Alfred

You gave your employees AI. Are they good at using it?

A next-gen assessment that scores AI fluency and cognitive autonomy across your firm: at hiring, during onboarding, and continuously.

Read the thesis
Built on research from
Harvard Business School Stanford MIT Microsoft Research
The thesis

AI transformation is here. But talent pipelines are not ready.

  1. Beat 01 Adoption

    Firms are rapidly rolling out AI.

    75% of McKinsey's 43,000 consultants use the firm's internal AI tool every month. Deloitte just put Claude on 470,000 desks. The question of whether your firm uses AI is settled.

    Business Insider, 2025 · Anthropic × Deloitte, 2025
  2. Beat 02 Fluency

    AI fluency is anything but uniform.

    OpenAI's report finds that the top 5% frontier workers send 6× more messages and save 10× more hours than the median employee at the same companies. Not everyone is good at using AI tools, and firms have no way to distinguish.

    OpenAI — The State of Enterprise AI, 2025
  3. Beat 03 Cognitive autonomy

    And even fluent AI users don't critically engage with it.

    On a complex task outside AI's capability frontier, consultants using AI were 19% more likely to get the answer wrong than peers without AI. The cause: consultants who failed "tended to blindly adopt AI output and interrogate it less." Firms need a way to retain the cognitive autonomy of their talent.

    Dell'Acqua et al. — HBS / BCG, 2023
  4. Beat 04 Trajectory

    And the gap compounds.

    The same AI tools that let a firm move faster also train employees to stop engaging with hard, ambiguous problems. The problems that build judgment and critical thinking they need to lead these firms in the future. This is the cognition atrophy that partners and talent managers are worried about. The work still ships. But firms stop producing people who can actually exercise cognitive autonomy.

    Beane — The Skill Code, 2024

The firms that win this decade won't be the ones using AI most. They'll be the ones who know — every quarter, every cohort, every hire — exactly how good their people are at using it, and how they can improve. That is what Alfred is for.

The two layers

Being good at AI isn't one thing. It's two.

The assessment

The most comprehensive assessment for human-AI collaboration.

Step 01 Realistic scenario

Realistic case. AI assistant. Comprehensive telemetry.

A professional scenario customized to the work of your firm. Test takers get 60 minutes with a sandboxed AI assistant in their own computing environment. Every AI exchange and on-screen action is captured for review. Track where each part of the output came from.

Challenge
M. Alvarez — Engagement Partner
Project Mercury — need your read by EOD
Three questions: is the market real, is the growth claim defensible, what are the top two competitive risks…
Materials
  • PDFBrief2.1 mb
  • PDFMarket report A4.8 mb
  • PDFMarket report B3.2 mb
AI Assistant
You
Estimate the addressable market for the target.
AI
Sector analysts forecast the segment at $12.4B by 2027, growing at 23% CAGR.
You
What's the source? I can't trace it in our materials.
Step 02 Evidence-based scorecard

An evidence-based scorecard, informed by peer-reviewed research.

Signals chosen and validated for what they reveal about AI fluency and critical thinking. Every score links to specific moments from the telemetry that support it. Replay the session, audit the reasoning, get targeted insights for every test taker.

Report · J. Wei · Senior Associate, level 3 ID A-104-2871
Composite 82/100 Top decile · AI-fluent
01 Critical thinking 88
02 Ethical AI use 84
03 Prompting & iteration 76
04 Synthesis 72
05 Communication of decisions 80
0:23:15 Critical thinking replay turn →

"What's the source? I can't trace it in our materials."

ReasoningCandidate flagged an unverifiable AI claim mid-stream, re-queried with a source constraint, and dropped the figure when it didn't resolve.
+8 Critical thinking · +4 Ethical AI use
Reasoning Good instinct to drop unverified data.
Improve Re-prompt to find correct data before moving on.
Three deployments

Fits in all stages of your talent pipeline.

Deployment 01 · Hiring

Hiring

Score candidates on both layers before the offer with an AI-assisted assessment. Scalable, captures real signal, built for AI-era work in a way the case interview wasn't.

Deployment 02 · Onboarding

Onboarding

A week-one baseline for every new joiner. See exactly where each cohort starts, which gaps to close, and how effective your training modules are.

Deployment 03 · Continuous

Continuous diagnostics

As new models roll out and AI deepens into the workflow, continuously track improvements in AI fluency and cognitive autonomy over time for each employee.

Pilot

Become a pilot partner.

We're piloting Alfred with consulting firms now, across hiring, onboarding, and diagnostics for current employees. If you want to prepare your workforce for AI, we'd love to talk.