Home/ Resources/ Knowledge Chaos Calculator
The Knowledge Chaos Calculator

How much is knowledge Chaos costing your organization?

Estimate the time and labor your team may be losing to searching, interruptions, and repeated questions.

Six assumptions, all yours to change. The math is on the page.

Chaos, the CurrentWave AI character built from stacks of paper and sticky notes, pointing ahead
Your assumptions

Start with your industry, then make it yours

Start here: choose your industry

Your inputs

Everyone who has to look things up.

$

Wages plus benefits and payroll taxes.

Your estimate

What searching costs you today, and what you can recover.

Spent searching today
13,000 hrs
a year · 250 hrs a week
Hours recovered with CurrentWave Portal™
4,550 hrs
a year · 88 hrs a week
Assumes 35% of search time is recovered, the conservative share McKinsey found a searchable knowledge base can remove.

That is about 2.8 full-time people for a whole year, back on the phone, at the branch, and with your customers or members: faster answers, shorter holds, less waiting.

About this estimate

This is an estimate built from the assumptions you entered, not a financial guarantee of actual savings. Recovered hours are paid time redirected to other work, not a reduction in payroll. Actual results depend on adoption, use case, documentation quality, and other factors specific to your organization. Future performance or savings cannot be guaranteed and past performance does not dictate future results. User inputs are not stored nor transmitted.

The math

Nothing here is hidden

One entry for each row of the estimate, numbered to match the markers beside it. A 52-week year, and no adjustment for seasonality, part-time schedules, or paid time off.

1

Spent searching today

search hours = employees × searches per week × minutes per search ÷ 60 help hours = employees × searches per week × share requiring help × minutes helping ÷ 60 weekly = search hours + help hours annually = weekly × 52 monthly = annually ÷ 12

Two people's time, counted separately: search hours is the person with the question, help hours is the colleague who supplies the answer. Monthly is derived from the year rather than from four weeks, so the three periods always agree with each other.

2

Hours recovered with CurrentWave AI

recovered weekly = search hours × 35% + help hours × 35% recovered annually = recovered weekly × 52 recovered monthly = recovered annually ÷ 12

The line under this row divides the annual figure by 2,080 (a 40-hour week for 52 weeks) to express it as full-time people for a year, or by 40 to express it as weeks of one person’s time when the total is under a year. The planning figure. 35% is McKinsey Global Institute's estimate of what a searchable record of organizational knowledge removes from search time. The optional customer-results view on the hours tab swaps in 98% and 60%, the reductions in answer retrieval time and interruptions we have measured with customers. The dollars tab always uses the 35% figure, never the customer-observed one.

3

Hours still spent searching

remaining = hours for that period − hours recovered for that period

What the estimate leaves in place, because the model does not assume every search disappears.

4

Hours recovered, per employee

per employee = hours recovered for that period ÷ employees

Row 2 divided across headcount, which is usually the figure a manager can picture. On the dollars tab it is the labor cost of those recovered hours per employee, at the hourly rate you entered.

$

Every figure on the dollars tab

dollars = the hours figure in that same cell × fully loaded hourly cost

The dollars tab adds no new assumptions and changes no hours. Each cell is the hours cell behind it multiplied by the rate you entered, so it prices time already being paid for rather than showing new spending.

Where the defaults come from

Every number has a source

The defaults are anchored to published research and federal wage data, and set below what that research would support. Read them and change them.

Tiers rank how independent each source is from anyone selling something.Tier 1 federal statisticsTier 2 consulting researchTier 3 vendor-sponsored surveysTier 4 our own customer data

Tier 1Federal statistics

The hourly cost default

Total employer compensation costs for private-industry workers averaged $46.60 per hour worked in March 2026: $32.60 in wages and salaries and $14.01 in benefits, meaning benefits are about 30% of what an hour of work costs an employer. At the median wage percentile, total compensation was $34.78 per hour.

U.S. Bureau of Labor Statistics, Employer Costs for Employee Compensation — March 2026, released June 12, 2026. bls.gov/news.release/ecec.nr0.htm

Tier 1Federal statistics

How each industry starting point is priced

Every industry preset sets its hourly cost the same way: take the BLS median hourly wage for the frontline roles that do the looking up in that sector, blend them, and gross them up so benefits and payroll taxes make up about 30% of the loaded hour (divide by 0.7). May 2025 medians behind the blends include tellers at $20.69, customer service representatives at $21.53, new accounts clerks at $22.92, loan interviewers and clerks at $24.05, loan officers at $36.87, compliance officers at $38.81, claims adjusters and examiners at $37.50, insurance claims and policy processing clerks at $23.67, hotel, motel, and resort desk clerks at $16.86, reservation and transportation ticket agents at $21.34, and exercise trainers and group fitness instructors at $22.67. That produces roughly $34 for a small bank or credit union, $37 for a mid-size institution, $41 for a large institution, $43 in insurance, $31 in a contact center, $26 in travel and hospitality, and $32 in studios and lesson schools. The exact recipe for each is in the table below, so the rates can be refreshed each May when new medians publish. The catch-all starting point uses the BLS median across all occupations, $24.51, grossed up the same way to about $35, for organizations outside those sectors.

The other four inputs do not vary by industry and are our own assumptions, not figures from any study. No published dataset reports how often a bank teller or a hotel front desk agent looks something up, so rather than guess sector by sector, every starting point uses one baseline: 8 lookups a week, 8 minutes each, 30% pulling in a colleague for 5 minutes. Sized against the research, that is about a sixth of McKinsey’s search time, and the colleague portion — roughly 12 minutes a week per employee — is a small fraction of the 5.3 hours a week the Panopto and YouGov survey above found employees waiting on coworkers. Treat it as a starting assumption to replace with your own.

Role weights behind each hourly rate. Rate = weighted median ÷ 0.7, rounded to the dollar.
Starting pointBlend (BLS May 2025 medians)Weighted medianRate
Small bank or credit unionTellers 30%, customer service reps 25%, new accounts clerks 20%, loan interviewers and clerks 10%, loan officers 15%$24.11$34
Mid-size bank or credit unionTellers 30%, customer service reps 20%, new accounts clerks 15%, loan interviewers and clerks 10%, loan officers 15%, compliance officers 10%$25.78$37
Large bank or credit unionTellers 20%, customer service reps 20%, new accounts clerks 5%, loan interviewers and clerks 10%, loan officers 25%, compliance officers 20%$28.99$41
InsuranceClaims adjusters and examiners 50%, claims and policy processing clerks 30%, customer service reps 20%$30.16$43
Contact center or BPOCustomer service reps 100%$21.53$31
Travel & hospitalityHotel, motel, and resort desk clerks 70%, reservation and ticket agents 30%$18.20$26
Studios & lesson schoolsExercise trainers and group fitness instructors 50%, customer service reps 50%$22.10$32
Something elseAll occupations 100%$24.51$35

U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025, national occupational employment and wage estimates, released May 15, 2026. bls.gov/news.release/ocwage.t01.htm
Benefits uplift from the BLS Employer Costs for Employee Compensation series cited above, where benefits are about 30% of what an hour of work costs an employer.

Tier 2Consulting research

Workers spend a large share of the week looking for internal information, and the newest research says it is not shrinking

The baseline everyone in this category still cites is McKinsey Global Institute's 2012 estimate that the average interaction worker spends "nearly 20 percent" of the workweek looking for internal information or tracking down colleagues who can help with a task — 19% in the underlying exhibit. The same report estimates a searchable record of organizational knowledge could cut that search time by as much as 35%. Newer research points the same direction: Gartner's 2022 survey of 4,861 digital workers found 47% struggle to find the information or data they need to do their jobs, and the number of applications a desk worker uses rose from six in 2019 to eleven. Coveo's 2025 survey of 4,000 U.S. and U.K. employees at large companies put search time at about three hours a day.

McKinsey Global Institute, The social economy: Unlocking value and productivity through social technologies, July 2012. Exhibit on p. 47; the chart is sourced to IDC data plus MGI analysis, so treat it as an informed estimate rather than a measurement. Summary page is public; the full report requires a free McKinsey account. mckinsey.com
Gartner, Gartner Survey Reveals 47% of Digital Workers Struggle to Find the Information Needed to Effectively Perform Their Jobs, May 10, 2023. Survey fielded September–November 2022 at organizations with 100 or more employees in the U.S., U.K., India and China. gartner.com
Coveo, EX Relevance Report: The Search for Relevance, April 29, 2025, with Arlington Research; companies of 5,000 or more employees. Coveo sells enterprise search, so read this as vendor research. coveo.com
The mid-size default in this calculator works out to about 1.3 hours per employee per week, including the colleague’s time. That is roughly a sixth of the oldest and lowest of these figures, and a fraction of the newest.

Tier 3Vendor-sponsored survey

The colleague’s time is a real and measured cost

The two inputs about pulling in a colleague rest on the best available survey of that behavior. Panopto commissioned YouGov to survey 1,001 U.S. workers at organizations of 200 or more employees; 60% said it is difficult, very difficult, or nearly impossible to get the information they need from colleagues, and the average employee reported spending 5.3 hours a week waiting for assistance or insights from coworkers, or recreating knowledge that already existed. This calculator counts only the colleague’s answering time, not the asker’s waiting time, so its default — about 12 minutes a week per employee — is deliberately far below that figure.

Panopto, Workplace Knowledge and Productivity Report, July 17, 2018, fielded by YouGov. Panopto sells video knowledge-sharing software, so read this as vendor research with a disclosed sample. panopto.com

Tier 4CurrentWave data

What an answer looks like when the searching stops

At Marine Credit Union, time to an answer is 4.24 seconds, down from about two minutes of searching. Rolled out to the call center first, then all 16 branches.

CurrentWave AI customer data, verified with the customer. Read the case study

Not usedDeliberately excluded

Three things this calculator will not count

The cost of losing focus after an interruption, the cost of a wrong or inconsistent answer reaching a member, and the cost of onboarding time. All three are real and all three are hard to source credibly at an organization-specific level, so they stay out of the arithmetic. If they matter to your business case, they belong in a conversation, not in a slider.

What changes

Where CurrentWave AI can help

The estimate measures two things: time spent searching and time spent interrupting a colleague. The first two cards are what the product does about them. The third is why the answers hold up.

Joulie, the CurrentWave AI assistant, waving

Faster access to answers

One question, one answer, sourced from your own policies and documentation — instead of four systems and a guess.

Fewer repetitive interruptions

The supervisor stops being the search engine. The same question answered the tenth time costs nothing.

Consistent, sourced answers

Every answer cites the document it came from, so five people asking get one answer and can see where it came from.

Want to see this with your own knowledge?

Bring your estimate and your hardest internal question. We will show you CurrentWave AI answering it from your documentation.