The Convenience Trap

AI is not saving us time. It is generating work.

The pitch is prolific. AI will handle the documentation, write the reports, will take care of the tedious knowledge work so you can focus on more important things. People who never published anything announce they can finally write, now that the machine does it for them. The framing is staging AI as a convenience technology, and convenience is an unqualified good.

Malls are dying, taken over by strip malls . The pattern is the same one that converted downtown department stores into Dollar Trees and vape shops — what was curated is downgraded for accessible and fast, and the result is more of everything and less of anything worth the trip. AI is the strip mall of knowledge work. It has made production cheap but has not made production good. And the volume it generates is burying the people it was supposed to free.

Photo by wyyz on Unsplash

The AI productivity discourse dances around the question, is there more work now, or less?

The data says more.

The Paradox Returns

Robert Solow won the Nobel Prize in economics for his work on growth theory. In 1987, he made an offhand remark that became one of the most cited sentences in the field: “You can see the computer age everywhere but in the productivity statistics.”¹ Nearly forty years later, the sentence works just as well if you swap in “AI.”

A February 2026 NBER survey of nearly 6,000 senior executives found that 89% of firms report no productivity impact from AI over the past three years.² McKinsey frames it as the gen AI paradox, where roughly 80% of companies use generative AI, but roughly 80% report no significant bottom-line impact.³ U.S. non-farm business productivity grew 2.1% in 2025 — modestly above the 2007–2019 trend of 1.5%, but nowhere near the revolution that was promised by trillions in investment.⁴ In early 2026, Apollo Chief Economist Torsten Slok said plainly that “AI is everywhere except in the incoming macroeconomic data”. ⁵

2026 NBER study

The micro-level studies that do show gains come with caveats that their evangelists omit. Erik Brynjolfsson’s 2023 Stanford and MIT study of 5,172 customer support agents found a 14% productivity increase — but gains were concentrated among novice workers; experienced agents saw minimal improvement.⁶ The MIT experiment by Noy and Zhang, published in Science, reported a 40% reduction in task time — and then the authors themselves warned that their “relatively short, self-contained” tasks “may inflate our estimates of ChatGPT’s usefulness.”⁷ The Harvard/BCG “jagged frontier” study found AI users completed tasks 25% faster when working within AI’s capabilities — but performed 19 percentage points worse on tasks outside that frontier or AI capabilities, and users could not reliably tell which zone they were in.⁸

METR conducted the most rigorous real-world study , which ran a randomized controlled trial with experienced open-source developers. AI tools made them 19% slower, not faster. The perception gap reveals that developers predicted a 24% speedup and believed they were 20% faster, even after the study demonstrated the opposite.⁹ A 39-point gap between perceived and actual performance. That number should concern everyone building organizational strategy on AI productivity assumptions.

METR study

Everyone Can Produce But No One Can Absorb

The market tells us the most measurable result of generative AI is efficiency but the data tells us otherwise. It is efficiency in creating volume. Or as Irina Malkova appropriately characterizes it—abundance.

An Ahrefs analysis of 900,000 newly detected English language web pages in April 2025 found that 74.2% contained AI-generated content.¹⁰ A longitudinal study of 65,000 articles showed AI-written content rising from roughly 10% in late 2022 to over 50% by late 2024.¹¹ Europol has warned that up to 90% of online content may be synthetically generated by 2026.¹² Merriam-Webster named “slop” its 2025 Word of the Year — low-quality AI-generated digital content — with usage of the term increasing ninefold from 2024 to 2025.¹³

And humans are in denial of the fact that AI generated, synthetic content is flooding the web, pointing to the fact that AI detection tools are unreliable, so how would anyone know that content is AI generated? The fallback argument is to argue that AI is an essential tool for expressing ideas, a tool that democratizes the expression of thought through content generation. Regardless of the reasons be it difficulty in reliable AI detection or the democratization of content generation, the facts remain. Synthetic content now outnumbers human generated content on the web. And intended audiences are overloaded with content, where the lines between truth and fiction are being blurred because of cognitive overload.

The flood is impacting every professional domain. GitHub Copilot now generates 46% of all code written by active users.¹⁴ Developers pushed nearly 1 billion commits in 2025, up 25% — but comments on those commits dropped 27%.¹⁵ The production went up. The review did not keep pace. In academia, submissions to NeurIPS jumped 61% in a single year, reaching 21,575 in 2025.¹⁶ A Science study found that researchers using AI posted 52–60% more papers than non-users.¹⁷ AAAI-26 received nearly 30,000 submissions, described as pushing the review system to the point of collapse.

In hiring, LinkedIn reported a 45% year-over-year increase in job applications in 2025.¹⁸ Greenhouse’s survey found that 34% of recruiters now spend up to half their workweek filtering spam and junk applications, with 74% of U.S. job seekers using AI to polish applications.¹⁹ Companies report receiving four to five times more applications than before, most of them indistinguishable because AI built them all from the same job descriptions.

This is the strip mall logic. When it becomes trivially easy to produce a cover letter, a report, a code review, a research paper, a proposal — you do not get fewer of them. You get more. And ultimately, someone has to read them.

The Verification Burden

A March 2026 Foxit/Sapio Research study of 1,400 workers found that end users save approximately 3.6 hours per week using AI but spend 3 hours and 50 minutes reviewing and correcting AI output — a net loss of 14 minutes.²⁰ Executives fared only slightly better with 4.6 hours saved, 4 hours 20 minutes spent validating, for a net gain of 16 minutes per week. Foxit coined the term “verification burden.” Workday’s 2026 research arrived at a similar figure, finding that 37–40% of time saved by AI is consumed reviewing, correcting, and verifying AI output.²¹ Simple logic, given the numbers, should stimulate discussions about the cost of increased output—but they don’t. The industry is still marching towards more output, not what happens to the volumes of output once released into the ether.

Foxit/Sapio study, January 2026

Examining studies that measure quality output data, a clearer picture as to why emerges. Stanford’s RegLab found large language models hallucinate 69–88% of the time on specific legal queries.²² A MedRxiv study reported a 64% hallucination rate on clinical case summaries without mitigation.²³ Even on grounded summarization — AI’s most reliable feature — the best models still hallucinate at a 0.7% rate, and OpenAI’s newest reasoning models paradoxically hallucinate more with GPT o3 at 33% and GPT o4-mini at 48% on factual questions.²⁴ MIT researchers found that when AI models hallucinate, they use 34% more confident language than when providing accurate information — making errors harder to catch.²⁵

In software development, where AI adoption is greatest, the quality data is particularly insightful. CodeRabbit’s analysis of 470 GitHub pull requests found AI-generated code contained 1.7x more issues overall, 1.4x more critical issues and 2.74x more cross-site scripting vulnerabilities.²⁶ Uplevel’s study of approximately 800 developers found a 41% increase in bug rates after Copilot adoption with no significant change in cycle time or throughput.²⁷

GitClear’s 2023 analysis of 153 million changed lines of code found code churn — lines reverted within two weeks — projected to double compared to the pre-AI baseline, with a 4x increase in code duplication.²⁸ The relatively older study is substantiated but the Stack Overflow 2025 Developer Survey which found that of 49,000+ respondents, trust in AI accuracy at an all-time low—only 29%, down from 40% in 2024. Sixty-six percent say they spend more time fixing AI code than they save.²⁹

The machine writes fast while human checks are comparatively slow.

Jevons Was Right

In 1865, William Stanley Jevons observed that James Watt’s more efficient steam engine did not reduce coal consumption, it increased it. When efficiency makes a resource cheaper to use, demand rises to consume the savings and then some. Economists call it the Jevons paradox. It applies to AI and knowledge work just as well as it did to the steam engine and coal consumption.

Jevons paradox

When Satya Nadella tweeted “Jevons paradox strikes again!” after DeepSeek’s January 2025 announcement, he meant it as a bullish signal for AI demand.³⁰ He was right, but perhaps not in the way he intended. The same logic applies to human labor. When AI expedites document production cutting time to create from an hour to ten minutes, the result is more documents, not less.

A February 2026 UC Berkeley Haas ethnographic study published in Harvard Business Review documented this dynamic over eight months at a 200-person technology company. Researchers Aruna Ranganathan and Xingqi Maggie Ye found AI intensification took three forms—workers expanded the scope of what counted as “their job,” work seeped into pauses and off-hours and workers ran multiple AI threads simultaneously. They identified a vicious cycle wherein increased capability leads to increased output, which leads to higher expectations, which results in pressures to further expansion.³¹

Stanford’s Social Media Lab gave calls this phenomena “workslop.” Their September 2025 survey of 1,150 U.S. desk workers found that 40% had received workslop — AI-generated content masquerading as good work but lacking substance — in the previous month. Average resolution time per incident was 1 hour 56 minutes. Estimated annual cost for a 10,000-person organization is $9 million dollars.³² The social costs compound the economic ones with 54% of workers characterizing workslop senders as less creative and 42% as less trustworthy. Ouch.

The historical parallels are everywhere, but we choose to ignore the signals. When ATMs reduced tellers per branch in the 1970s, banks opened 43% more urban branches — and net teller employment rose. Spreadsheets eliminated 400,000 bookkeeping positions but created over 600,000 new accounting roles, because cheap analysis meant every business wanted dramatically more accounting on the cheap.³³

Tim Wu diagnosed the deeper pattern in his 2018 New York Times essay “The Tyranny of Convenience,” noting Betty Friedan’s observation that even with all the new labor-saving appliances, the modern housewife probably spent more time on housework than her grandmother.³⁴

The convenience trap is not so much about a failure in technology as it is a feature of the incentive structure.

Exhausted Users

Upwork’s 2024 survey of 2,500 workers found that 77% of employees using AI report the tools have added to their workload, with 39% spending more time reviewing AI-generated content and 47% saying they have “no idea” how to achieve expected productivity gains.³⁵ The burnout numbers speak for themselves, with 71% of full-time employees reporting burnout, and GenZ with the highest rates at 83%. The Upwork’s 2025 follow-up study discovered the cruelest paradox, finding that 88% of workers reporting the highest AI productivity gains were also the most burned out, and were twice as likely to consider quitting.³⁶

BCG’s research on what it calls “AI Brain Fry” found that workers constantly supervising multiple AI tools experience 12% more mental fatigue and significantly more information overload.³⁷ A 2025 study in the International Journal of Information Management confirmed that AI technostress increases exhaustion and exacerbates work-family conflict, even when it contributes to productivity.³⁸ ManpowerGroup’s 2026 Global Talent Barometer found that across 14,000 workers in 19 countries, AI use increased 13% in 2025, but confidence in AI’s utility plummeted 18%.³⁹

Workers are using more AI while trusting it less. This is what happens when the tool becomes mandatory and the work it generates becomes someone else’s problem, at the expense of humans.

Wharton visiting scholar Cornelia Walther formalized this pattern with the “AI Efficiency Trap” — a four-stage cycle where initial productivity gains trigger managerial recalibration, then dependency acceleration and skill atrophy, and finally performance expectation lock-in.⁴⁰ The endpoint of this pattern is what she calls “agency decay.” Workers become unable to function without the tools that were supposed to augment them, while the tools generate a continuous stream of output that someone must review, edit, and approve.

Walther’s AI Efficiency Trap

The American Psychological Association’s Dennis Stolle described “Employees are as busy as they can possibly be during the work day, and then they’re feeling like they need to spend their evenings learning about AI so that they can keep up and make the next day even more busy.”⁴¹ This is the lived experience of so many workers, and the conditions are being normalized which accelerates the AI productivity paradox.

The Strip Mall Has Won

Microsoft titled its 2025 Work Trend Index report “Breaking Down the Infinite Workday”, an appropriate title given the findings of the report.⁴² The average employee now receives 117 emails and 153 Teams messages per day. Workers are interrupted every 2 minutes during core work hours. After-hours chats rose 15%. Meetings after 8 PM climbed 16%. Forty percent of employees are checking email by 6 AM. While the AI industry calls this engagement, from where the knowledge worker sits, it is an infinite supply of low-grade output demanding finite human attention.

The defenders of the productivity gains claims say this is a J-curve — that general-purpose technologies always depress measured productivity before they deliver gains, that electrification took 30 years and personal computers took 20.⁴³ And so maybe we are in the valley of depression that comes before immense productivity. But will those productivity gains be worth the current quality and volume dearth?

Daron Acemoglu, the 2024 Nobel laureate in economics, estimates AI will boost total factor productivity by no more than 0.53–0.66% over an entire decade — roughly 0.05% per year — because only about 5% of economy-wide tasks can be profitably automated.⁴⁴ Set that number against the scale of current corporate AI expenditure and the J-curve comes into sharp focus. Firms across every sector are pouring capital into infrastructure, platform licensing, integration engineering and workforce restructuring, all in pursuit of productivity gains that Acemoglu's analysis suggests will be marginal at best.

The gap between what organizations are spending now and what they can reasonably expect to recoup is the downward slope of the J-curve evidenced in balance sheets. Investment is running far ahead of measurable output, which is what is described by this J-curve pattern—a period of net loss driven by costly reorganization, followed by a delayed and often underwhelming recovery. Acemoglu's work presents a constraint on the recovery from the era of AI overabundance.

The classic J-curve narrative promises that the trough is temporary and the upswing will eventually justify the investment. But if only 5% of tasks are profitably automatable, the upswing has a very low ceiling. The fallacy becomes that organizations are experiencing a lag between investment and return. The reality is many of them may be investing in excess of any return that will ever materialize. The J-curve, in other words, is real, but flatter than its evangelists need it to be.

Developers are losing valuable time to non-coding tasks: 50% report losing 10+ hours per week, and 90% lose 6+ hours or more, largely due to organizational inefficiencies.

The evidence to date suggests most organizations cannot resist the seductive allure of increased productivity yet ignore the cost of any surface-level productivity gains. As Atlassian’s research concluded, “Without intentional reallocation, savings get absorbed into increased email volume, delivering no net productivity gains.”⁴⁵ Seems organizations do not want to face this reality, perhaps because AI is an arms race, with high-stake claims outweighing the reality on the ground.

The truth is that AI is a convenience technology. It has made producing knowledge artifacts trivially easy yet it has not made generated artifacts good. Research concludes that people in the AI productivity loop are not happier or less busy. In fact, data shows that It has made for more work, more busyness. Now humans are drowning in more of everything—documents, code, applications, proposals and messages. The volume of content requiring human attention has exploded yet the hours available for that attention have not.

---

Footnotes

1. Robert Solow, “We’d Better Watch Out,” New York Times Book Review, July 12, 1987. The original observation was about computers; it has been widely applied to subsequent waves of information technology.

2. Nicholas Bloom et al., NBER Working Paper, February 2026. Surveyed 5,937 senior executives across industries. Reported in Fortune, February 17, 2026; Rich Turrin, “NBER: 80% of Companies Report No Productivity Gain from AI Despite Billions in Investment,” Substack, 2026. See also Irving Wladawsky-Berger, “The AI Productivity Paradox,” October 2025.

3. McKinsey framing as reported in the NBER study context and Fortune coverage of the February 2026 CEO survey. The “80/80” formulation — 80% adoption, 80% no impact — circulated widely in early 2026.

4. U.S. Bureau of Labor Statistics, “Fourth Quarter and Annual Averages 2025, Revised,” Productivity and Costs report, 2026.

5. Torsten Slok, Apollo Global Management, as quoted in Fortune, February 17, 2026. The St. Louis Fed’s measurement of 1.9 percentage points cumulative excess productivity growth since late 2022 is reported in Irving Wladawsky-Berger, “The AI Productivity Paradox,” October 2025.

6. Erik Brynjolfsson, Danielle Li, and Lindsey Raymond, “Generative AI at Work,” NBER Working Paper No. 31161, 2023. Published findings based on a deployment of an AI conversational assistant at a Fortune 500 company’s customer support operation.

7. Shakked Noy and Whitney Zhang, “Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence,” Science 381, no. 6654 (2023): 187–192. The authors’ caveats about task design appear in the paper’s discussion section.

8. Fabrizio Dell’Acqua et al., “Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality,” Harvard Business School Working Paper, 2023. Published in Organization Science, 2025. Study conducted with 758 BCG consultants using GPT-4.

9. METR, “Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity,” July 2025. Randomized controlled trial. The 19% slowdown was measured against a control group without AI access; the 39-point perception gap (24% predicted speedup + 20% perceived speedup vs. 19% actual slowdown) is derived from the study’s self-report data. See also arXiv:2507.09089.

10. Ahrefs analysis of 900,000 newly detected English-language web pages, April 2025. Reported in Stan Ventures, “74% of New Web Pages Now Contain AI-Generated Content,” 2025.

11. Graphite longitudinal study of 65,000 articles, as reported in multiple 2025 analyses of AI content prevalence. See also Futurism, “Over 50 Percent of the Internet Is Now AI Slop, New Data Finds,” 2025.

12. Europol projection, cited in Euronews, “2025 Was the Year AI Slop Went Mainstream,” December 28, 2025. See also Meltwater, “What the Rise of AI Slop Means for Marketers,” 2025.

13. Merriam-Webster, 2025 Word of the Year announcement. Usage increase reported in Euronews, December 2025.

14. GitHub Octoverse 2025 report. See Gajanan Chandgadkar, “Reading Between the Commits: What GitHub’s Octoverse 2025 Tells Us About the Next Developer Decade,” Medium, 2025.

15. Ibid. The 27% drop in code comments alongside 25% growth in commits is drawn from the same Octoverse 2025 data.

16. NeurIPS 2025 submission data. See IntuitionLabs, “NeurIPS 2025: A Guide to Key Papers, Trends & Stats,” 2025. See also arXiv:2412.07793, “Publication Trends in Artificial Intelligence Conferences: The Rise of Super Prolific Authors.”

17. Reported in Euronews, “Scientists Are Publishing More Than Ever with AI. But Not All Papers Measure Up, Study Finds,” December 31, 2025. See also Phys.org coverage of the same Science study.

18. LinkedIn data reported in Greenhouse, “An AI Trust Crisis: 70% of Hiring Managers Trust AI to Make Faster and Better Hiring Decisions, Only 8% of Job Seekers Call It Fair,” 2025.

19. Greenhouse 2025 hiring survey. The “AI doom loop” framing comes from Fortune, “Hiring Platform CEO Says Talent Acquisition Is in an ‘AI Doom Loop,’” November 18, 2025.

20. Foxit/Sapio Research, March 2026, survey of 1,400 knowledge workers. Reported in IT Brief Asia, “AI Productivity Gains Erased by Document Verification,” 2026, and BetaNews, “AI at Work Saves Executives Just 16 Minutes a Week, While Employees Lose 14 Minutes,” 2026.

21. Workday 2026 research on AI verification overhead, as reported in multiple enterprise technology publications.

22. Stanford RegLab, hallucination rate study on legal queries. Reported in Suprmind, “AI Hallucination Statistics: Research Report 2026,” 2026.

23. MedRxiv study on clinical case summary hallucination rates. Reported in Suprmind and All About AI, “AI Hallucination Report 2026,” 2026.

24. Hallucination benchmarking data from multiple sources. The o3 (33%) and o4-mini (48%) figures are from 2026 benchmarks. The 34% increase in confident language during hallucinations is from MIT research. Reported in Suprmind, “AI Hallucination Rates & Benchmarks in 2026,” and All About AI, 2026.

25. Ibid. MIT research on confidence calibration in language model outputs.

26. CodeRabbit analysis of 470 GitHub pull requests. Reported in The Register, “AI-Authored Code Needs More Attention, Contains Worse Bugs,” December 17, 2025.

27. Uplevel, “Gen AI for Coding Research Report,” 2024. Study of approximately 800 developers before and after Copilot adoption. See also Computing, “New Study Shows Minimal Impact on Developer Productivity,” 2024.

28. GitClear, “Coding on Copilot: 2023 Data Suggests Downward Pressure on Code Quality,” analysis of 153 million changed lines of code including 2024 projections

29. Stack Overflow 2025 Developer Survey, 49,000+ respondents. See Stack Overflow Blog, “Developers Remain Willing but Reluctant to Use AI,” December 29, 2025. See also ADTmag, “Developers Lean on AI More, But Report Growing Doubts About Accuracy,” January 2026.

30. Satya Nadella tweet following DeepSeek’s January 2025 announcement. The Jevons paradox framework as applied to AI is discussed extensively in Northeastern Global News, “What Is Jevons Paradox? And Why It May — or May Not — Predict AI’s Future,” February 7, 2025.

31. Aruna Ranganathan and Xingqi Maggie Ye, UC Berkeley Haas School of Business, ethnographic study published in *Harvard Business Review*, February 2026. See Berkeley Haas Newsroom, “AI Promised to Free Up Workers’ Time. UC Berkeley Haas Researchers Found the Opposite,” 2026. The Manos quote appears in the same coverage.

32. Jeffrey Hancock et al., Stanford Social Media Lab, September 2025. Survey of 1,150 U.S. desk workers. Reported in Axios, “AI ‘Workslop’ Is Crushing Workplace Efficiency, Study Finds,” September 24, 2025; Entrepreneur, “AI Workslop Is a $9 Million Issue: Stanford, BetterUp Study,” 2025; Futurism, “Companies Are Being Torn Apart by AI ‘Workslop,’ Stanford Research Finds,” 2025.

33. The ATM/bank teller and spreadsheet/bookkeeper examples are standard illustrations of the Jevons paradox in labor markets, discussed in 367 Ventures, “Jevons Paradox and the AI Workforce: Why Efficiency Creates More Demand, Not Less,” 2025, and MindStudio, “What Is Jevons Paradox in AI?” 2025.

34. Tim Wu, “The Tyranny of Convenience,” *New York Times*, February 16, 2018. The Friedan reference appears in the essay. See also Open Markets Institute republication.

35. Upwork, “Employee Workloads Rising Despite Increased C-Suite Investment in Artificial Intelligence,” July 23, 2024. Survey of 2,500 workers including C-suite executives, freelancers, and full-time employees.

36. Upwork, “New Insights Into the AI-Human Work Dynamic,” 2025. The finding that the most AI-productive workers were the most burned out (88%) and twice as likely to consider quitting is from this follow-up study.

37. BCG “AI Brain Fry” research, as reported in CNBC, “Using AI Can Add Extra Labor and Cause ‘Brain Fry’ for Workers, Experts Say,” April 6, 2026.

38. “Insights from the Job Demands–Resources Model: AI’s Dual Impact on Employees’ Work and Life Well-Being,” International Journal of Information Management, 2025.

39. ManpowerGroup, 2026 Global Talent Barometer, surveying 14,000 workers in 19 countries.

40. Cornelia Walther, “The AI Efficiency Trap: When Productivity Tools Create Perpetual Pressure,” Wharton, Knowledge@Wharton, 2025–2026.

41. Dennis Stolle, American Psychological Association, as quoted in CNBC, April 6, 2026.

42. Microsoft, “Breaking Down the Infinite Workday,” Work Trend Index, 2025. Email/Teams message volume, interruption frequency, and after-hours work data are from Microsoft’s M365 telemetry. See also Microsoft, “2025: The Year the Frontier Firm Is Born,” Work Trend Index, 2025.

43. Erik Brynjolfsson, Daniel Rock, and Chad Syverson, “The Productivity J-Curve: How Intangibles Complement General Purpose Technologies,” American Economic Journal: Macroeconomics 13, no. 1 (2021). The electrification and PC analogies are standard in this literature. See also Brynjolfsson and colleagues, “Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics,” NBER Working Paper No. 24001, 2017.

44. Daron Acemoglu, “The Simple Macroeconomics of AI,” Economic Policy 40, no. 121 (2025): 13. See also MIT Technology Review, “A Nobel Laureate on the Economics of Artificial Intelligence,” February 25, 2025.

45. Atlassian research on productivity reallocation, as cited in analyses of AI time savings absorption. See also Superhuman, “How to Use an AI Email Assistant to Take Control of Your Inbox,” which discusses the pattern of savings being absorbed into increased communication volume. https://www.atlassian.com/blog/developer/developer-experience-report-2025

---

about me. I’m a Semantic Engineer, Information Architect, and knowledge infrastructure strategist dedicated to building information systems. With more than 25 years of experience in enterprise architecture, e-commerce content systems, digital libraries, and knowledge management, I specialize in transforming fragmented information into coherent, machine-readable knowledge systems.

I am the founder of the Ontology Pipeline™, a structured framework for building semantic knowledge infrastructures from first principles. The Ontology Pipeline™ emphasizes progressive context-building: moving from controlled vocabularies to taxonomies, thesauri, ontologies, and ultimately fully realized knowledge graphs.

Professionally, I have led semantic architecture initiatives at organizations including Adobe, where I architected an RDF-based knowledge graph to support Adobe’s Digital Experience ecosystem, and Amazon, where I worked in information architecture and taxonomy. I am also the founder of Contextually LLC, providing consulting and coaching services in ontology modelling, NLP integration, knowledge graphs and knowledge infrastructure design.

I am also a curriculum designer, teacher and founder of The Knowledge Graph Academy, a cohort-based educational program designed to train and up skill future semantic engineers and ontologists. The Academy is the the perfect balance of ontology and knowledge graph theory and practice, preparing graduates to confidently work as ontologist and semantic engineers.

An educator and thought leader, I publish regularly on my Substack newsletter, Intentional Arrangement, where my writing frequently explores the relationship between semantic systems and AI.

Connect with me on LinkedIn-!

---

More from Intentional Arrangement