Podcast Summary: "AI tools are overdelivering: results from our large-scale AI productivity survey"

Podcast: Lenny’s Reads
Host: Lenny Rachitsky
Date: December 24, 2025

Episode Overview

In this episode of Lenny’s Reads, Lenny Rachitsky shares results from a large-scale, independent survey conducted with Noam Siegel (AI Insights Manager at Figma), exploring how AI tools are truly impacting productivity for tech workers. The survey polled 1,750 respondents across product, engineering, design, and founder roles to uncover not only if AI is delivering productivity gains, but how and where it’s providing value—and where gaps remain. The results paint a picture of AI “overdelivering” on expectations, driving notable time savings and shifting job boundaries, though not without drawbacks and clear opportunities for improvement.

Key Discussion Points & Insights

1. AI Tools Are Surpassing Expectations

• 55% say AI has exceeded expectations
• ~70% report AI improves the quality of work
• AI tools are saving half a day per week or more for a majority—unprecedented vs. prior productivity technologies

"We've never seen a tool deliver a productivity boost like this before." – Lenny (02:23)

2. Role-by-Role Findings

Product Managers (PMs)

• Top Tasks Enhanced: PRD writing (21.5%), mockups/prototyping (19.8%), communication (18.5%)
• AI is mostly helping “production” (writing, prototyping) rather than upstream strategy (user research: 4.7%, roadmap: 1.1%)
• PMs are encroaching on traditional design/engineering work, especially in prototyping

"AI is helping product managers produce, but it lags in helping them think." – Lenny (04:00)

Designers

• Top Tasks: Research synthesis (22.3%), content/copy (17.4%), ideation (16.5%), visual design much lower (3.3%)
• Fewer report a positive ROI from AI (45%) vs. founders (78%)
• Designers see less benefit because AI supports surrounding tasks more than “pushing pixels” (core creative work)
• Notable Dissatisfaction: 31% say AI has fallen below expectations, 3x the rate among founders

Founders

• AI as Thinking Partner: Productivity & decision support (32.9%), product ideation (19.6%), vision/strategy (19.1%)
• Use AI most for upstream/strategic work, unlike other roles
• Highest satisfaction rates—leveraging AI for higher-leverage, strategic gains
• Surprising underuse in financial modeling (1.8%) and recruiting (1.3%)

Engineers

• AI’s Main Job: Code writing (51%)
• Also assists with documentation (7.7%), testing (6.2%), and code review (4.3%)
• Engineers are eager for AI to automate tedious “after-code” tasks (docs, tests)
• Mixed feelings on quality: 51% say AI improves code, but 21% say it decreases quality—the highest “worse” rating of any group

"Engineers have accepted AI as a coding partner. Now they want it to handle the tedious work that comes after the code has been written." – Lenny (07:00)

3. AI Tool Preferences by Role

ChatGPT: #1 for PMs (57.7%), Designers (49.6%), Founders (72.1%)

• Engineers prefer newer tools: Cursor (33.2%), ChatGPT (30.8%), Claude Code (close behind)
• Claude Code now beats the Claude chat interface among engineers
• Specialized, purpose-built AI is outpacing general chatbots in engineering
• Perplexity ranks high for PMs (research strength); prototyping tools (Lovable, Cursor) crack PM’s top 7

"Purpose-built tools are winning, but Claude is also helpful with several core coding related tasks..." – Lenny (10:10)

4. Time and Quality Gains

• Time saved (>4hrs/week): PMs 63%, Founders 83%, Designers 47.5%
• Quality gains: PMs/founders >70% say better, Designers 60%, Engineers 51% better/21% worse
• Lowest time/quality improvements: Designers (both metrics)
• Engineers’ mixed ratings likely linked to code correctness being easier to verify and more critical

"A somewhat better first draft of a PRD is useful, but a somewhat better but buggy function is not." – Lenny (13:30)

5. Wish List: Where Users Want More AI

• PMs: Big demand gap for user research (current 4.7% —> desired nearly 33%), and especially for prototyping (current 19.8% —> 44.4%)
• Designers: Want more in prototyping/interaction design
• Engineers: Want AI for documentation, test writing, code review (demand gaps: 23-25%)
• Founders: Want AI for product ideation (19.6% —> 48.6%), plus market analysis and GTM planning (gaps of ~24%)

"Founders are looking for a strategic collaborator to test ideas, explore markets and think through go-to-market. It's AI as a co-founder, not just an assistant." – Lenny (16:30)

• Conclusion: The next wave will require better workflows, not just smarter models, especially for ambiguous/problem-solving tasks.

6. Product Market Fit: Which Tools Would You Miss?

• 83.6% of respondents would be “very disappointed” if at least one AI tool went away—signals deep integration
• ChatGPT still dominates, but only about half say they’d be “very disappointed” to lose it, lower than regular usage rates (shows high switching/low lock-in)
• Engineers: 3 of top 4 most-missed tools are coding-specific (Cursor, Claude Code, Claude)
• Specialization Wins: Tools like Cursor have deep PM/engineering embedment; designers/founders showing affinity for Claude ecosystem

"Engineers have found and want to hold on to specialized tools that fit their needs rather than relying on general purpose chat interfaces." – Lenny (20:40)

7. Downsides and Open Challenges

• High Prevalence of Issues: 92.4% reported at least one significant downside to current AI tools—plenty of unmet needs, room for improvement
• Caution: “AI is far from the novelty it was...but tools are improving at a breathtaking pace.”

"If AI is already giving most people back at least half a day per week in late 2025, what does 2026 or 2027 look like?" – Lenny (03:00)

Memorable Quotes & Moments

• On unprecedented impact:

  "We've never seen a tool deliver a productivity boost like this before." (02:23)

• On role differences:

  "AI is helping product managers produce, but it lags in helping them think." (04:00)

• On strategic use:

  "Founders are treating AI as a thought partner and sounding board, not just a tool for specific deliverables." (07:45)

• On new opportunities:

  "The next wave of AI adoption will require not just better models, but better workflows for human-AI collaboration on fuzzy problems." (18:30)

Timestamps for Key Segments

• [02:00] – Survey purpose and methodology overview
• [03:00] – Top 7 overall survey insights (“AI is overdelivering”)
• [04:00] – Role-by-role breakdown: Product Managers
• [05:45] – Designers’ results and emerging patterns
• [07:00] – Founders’ and Engineers’ insights, focus on strategic/thinking vs. production tasks
• [10:10] – AI tool preferences, role differences
• [13:30] – Time & quality impact, nuanced by role
• [16:30] – Where users want more from AI—wishlist/demand gap
• [18:30] – Workflows, ambiguity, and future opportunities
• [20:40] – Which tools really have product market fit, per role

Closing Thoughts

Lenny wraps up by emphasizing that AI is no longer a novelty—it’s foundational to knowledge work, with productivity benefits compounding rapidly. Tools are getting better, adoption is deepening, and the biggest opportunities lie in helping people think and collaborate, not just produce.

For a comprehensive appendix of the survey and charts, see the written post (link in show notes).