Your Slack Is Not Internal Communications

It is Friday, 4:47 PM. Your COO opens Slack to catch the week. The sidebar reads 94 unread channels and 182 unread DMs. Search for "pricing change" returns 42 threads across six channels dating back to August. Two of them contradict the deck Sales shipped on Tuesday. One of them is the CEO signing off on a number nobody rolled into the Q4 commit. The COO screenshots the three messages and pastes them into a Monday agenda item nobody will read until Tuesday at 11:00 AM.

Slack ships 42,000 messages a week across 240 channels in a Series B company of 94 heads. The function the tool was sold to run is internal communications. Internal communications is the work that lands 94 employees on the two signals that move the quarter and the four decisions the exec team signed. The tool ships the flood. The function closes the flood into a weekly cut the whole company reads in six minutes on Monday.

The 42,000-message week that lands nobody on the same page

Pull the trailing 30 days of Slack activity. 168,000 messages across 240 channels. 1,840 of them carry a decision verb (approved, signed, killed, moved, committed). 94 of them carry a date the function heads treat as binding. Zero of them ship into a weekly memo the 94 employees read against a plan.

The CEO announces the pricing change in #exec on day three of the quarter. Sales reads it on day eight in a one-off DM. Finance reads it on day 14 when the forecast misses by $420K. Support reads it on day 21 when a renewal account opens a ticket against the invoice they got Friday. Three functions, three different pricing numbers, four weeks of version drift the exec ledger was supposed to close on Monday.

The stuck version runs six company all-hands a quarter, two town halls a month, and an email from the CEO every other Thursday that opens 42 percent of the time. The all-hands slide deck lands at 4:02 PM on the day of the meeting against a prep thread that ran from Monday. The 94 employees read the deck once and the pricing change surfaces on slide 28 of 34 in a bullet the Sales lead edits on the fly. The 11 function heads walk out with 11 different versions of the number and the Monday forecast reads the drift back inside a week.

What Slack stores vs what the function ships

Slack stores the firehose. 42,000 messages a week against 240 channels, 94 DMs, and 14 huddles a day. Search works against the string the employee remembers typing in June, not against the decision the exec team signed in September. Threads bury the signoff two levels deep against a reply from the Head of People who muted the channel in week three.

The function ships four cuts a week against the firehose. Monday company memo. Thursday decision ledger. Weekly function digest to the 11 function heads. Quarterly all-hands brief that lands on day 7 of the quarter against the plan the board signed, not day 42 against the plan the exec team forgot.

The Monday memo reads the prior week in 420 words. The four decisions signed, the two decisions slipped, the three asks the CEO is routing to the function heads, and the one number every employee carries to their function cut by Tuesday 10:00 AM. 94 employees read the memo at 7:02 AM Monday on their phone, the same cut, the same owners, the same dates. The Thursday decision ledger closes the week on signoffs shipped, and the weekly function digest routes the three buried signals before the Friday QBR opens.

The three signals Slack buries every week

Pull every message in the trailing 14 days tagged with a decision verb. Sort by channel. Three signals surface every time against a Series B queue, and no human reliably routes any of them inside 48 hours of the post.

The pricing or packaging call. The CEO ships the number in #exec on Tuesday. Sales ships the deck update on day nine. Finance rebuilds the forecast on day 14. The signal is one line in one channel on one Tuesday. The function routes the line into the Monday memo, into the Sales battle card, and into the Finance forecast model inside 24 hours of the signoff.

The hire or kill call. The Head of People closes the req in #hiring on Thursday. Greenhouse never gets the status update. The candidate pipeline runs seven more interviews against a role that closed three weeks ago. The signal is two sentences in one channel on one Thursday. The function routes the kill into the hiring plan brief, into the ATS, and into the 11-req reconciliation the CFO reads Friday.

The customer escalation call. The CS lead posts the Tier-1 account at risk in #customer-success on Monday. The exec team sees it in the Friday QBR seven days late. The signal is three sentences in one channel on one Monday. The function routes the risk into the Monday exec ledger, the Thursday decision ledger, and the Zendesk-to-renewal brief the customer success function runs on cadence.

Three signals a week, buried in a 42,000-message flood. The Head of People at $180K, the COO at $240K, and the Chief of Staff at $220K rebuild the three cuts by hand every Monday. Loaded time runs $14K to $22K a month on the search-and-synthesize work Slack was priced as handling. The function ships none of it on cadence, and the Monday forecast carries the version drift into the next quarter every time.

What a fractional AI internal communications agent does

A fractional AI internal comms agent reads against Slack, Notion, Linear, Greenhouse, Salesforce, and NetSuite on day one. 168,000 messages across 240 channels. Tags every message against the 12 decision-verb patterns and the 42 signal channels. Routes the tagged signal into the Monday memo, the Thursday decision ledger, the weekly function digest, and the quarterly all-hands brief inside 60 minutes of the signoff.

Week one, the agent ships the Monday memo at 7:02 AM against the trailing week. Week two, the Thursday decision ledger lands with the four signed decisions reconciled and the two slips flagged. Week three, the weekly function digest reroutes the three signals the function heads buried against the stuck channels. Week four, the quarterly all-hands brief ships on day 7 of the quarter against the plan the board signed.

Write access sits on the four cuts only. Every reroute and every kill call routes through a human signoff before it ships. The Head of People keeps the people judgment, the COO keeps the function calls, and the Chief of Staff keeps the CEO relationship. The agent clears the firehose the three titles were triaging by hand against 42,000 messages a week.

The unit economics against a fifth comms hire

The stuck version opens a Head of Internal Communications req at $160K base plus benefits. Loaded year one lands $220K to $260K. The hire runs six all-hands a quarter, drafts the Thursday CEO email, and chases the function heads on Slack for the Monday memo the exec team reads at 11:00 AM instead of 7:02 AM. The ramp runs two quarters before the Monday cadence holds for three weeks in a row.

The fractional AI department runs the four cuts on cadence against the Slack firehose, Notion, Linear, Greenhouse, Salesforce, and NetSuite. Build cost on a 14-day sprint lands $42K to $72K. Ongoing cost sits at $3K to $5K a month on API spend, read access against the six systems, and the Head of People who signs the Monday memo. Year one lands $78K to $132K against the $220K to $260K loaded line on the fifth hire, and the Monday memo ships at 7:02 AM against 94 employees instead of 11:00 AM against 42 percent open rate.

What the sprint ships in 14 days

Day one, the agent reads against Slack, Notion, Linear, Greenhouse, Salesforce, and NetSuite. Day seven, the first Monday memo ships at 7:02 AM against the three signals the firehose buried last week. Day fourteen, the Thursday decision ledger lands with the four signed decisions reconciled and the weekly function digest scheduled to ship on cadence.

If your COO opens Slack at 4:47 PM Friday to catch a pricing change from day three of the quarter, the version where the Monday memo lands at 7:02 AM against the 94 employees is 14 days away. Slack still ships the firehose. The agent closes the function the tool was never going to close on its own. You can keep running six all-hands a quarter against a 42 percent open rate, or you can scope a sprint and read the Monday memo at 7:02 AM against the plan the board signed.