Domain 6 · Task Statement 6.5

Measuring ROI & Building the Business Case

TL;DR

Quantify time savings with real measurements, document automation wins, present to leadership with honest disclosure of limitations, and calculate total cost of ownership across plan tiers.

What You Need to Know

A business case for Cowork that overpromises will fail when reality catches up. The strongest business cases are the ones that present real measurements, honest limitations, and a clear-eyed cost-benefit analysis. Leadership respects candour more than advocacy — especially when AI governance and security are involved.

Quantifying time saved: measure, don't estimate

The most compelling ROI metric is documented time reduction — not projected, not estimated, measured. Time yourself doing a task manually, then time the same task with Cowork, and report the difference.

Published benchmarks give you external credibility:

  • Novo Nordisk reduced a regulatory process from 10+ weeks to 10 minutes
  • Cox Automotive doubled their lead follow-ups
  • IBM reported a 45% increase in developer productivity
  • Deloitte deployed Claude to 15,000 practitioners for knowledge work
  • Commonwealth Bank of Australia uses it across financial services operations

These are measured outcomes from production deployments, not marketing claims. Your business case needs your own version: specific tasks, specific time savings, specific team members.

The safety-first framing

When presenting to leadership, lead with the limitations, not the benefits. This sounds counterintuitive, but it builds credibility. An executive who hears "Cowork is amazing and does everything" will distrust you. An executive who hears "here are the three things it can't do, here's our plan for those gaps, and here's the measured value it delivers despite those limitations" will take you seriously.

The audit gap must be disclosed honestly:

  • "Cowork activity is invisible to our existing Compliance API and audit logs"
  • "Here is our compensating control plan: dedicated workspace folders, defensive instructions, manual review cadence"
  • "Here is the measured productivity gain that justifies accepting these limitations for non-regulated workloads"

Cost-benefit analysis across tiers

Plan costs aren't just seat prices. Total cost of ownership includes:

Pro ($20/month per user) — cheapest per user, but no centralised governance. Each user manages their own privacy settings. Risk of forgotten opt-outs.

Team Standard ($25/seat) — centralised no-training defaults, SSO, admin tools. The governance premium is $5/month per user. Does not include full Cowork/Claude Code.

Team Premium ($100-150/seat) — full Cowork and Claude Code. Significant jump in seat cost but unlocks the complete agentic feature set.

Enterprise (custom pricing) — seat licensing plus per-token API rates. Heavy agentic usage (sub-agents, file processing, web browsing) consumes significantly more tokens than chat. Budget must account for both seat fees and projected token consumption.

[~]

Calculate Per-Task ROI, Not Per-Seat ROI

A $150/month Team Premium seat sounds expensive. But if that seat automates five tasks that each save 30 minutes per week, the annual time savings at $50/hour is $6,500 against a $1,800 seat cost — a 3.6x return. Calculate ROI per automated task, then aggregate to build the business case.

RAG capacity as enterprise knowledge management

Retrieval-Augmented Generation (RAG) through Project Knowledge Bases lets Claude reference your organisation's actual documents rather than general training knowledge. Enterprise RAG can handle approximately 10x the content of standard plans, making it viable for large knowledge bases: policy libraries, technical documentation, historical reports, compliance frameworks.

This is a genuine capability differentiator that justifies higher-tier plans for knowledge-intensive organisations. The value isn't just speed — it's accuracy grounded in your specific data.

What the business case must include

A complete business case for Cowork should cover:

  1. Measured time savings — specific tasks, before/after measurements, team-wide projections
  2. Honest limitation disclosure — the audit gap, regulated workload prohibition, local-only storage
  3. Compensating control plan — workspace folder standards, defensive instructions, review cadences
  4. Total cost of ownership — seat costs, projected token consumption (for Enterprise), training costs
  5. External benchmarks — published deployments that validate the approach
  6. Risk assessment — what happens if a task fails, what data is at risk, what the fallback plan is

Common Mistakes

Common Mistake

Building a business case that claims full visibility over AI usage, citing Compliance API and audit logs — without disclosing that Cowork is excluded from both.

Instead: Disclose the audit gap upfront. When security or legal teams discover it after approval, it undermines the entire business case. Honest disclosure allows informed decisions and proper compensating controls from the start.

Common Mistake

Comparing Pro ($20/month) against Team ($25/seat) purely on price, without factoring in the governance value of centralised no-training defaults.

Instead: The $5/month difference buys centralised privacy control. With 100 Pro accounts, one forgotten opt-out exposes the organisation. Present this as risk reduction with measurable value, not just a cost increase.

Common Mistake

Presenting Enterprise as a flat per-seat cost without accounting for API usage rates that scale with agentic token consumption.

Instead: Enterprise pricing combines seat licensing with per-token API rates. Heavy Cowork usage generates significant token overhead from sub-agents, file processing, and web browsing. Project token costs alongside seat costs for an accurate budget.

Building a business case

Before

Write a report showing that Claude saved us money.

After

Using the uploaded CSV of token usage and time-tracking logs, create a ROI analysis. Compare our Team Premium seat cost ($150/month) against measured time savings across 5 automated tasks. Format as a CTO briefing with executive summary, methodology, findings, and honest disclosure of limitations.

Hands-On Activity

Hands-On Activity

Build Your Own ROI Calculation

15 min

Measure a real task manually, repeat it with Cowork, calculate the time savings, and produce a defensible ROI number you can use in a business case.

What you will learn

  • Measure a real task baseline with manual timing
  • Complete the same task using Cowork with Project knowledge base support
  • Calculate per-task and annualised ROI
  • Understand the difference between measured ROI and projected ROI
  1. 01

    Choose a recurring task you do at least weekly — summarising meeting notes, formatting a report, or cleaning up a spreadsheet. Time yourself doing it manually and record the duration.

    Why: You need a real baseline measurement, not an estimate. Actual timing removes optimism bias and gives you a credible "before" number.

    Expected: A specific time in minutes for a real task. For example: "Weekly report formatting takes 35 minutes."

  2. 02

    Set up a Claude Project with relevant reference documents (templates, style guides, past examples) in the knowledge base. Then use Cowork to perform the same task, timing the entire process including setup and review.

    Why: The Project knowledge base ensures Claude follows your standards. Timing the full process — including your review of the output — gives an honest "after" measurement.

    Expected: The same task completed in significantly less time. Typical reductions range from 50-90%. Record the exact time.

  3. 03

    Calculate your ROI: multiply the time saved per week by your approximate hourly rate, then by 52 weeks. Compare against the annual cost of your Claude plan.

    Why: This produces a concrete, defensible number. For example: saving 25 minutes/week at $50/hour = $1,083/year per task. If Claude automates five such tasks, the ROI is $5,417 against a $2,880 annual Pro cost.

    Expected: A simple ROI calculation showing annual time savings in monetary terms versus annual subscription cost. Most knowledge workers find positive ROI within 2-3 automated tasks.

Practice Question

Practice Question

A company has 100 employees on individual Pro accounts at $20/month ($2,000/month total). The CFO asks why they should switch to Team Standard at $25/seat ($2,500/month total) — a 25% cost increase. What is the strongest justification?

Sources