent tie confidence to clarity and specificity; 46 percent tune out repeated cross-channel messages; 92 percent want AI to reduce information overload.

Appendix B: Prompt Library

Copy-Paste AI Instructions for Every Stage

Every prompt below is tool-agnostic and ready to copy. Replace anything in [brackets]. Prompts are grouped by workflow stage. Save this appendix wherever you write.

Stage 1: Filtering incoming communication

The triage summary (for any long document, deck, or thread)

Read this and answer three questions only:
1. What does it conclude?
2. What does it want from me, specifically, and by when?
3. What verifiable claims does it make (numbers, dates,
   names, commitments)?
If any answer is missing from the document, say
"not stated" rather than inferring one.

The workslop detector

Assess this document for substance. Identify:
1. Claims that are specific and checkable
2. Sentences that carry no information and could be
   deleted without loss
3. Whether it contains an actual position or
   recommendation, and if so, what it is
Give a one-line verdict: is this worth a full read,
a skim, or a reply asking for the point?

The thread condenser (for long email or chat threads)

From this thread, extract only:
1. Decisions made, with who made them
2. Open questions and who owns each
3. Action items with owners and dates
4. Anything directed at me specifically
Ignore pleasantries and repetition. Flag any point
where participants contradict each other. Do not
promote a suggestion to a decision unless the thread
shows agreement.

The meeting extract (for raw notes or transcripts)

Extract from these notes in three short bullets each:
decisions made, action items with owners, and open
questions. Nothing else. Write "not stated" for any
missing owner or date.

Stage 2: Briefing before you draft

The full brief (for any message that matters)

Draft [message type] to [audience].

The stakes: [why this matters to them, what it affects,
what happens if they do nothing]
My position: [what I recommend or have decided, and why]
Must include: [the specific facts: numbers, dates, names]
This reader: [what they care about, what they pushed
back on before, how formal we are with each other]
My voice: [how you write, described concretely, or
paste your identity file]

Put the stakes and the ask in the first four lines.
Do not add facts I have not given you. Where you need
a fact I have not supplied, insert [NEEDED: description]
instead of inventing one.

The short-form brief (for quick messages)

Draft a short [email / chat message] to [person].
Point: [the one thing it must land]
Ask: [what I need from them and by when]
Tone: [how we talk to each other]
Four sentences maximum. Ask first, context after.

The bad-news brief

Draft a message delivering this news: [the news, stated
plainly, with the real numbers and dates].
Structure: the news in line one, the cause in one
sentence, my recommendation, then the one decision I
need from the reader and by when.
No softening preamble. No "I hope this finds you well."
Acknowledge this specific impact on the reader: [what
it actually costs them].

The request brief

Draft a request to [person] for [resource / decision /
action].
What I am asking for, exactly: [specifics, cost, duration]
What happens without it: [the concrete consequence]
My recommended way to fulfill it: [option and reason]
Their planning constraint: [their deadline or cycle]
Ask in the subject line and first sentence. Under
120 words.

Stage 3: Pressure-testing the draft

The signal check

Score this draft against four criteria:
1. Stakes: does it state consequences for the reader?
2. Specificity: does it make claims someone could
   verify or dispute?
3. Stance: does it take a clear position?
4. Signature: does it sound like a specific person
   rather than anyone?
Identify the weakest of the four and show me exactly
which sentences fail it.

The skeptic

You are the most skeptical person in the room reading
this. What would you push back on, what would you not
believe, and what question would you ask that this
draft cannot answer?

The compressor

Cut this by a third without losing any specific claim,
any part of the recommendation, or the reasoning.
Flag every sentence you cut because it carried no
information.

The recipient simulation

You are [role/name], who cares most about [their
priority] and previously raised [their objection].
You have 20 seconds. React to this message: what do
you take away, what do you skip, and what do you
reply?

The misread test

List every way this message could be misunderstood or
quoted out of context, especially: [the sensitive
point]. For each, show the sentence at fault and a
tighter version.

Stage 4: Revising toward signal

The de-hedge

This draft hedges. Rewrite it to commit: pick the
strongest supportable version of each claim, state my
recommendation plainly, and move every "on the other
hand" into a single short risks section with my
response to each risk.

The stakes-first restructure

Restructure this so the first four lines contain: what
this is about, why it matters to the reader, and what
I need from them. Everything else becomes supporting
detail below, in descending order of importance.

The filler sweep

Remove all of the following from this draft: filler
openers, importance words with no number or example
attached ("significant," "robust," "impactful"),
praise inflation, passive constructions that hide who
decided, and closing sentences that conclude nothing.
Show me what was removed.

The voice pass

Here are two examples of how I actually write:
[paste two real messages you sent and liked].
Rewrite this draft to match that voice: same sentence
length, same level of formality, same kind of
directness. Do not change any facts or the
recommendation.

The subject line

Write 3 subject line options for this message. Each
must contain the decision or the deadline, not a
topic label. "Vendor decision needed by Fri" beats
"Vendor Update."

The slide-title converter (for decks and pre-reads)

Here are my slide titles: [paste]. Rewrite each as a
claim the slide proves, in one sentence, rather than
a topic label. Flag any slide whose content does not
actually support a claim.

Stage 5: Final gates before sending

The fact gate

List every specific claim in this draft: numbers,
dates, names, quotes, and statements about what other
people said or decided. Present them as a checklist
for me to verify. Mark any that you generated or
inferred rather than took from my input.

The three-sentence test

Summarize this message in three sentences: the point,
the reasoning, the ask. If any of the three is
unclear from the draft, tell me which and why.

The commitment scan

List every promise, deadline, or commitment this
message makes on my behalf, explicit or implied.

Stage 6: Influence

The objection generator

I am going to recommend [recommendation] to [group].
The people in the room: [roles and what each cares
about]. Generate the strongest objection each person
is likely to raise, and mark the one I most need to
address before the meeting.

The pre-wire drafter

Draft a three-line message to [skeptic], who raised
[past objection] last time. Tell them what I will
recommend on [date], how I have addressed their point,
and ask what I am missing. Informal, no preamble.

One-line snippets for the incoming side

These are replies you send to people, not prompts for AI. Keep them ready:

• Missing stakes: "What do you need from me on this?"
• Missing stance: "What is your recommendation?"
• Suspected workslop: "Before I dig in, can you give me the three-line version of what this concludes and what you want me to do with it?"
• Buried signal: "Can you put the ask in the first line next time? Makes it easier for me to move fast on it."
• Rewarding a 4: "This was easy to act on. The ask in line one helped."

Standing instructions worth saving

Add these to your AI tool's custom instructions, or to preferences.md if you built a second brain:

When drafting communication for me:
- Put the point and the ask in the first four lines
- Never invent facts, numbers, or quotes; mark gaps
  as [NEEDED: description]
- No filler openers, no importance words without
  content, no conclusions that conclude nothing
- If I have not stated a position, ask me for one
  before drafting; do not write around the absence
- Default to shorter; flag any sentence that exists
  for tone rather than information
- Before finalizing, list every specific claim you
  did not take directly from my input