2026 Decision Intelligence Benchmark — Special AI Report

The State of AI in Data Privacy

How data privacy leaders are adopting AI — or not — to achieve their business objectives

Table of Contents

This report is interactive. Use the controls to filter all charts by role level. Charts update instantly. Hover over bars to see exact values. Click All Roles to return to the overall view, which is the view reflected in the written narrative.

View by segment
Executive Summary

Data Privacy Leaders Are First on AI Ambition — and Last on Trust

Data privacy leaders hold a contradiction that no other function in this benchmark holds quite the same way. On automation outlook, they rank 1st of the 11 functions surveyed — 19% expect 40% or more of the function's work to be automated within 24 months, more than double the 9% cross-functional average. On trust, the same group sits at the opposite end: public GenAI draws moderate or significant trust from just 8% of leaders, the lowest of any function on that category and 21 points below the 29% average, while company-owned AI trust (35%) ranks 10th of 11 and enterprise AI trust (46%) ranks 9th of 11.

That gap between ambitious automation expectations and trust runs through usage as well. Data privacy's team usage expectations for enterprise (65%) and company-owned AI (56%) both rank 4th of 11 functions, comfortably above their respective averages — but actual daily use tells a split story: enterprise AI use (46%) holds up well at 3rd of 11, while company-owned AI use (19%) falls to 10th of 11, the second-lowest in the benchmark. The function's strongest proficiency also sits in company-owned AI (71% rated competent or better, 2nd of 11), while public GenAI proficiency (50%) ranks 10th of 11 — a function that has built real fluency with the tools it controls, and has not yet extended that fluency, trust, or daily habit to the public tools used more commonly by other teams.

That tension shows up across every level of the team. Senior leaders report meaningfully higher dependence on AI and higher proficiency with company-owned AI than directors, while managers register the highest share of significant mistrust in public generative AI of any role level — a different pattern than many other functions in this benchmark, where seniority tracks with skepticism. The function is bracing for AI to take on more of the work — reviews, intake, documentation — while remaining, by a wide margin, the benchmark's most skeptical and least trusting of AI tools.

What this means for leaders
Different Perspectives on Different Types of AI

For the purposes of this report, and to better understand how leaders think about different kinds of AI, we grouped tools into three categories:

Public Generative AI
Broadly available tools such as ChatGPT, Claude, Gemini, and similar public-facing generative AI platforms.
Enterprise Platform AI
AI features embedded in vendor-supplied systems, such as CRM tools, productivity suites, or other enterprise software.
Company-Owned AI
Proprietary, internally governed, or organization-controlled AI tools built, licensed, or configured specifically for the company or its teams.
Part 1: AI Adoption

Data Privacy Leaders Have High Expectations — but Actual Use Trails Furthest on the Tools They Built Themselves

Data privacy leaders set their highest usage expectations for enterprise platform AI, not public tools. Reviewing expectations, actual use, and the gap between them — and how each compares with the other ten functions in this benchmark — shows where the function's AI expectations are running ahead of practice.

Highlights from the data
  • Enterprise AI carries the highest expected daily use: 65% of leaders expect daily enterprise AI use, ahead of 57% for public GenAI and 56% for company-owned AI. Data privacy ranks 4th among 11 functions in both enterprise and company-owned daily use expectations.
  • Actual daily use trails expectations across every category: 46% report daily enterprise AI use (3rd of 11 functions, against a 37% cross-functional average), 33% report daily public GenAI use (9th of 11, against a 46% average), and just 19% report daily company-owned AI use (10th of 11, against a 26% average).
  • Some of the largest gaps between usage and expectations: The widest gap belongs to the tools built in-house. Company-owned AI trails expectation by 37 points (4th-largest gap of 11 functions, against a 24-point average), more than enterprise AI (19 points, 7th-largest of 11, against a 24-point average) or public GenAI (24 points, 4th-largest of 11, against a 17-point average).

Expectations of AI Use

Data privacy leaders set their highest bar for enterprise platform AI: 65% expect their teams to use it daily, ranking 4th of 11 functions and five points above the 60% cross-functional average. Public GenAI expectations (57%) rank lower — 7th of 11, six points below the 63% average — while company-owned AI (56%) ranks 4th of 11, six points above its 50% average. Data privacy is part of a small minority: one of only three functions — alongside talent marketing and supply chain — where enterprise platform AI leads expected usage across our three categories.

In data privacy, senior leaders are setting the pace. Averaged across all three AI types, data privacy's expected-usage figure of 59% lands seventh of eleven functions, close to the 58% cross-functional average — squarely middle of the pack.

Expected Daily Use by AI Type
Q: What is the frequency of use you expect from your team for the following types of AI tools? (Daily responses shown)
(N=26) · "N/A" responses excluded; distribution normalized to valid responses per AI type.
Source: Assemble’s Decision Intelligence Benchmark — Q1 2026

Actual AI Use

Actual use lands well below expectations in every category, and data privacy's standing relative to peers shifts sharply by type. Enterprise AI leads with 46% reporting daily use — 3rd of 11 functions, nine points above the 37% cross-functional average. Public GenAI sits at 33%, ranking 9th of 11 and 13 points below the 46% average. Company-owned AI trails at just 19%, the second-lowest of the 11 functions and seven points below the 26% average. Even the category with the most institutional backing — internally built AI — produces both the lowest daily habit in the function and one of the weakest showings in the benchmark.

Averaged across all three AI types, data privacy's actual-use figure of 33% ranks eighth of eleven functions, slightly below the 36% cross-functional average.

Actual Daily Use by AI Type
Q: What is the actual frequency of use from your team for the following types of AI tools? (Daily responses shown)
(N=26) · "N/A" responses excluded; distribution normalized to valid responses per AI type.
Source: Assemble’s Decision Intelligence Benchmark — Q1 2026

The Gap

Every category shows a gap between expectation and practice, and the largest gap sits with the AI the organization owns outright. Company-owned AI shows a 37-point gap between 56% expected daily use and 19% actual. Enterprise AI shows the smallest gap, at 19 points, while public GenAI sits in between at 24 points.

The pattern inverts what shows up elsewhere in this benchmark series: The more control an organization has over an AI tool, the further actual use lags what leaders expect of it.

The Gap: Expected vs. Actual Daily Use
Percentage-point difference between expected and actual daily use, by AI type
(N=26)
Source: Assemble’s Decision Intelligence Benchmark — Q1 2026
Our Take

The gap here may represent a build problem. Data privacy leaders trust enterprise and company-owned tools more than public ones, yet actual use of the tools they trust most lags furthest behind. That combination points to workflow design as the missing piece, rather than skepticism. A function this exposed to AI claims from vendors knows the difference between a tool that exists and a tool that is actually load-bearing in daily work — and right now, the tools it owns are not yet load-bearing.

Three actions for leaders
Part 2: AI Dependence & Automation Outlook

Data Privacy Expects More Automation Than Any Function in the Benchmark — but Feels Least Disrupted Today

Data privacy leaders generally say losing AI would not be highly disruptive. But on the two-year horizon, automation expectations outpace every other function.

Highlights from the data
  • Current dependence on AI is low: Just 16% of data privacy leaders say AI disappearing would cause moderate or major disruption, ranking tenth of eleven functions against a 28% cross-functional average.
  • Senior leaders report meaningfully higher dependence than the rest of the team: 34% of senior leaders cite moderate or major disruption, compared with 17% of directors and 0% of managers.
  • Automation outlook ranks first of all eleven functions: 19% of data privacy leaders expect 40% or more of the function's work to be automated within 24 months, more than double the 9% cross-functional average.

Dependence on AI

Data privacy reports the lowest dependence on AI of any function in this benchmark except one. Just 16% of leaders say AI disappearing tomorrow would cause moderate or major disruption — well under the 28% cross-functional average. That reading is consistent with members' own description of the work as still fundamentally a human judgment call. Senior leaders predict 34% moderate-or-major disruption, compared with 17% of directors and 0% of managers.

Disruption if AI Disappeared Tomorrow
Q: If AI were suddenly unavailable tomorrow, how disrupted would your function be?
(N=26)
Source: Assemble’s Decision Intelligence Benchmark — Q1 2026

AI Automation Outlook

The 24-month outlook tells a different story than today's dependence. 19% of data privacy leaders expect 40% or more of the function's work to become AI-powered within two years — the highest share of any function in the benchmark, with supply chain (18%) and L&D (15%) the next-closest functions.

That ambition is consistent with what data privacy leaders describe in practice: a team buried in volume — contract reviews, privacy impact assessments, intake requests — that could welcome real automation if vendors could deliver it. The bullishness on automation's horizon and the caution about dependence today represent an interesting conflict that is worth future exploration.

Percent of Data Privacy Work Automatable by AI in the Next 24 Months
Q: Approximately what percentage of your function's work could be automated by AI over the next 24 months?
(N=26)
Source: Assemble’s Decision Intelligence Benchmark — Q1 2026
Our Take

Members describe this gap directly: AI dependence is low today because nobody has shown this group automation that works at the volume they need. This points to skepticism from the vendor market. Data privacy leaders are willing bettors on automation's future; they may be waiting for a vendor to prove it.

Three actions for leaders
Part 3: Team Proficiency with AI

Data Privacy Is Most Proficient on the AI It Owns

Unlike most functions in this benchmark, data privacy's strongest proficiency sits in company-owned AI instead of the widely available public tools.

Highlights from the data
  • Company-owned AI proficiency rated highest: 71% of data privacy teams rate competent or better on company-owned AI — highest for the function and the second-highest share of any function in the benchmark, twelve points above the 59% cross-functional average.
  • Public GenAI proficiency (50%) is the outlier: It ranks 10th of 11 functions, 18 points below the 68% cross-functional average.

Proficiency Across AI Types

Data privacy breaks from the pattern this benchmark sees elsewhere, where public GenAI typically holds the strongest proficiency simply on account of easier access and longer time in the market. Here, company-owned AI leads: 71% of leaders rate their teams competent or better, second-highest among 11 functions. Conversely, public GenAI carries the highest "beginner" concentration of the three, at 50%.

That ordering is consistent with a function that has been deliberate about which AI environments it uses. Averaged across all three AI types, data privacy's proficiency figure of 60% ranks ninth of eleven functions, close to the 61% cross-functional average. When looking at the data by level, senior leaders report a higher aptitude among their teams than managers or directors.

Team Proficiency by AI Type
Q: How would you rate your team’s overall proficiency in using each AI type?
Beginner / No experience
Competent
Advanced
Expert
(N=26) · "N/A" responses excluded; rows normalized to valid responses per AI type.
Source: Assemble’s Decision Intelligence Benchmark — Q1 2026
Our Take

Members note that privacy professionals often arrive from legal backgrounds already managing a steep technical learning curve — black-box algorithms were not on the law school syllabus. That context reframes the proficiency numbers: competence with company-owned AI at the senior level is not a given, it is the product of leaders who have had to build AI fluency on top of an already-demanding legal foundation. The question for this function is whether that fluency reaches the directors and managers.

Three actions for leaders
Part 4: AI Trust & Mistrust

Data Privacy Trusts Public GenAI Least of Any Function

Trust in this function runs almost entirely along category lines: public GenAI earns almost none, enterprise and company-owned tools earn considerably more.

Highlights from the data
  • Public GenAI trust is the lowest of any function in the benchmark: Only 8% of data privacy leaders report moderate or significant trust, against 46% for enterprise AI and 35% for company-owned AI.
  • Data privacy and security concerns lead mistrust drivers for public tools: 88% cite data privacy or security concerns for public GenAI, ahead of 81% citing inaccurate or hallucinated outputs.

Trust Across AI Types

This function trusts AI less than almost any other in the benchmark — near or at the bottom in all three categories. Public GenAI draws moderate or significant trust from just 8% of leaders — the lowest public-AI trust level of any function in the benchmark, 21 points below the 29% cross-functional average. Company-owned AI sits at 35%, ranking 10th of 11, fifteen points below its 50% average. Enterprise AI fares far better at 46%, though that still ranks 9th of 11 functions, four points below the 50% average.

Trust by AI Type
Q: For each type of AI, what is your level of trust with it?
Significant mistrust
Moderate mistrust
Neutral
Moderate trust
Significant trust
(N=26) · Rows normalized to valid responses per AI type.
Source: Assemble’s Decision Intelligence Benchmark — Q1 2026

Drivers of Mistrust

Data privacy or security concerns lead mistrust drivers for public GenAI, cited by 88% of leaders — unsurprising for a function whose job is precisely those concerns for everyone else's tools. Inaccurate or hallucinated outputs follow closely at 81%, and misalignment with internal policy sits at 73%.

Threat to job security stays low across all three categories, never reaching 20% — though it's misalignment with internal policy, not job security, that becomes the least-cited concern as AI moves into enterprise and company-owned environments. That ordering matches what members describe directly: less concern about AI replacing them, more concern about whether the organization's own tools are governed well enough to use responsibly.

Drivers of AI Mistrust — Public GenAI
Q: When you have trust concerns with Public GenAI, what are the primary reasons? (Select all that apply)
(N=26) · Multi-select; denominator is total respondents. Reflects drivers cited for Public GenAI specifically.
Source: Assemble’s Decision Intelligence Benchmark — Q1 2026
Our Take

Members describe this group's mistrust less as fear and more as professional pattern recognition. They sit on the governance committees that review every AI use case the rest of the organization wants approved, which means they see the full breadth of hype, overpromising, and quiet under-delivery before anyone else does. The line privacy leaders keep returning to — that none of the privacy tools are great; they all promise more than they deliver — is the same lens this group applies to AI generally. Low trust here is informed skepticism from the people whose job is to find the gap between marketing and reality.

Three actions for leaders
Part 5: How Data Privacy Compares Across Functions

Data Privacy Leads the Benchmark on Automation Outlook — Yet Trails on Trust and Dependence

Compared with the other ten functions in this benchmark, data privacy expects the most automation of any group surveyed. However, trust and dependence are both near the bottom of the eleven functions.

Highlights from the data
  • Data privacy ranks 1st of 11 functions on automation outlook: 19% expect 40%+ of the function's work automated within 24 months, more than double the 9% cross-functional average.
  • Data privacy ranks 10th of 11 on AI trust: a 30% average trust share across AI types, against a 43% cross-functional average.
  • Data privacy ranks 10th of 11 on AI dependence: Just 16% predict moderate or major disruption if AI disappeared, against a 28% cross-functional average.
Where Data Privacy Stands Out on AI
Data privacy’s distance from the cross-functional average on six AI vectors, aggregated across all three AI types
Ahead of average
Behind average
The benchmark average is the mean of the other ten functions surveyed: Social Media, Healthcare Social Media, ESG & Sustainability, CSR & Social Impact, Talent Marketing, Learning & Development, DEI, Data Strategy, Supply Chain, and Employee Experience. Each vector is the percent of valid responses on a standard benchmark question, averaged across the three AI types where applicable. Hover any bar for Data Privacy’s score, the benchmark average, and its rank.
Source: Assemble’s Decision Intelligence Benchmark — Q1 2026 (N=26 Data Privacy; 11 functions, cells n≥5).

Cross-Functional Position

Data privacy's position in this benchmark is unusual: The function that ranks first on automation ambition ranks near last on trust and dependence. Most functions surveyed pair high automation expectations with at least moderate trust in the tools meant to deliver it. Data privacy decouples the two — it is willing to bet on AI's future capability without yet extending much trust to AI's present capability, particularly in public form.

Where most functions build trust and dependence together as adoption deepens, data privacy's skepticism appears to be holding steady even as its automation expectations are solid — something that rewards vendors who can actually address data privacy and hallucination concerns.

Our Take

For data privacy, trust and dependence have not moved together the way they do in most functions. It is the function most optimistic about automation and yet one of the functions least willing to extend trust to AI's current tools. That may be the function's real advantage — and leaders who convert it into reliable processes, with the right guardrails, can make it durable.

Three actions for leaders