TL;DR: Companies mandating AI proficiency achieve 38% productivity gains—11 points higher than peers. Despite gains, 69% maintain headcount to amplify output rather than cut costs. Entry-level roles face the steepest displacement (28% expected within two years), while 75% of companies ban AI in interviews, creating a disconnect with the top productivity driver.

1. AI Mandates Drive 11-Point Productivity Lift

‍

Finding: Only 15% of startups require AI proficiency for technical hires, but these companies achieve 38% productivity gains vs. 27% for peers without this requirement.

What This Means: Hiring for AI capability delivers more value than upskilling current staff. Companies with formal training programs saw 42% lower productivity, suggesting training is remediation rather than optimization.

Suggestion Actions:

• Add AI tool proficiency as a required skill in technical job descriptions. 
• Screen candidates on AI literacy during interviews (tool usage, prompt engineering, workflow optimization, etc.).
• Skip broad training programs; focus on hiring people who already leverage AI effectively or self-led learning for existing teams.

2. Companies Are Scaling Output, Not Cutting Headcount

‍

Finding: Despite 27% productivity gains from AI, 69% of companies maintain their current headcount over the next 12 months. Only 20% are reducing teams (average 10% cuts, primarily in Customer Support).

What This Means: The AI dividend is being captured as expanded capacity, not efficiency gains. Companies are betting on doing more with the same team rather than doing the same with fewer people.

Suggestion Actions:

• Reallocate productivity gains toward strategic initiatives (new products, market expansion).
• Evaluate where AI can truly replace headcount and shift those resources to other, more strategically impactful, teams.
• Communicate this strategy clearly to retain talent who are nervous about AI displacement.

3. Entry-Level Roles May Vanish Within Two Years

‍

Finding: 28% of entry-level positions (both technical and non-technical) are expected to be displaced by AI within 24 months. Companies will hire more experienced talent across functions.

What This Means: The traditional career ladder is breaking. Junior roles that once provided on-ramps are disappearing, forcing companies to compete for mid-level and senior talent in a tighter market.

Suggestion Actions:

• Carefully consider the ideal organizational structure, both today and in the next several years. Having a high-potential bench of more junior employees may be impactful in the long term.
• With lower demand, hiring the most talented entry-level employees will become easier for companies willing to invest in their development.
• Expect headcount costs to increase if the focus of hiring becomes sr. employees.

4. Most Interview Processes Ignore the Top Productivity Driver

‍

Finding: 60% of companies revamped interviews for the AI era, with 75% limiting or banning AI usage during evaluations. Only 1 in 5 encourage AI in virtual interviews. Yet, AI proficiency is the strongest predictor of productivity.

What This Means: Companies want to assess "core capabilities", but may fail to evaluate the ability to effectively leverage AI proficiency. This creates a mismatch between hiring criteria and on-the-job performance.

Suggestion Actions:

• Split interviews into AI-allowed and AI-restricted sections to test both capabilities.
• Add explicit AI-assisted challenges (e.g., "solve this problem using any tools you'd normally use").
• Weight AI proficiency assessments as part of the total evaluation score.
• Train interviewers to distinguish between AI-augmented performance and AI-dependent work.

5. Smaller Companies Can Extract More Value from AI

‍

Finding: Companies under 100 employees report 33% technical productivity gains versus 22% for 500+ employee companies. Smaller teams move faster on AI adoption.

What This Means: Bureaucracy and legacy systems slow AI integration at scale. Early-stage companies have structural advantages in capturing AI value.

Suggestion Actions:

• If you're <100 employees: double down on AI tools before adding layers of process.
• If you're 500+: determine how to upskill employees to leverage the latest AI tools. 
• Measure productivity gains by team size—identify pockets of high performance and replicate their approach.
• Resist the urge to standardize tools too early; let teams experiment.

6. HR Tools Are Falling Short 

‍

Finding: While AI tool deployment is widespread, only 37% of implementations met or exceeded HR team expectations. Top barriers: data security (61%) and system integration (54%).

What This Means: The HR tech stack isn't ready for AI. Most tools were built for manual workflows, and bolting on AI creates friction rather than efficiency.

Suggestion Actions:

• Prioritize recruiting tools (79% adoption) where AI delivers the clearest value.
• Require AI vendors to demonstrate compliance and integrations before purchasing.
• Start with standalone AI tools rather than waiting for your HRIS to add features.

Final Recommendations

‍

For executives: Make AI proficiency a hiring requirement now, not in six months. The 11-point productivity advantage compounds quickly.

For HR leaders: Fix your interview process disconnect—test for AI capability, not just core skills. And, don't expect your HR tech stack to deliver on AI promises without significant data cleanup.

For finance leaders: Model headcount as output capacity, not just cost. Companies that amplify teams with AI will outpace those cutting headcount for short-term savings.

For everyone: Entry-level displacement is happening faster than most expect. Adjust talent pipelines, compensation strategies, and training programs accordingly.

‍

The content here should not be viewed as investment advice, nor does it constitute an offer to sell, or a solicitation of an offer to buy, any securities. The views expressed here are those of the individual Lightspeed Management Company, L.L.C. (“Lightspeed”) personnel and are not the views of Lightspeed or its affiliates; other market participants could take different views.

‍