CRM Costs CRM pricing, add-ons, and support staffing calculators

Thomson Reuters Future of Professionals

Thomson Reuters Says Shadow AI Is Putting $143B of Work at Risk

The major impact is governance: AI adoption is already happening inside teams, but many organizations cannot prove which tools are approved, which data is safe, or which outputs can be defended.

Open cost calculator Plan cleanup and agents

Direct answer

Thomson Reuters shadow AI revenue risk: what CRM buyers should take from it

Thomson Reuters released its 2026 Future of Professionals report on June 22, 2026 and warned that up to $143 billion in U.S. professional-services revenue is under active reassessment based on AI delivery. The report also says one third of lawyers, accountants, and compliance professionals are using unsanctioned AI tools. For CRM and support operations leaders, the lesson is clear: AI governance is now an operating control, not a policy document.

Shadow AI control map visual card for CRM and support operations governance.

Published 6/24/2026. News event: 6/22/2026.

What happened

  • Thomson Reuters said its 2026 Future of Professionals report is based on a global survey of 1,800 professionals.
  • The company warned that up to $143 billion in U.S. legal and accounting revenue is under active reconsideration as clients expect AI-enabled improvements.
  • The report says 74% of professionals use AI weekly, but 91% believe their organizations are falling short of what AI can deliver.
  • Thomson Reuters also said one third of lawyers, accountants, and compliance professionals use unsanctioned AI, creating risk their organizations cannot monitor or control.

Why this is trending

  • The story turns AI from a productivity experiment into a measurable revenue, client-retention, talent, and compliance risk.
  • Shadow AI is now a day-to-day operations problem: employees may paste sensitive records, ticket details, contracts, or client context into unapproved tools because official workflows are too slow.
  • CRM and support operations teams are often where the risk appears first, because customer records, internal notes, case histories, and service data move through many tools.

The CRM Costs take

CRM and support buyers should not respond by banning AI with a policy nobody follows. They need a control layer: approved tools, field-level data rules, workflow owners, audit logs, prompt/output review, and a way to measure whether automation improves customer outcomes instead of creating hidden risk.

Shadow AI Control Map

A support operations framework for finding unsanctioned AI use, mapping confidential workflows, approving tools, measuring value, and preventing hidden CRM or helpdesk risk.

Cost layer
Buyer question
Risk signal and next step
Tool inventory
Can the business list every AI tool used for tickets, CRM notes, call summaries, reporting, proposals, and internal service?
Employees solve work in personal or browser tools because official systems are too slow.

Create an AI tool register with owner, approved use case, data boundary, and renewal or access status.

Data exposure
Which customer fields, transcripts, support notes, contracts, and attachments may enter an AI tool?
Confidential content is copied into prompts without logging or approval.

Define no-paste fields and safe summary patterns before enabling broad AI use.

Output trust
Who verifies AI-generated answers before they affect customers, accounts, billing, or compliance?
Teams accept polished outputs because they look complete.

Require source links, reviewer approval, and exception flags for customer-facing or regulated outputs.

Workflow ownership
Does each AI-assisted workflow have a human owner for exceptions and quality?
Automation spans CRM, helpdesk, phone, and documents, but no person owns the full path.

Assign workflow owners for intake, routing, resolution, escalation, QA, and correction.

ROI measurement
Is AI value measured by outcome quality, rework, customer effort, and revenue risk rather than activity volume?
The team reports AI usage without knowing whether it reduces mistakes or improves service.

Track reopened cases, correction rate, cycle time, escalation accuracy, and client-impact evidence.

What buyers should do next

Step 1 Audit where employees already use AI across CRM, helpdesk, phone summaries, spreadsheets, and internal service requests.
Step 2 Classify approved, restricted, and prohibited data before adding more automation to customer workflows.
Step 3 Assign a human owner and review rule for every AI output that affects customers, money, access, legal claims, or compliance.
Step 4 Measure AI by service outcome quality, not by the number of prompts, summaries, or automated resolutions generated.

Buyer FAQs

What did Thomson Reuters report?

Thomson Reuters reported that professional firms face major AI execution pressure, including up to $143 billion in U.S. revenue under active reassessment and widespread use of unsanctioned AI tools.

Why does shadow AI matter for CRM and support operations?

CRM and support workflows contain customer records, notes, transcripts, account details, contracts, billing context, and internal decisions. Unsanctioned AI use can expose that data or produce unsupported outputs.

Should teams ban AI tools to reduce risk?

A blanket ban often drives more shadow usage. A better control is to approve specific tools and workflows, define data boundaries, log use, assign owners, and verify outputs before they reach customers.

What should an operations leader check first?

Start with an AI tool inventory, then map which customer data can enter each tool, who approves output, and how the team measures whether AI improves service quality.