Author name: Catalyst Marketing

Ask an AI tool to write the report, fix the code, or draft the pitch, and it will hand back something polished in seconds. Ask the person who submitted that work to explain why it’s structured that way, what assumption it’s built on, or what happens if one input changes, and the answer often falls apart. AI vs human explanation of work has quietly become one of the sharpest tests in offices, classrooms, and interview rooms in 2026, and a growing stack of research suggests a lot of people would fail it. The Interview Question That Exposes This Directly The clearest, most literal version of this test now happens in job interviews themselves. Meta began rolling out AI-enabled coding interviews in late 2025, handing candidates an assistant like Cursor, Copilot, or Claude Code instead of a blank editor, and companies including Shopify, Rippling, LinkedIn, Canva, and Uber have adopted similar formats through 2026. The AI is allowed and expected. What’s being graded isn’t whether the candidate can produce working code; it’s whether they can direct, verify, and explain it. The failure pattern shows up consistently across these interviews: at Canva, interviewers pause after each AI generation and ask candidates to walk through what the code does, and a Rippling candidate was explicitly rejected for relying too heavily on AI even though their initial approach was correct, because they let the AI make the implementation decisions instead of directing them. Getting the right answer stopped being the bar. Being able to explain and defend the answer became the actual bar. Botshitting: Shipping Work You Can’t Defend There’s now a specific term for the pattern this creates outside interview rooms, too. A 2026 report from Glean found that nearly 7 in 10 AI users admit to what researchers call botshitting, shipping AI-generated work they haven’t reviewed, don’t fully understand, or couldn’t defend if asked, with heavy AI users, Gen Z employees, men, and managers most likely to do it. The same report found digital workers now spend nearly a full workday every week botsitting, checking outputs, debugging mistakes, and fixing confidently wrong answers, a workload that itself seems to be pushing people toward skipping the review step entirely rather than adding more of it. When the Work Fails, Who Actually Takes the Blame The gap between doing the work and explaining it has a psychological cost that researchers can now measure directly. When AI-generated work fails, 40 per cent of workers blame the AI, while only 29 per cent admit it was their own fault, a pattern researchers describe as moral disengagement, the gradual process by which people stop holding themselves accountable for outcomes they had a hand in. A related Boston Consulting Group experiment found something that makes this worse rather than better: when AI was framed as an employee rather than a tool, workers felt less accountable for what it produced and reviewed its output less carefully. The more AI gets treated like a colleague, the less anyone feels obligated to explain what it actually did. The Accountability Gap Shows Up at the Company Level Too This isn’t only an individual habit problem. It’s showing up in how organisations track their own AI-generated output. Lanai’s 2026 AI Labour Report found that 92 per cent of leaders say their organisation tracks the financial and efficiency impact of AI-generated work, but only 2 per cent say more than half of that work is actually recorded as a measurable business outcome, a pattern researchers call AI labour orphaning. Separately, Grant Thornton’s 2026 AI Impact Survey found that 78 per cent of business executives lack strong confidence they could pass an independent AI governance audit within 90 days. If leadership can’t reconstruct how an AI system reached a decision, asking an individual employee to explain and defend it becomes an almost impossible position to put them in. Frontline Workers Are Explaining Decisions They Didn’t Make A July 2026 Harvard Business Review study, based on multi-year fieldwork across banking, recruitment, and biotechnology, found that frontline employees are increasingly expected to communicate, justify, and defend AI-generated decisions they neither created nor fully understand. The researchers found workers respond in one of three ways: they relay the AI’s output verbatim without really owning it, they mask it by presenting it as their own reasoning even when it isn’t, or they amplify and complement it by adding genuine judgment on top. Only the third pattern actually closes the gap between AI vs human explanation of work in any meaningful way, and it’s also the pattern that takes the most deliberate effort to practice. The Same Pattern Is Showing Up in Classrooms This isn’t strictly a workplace phenomenon either. The interview trend at Meta, Canva, and Rippling has a direct parallel in how universities and bootcamps have started grading assignments: rather than banning AI outright, instructors increasingly allow it and then grade the student’s ability to walk through and defend what the tool produced, on the logic that this is now the actual skill being hired for. Reports from hiring managers running technical interviews describe a consistent tell: a candidate who produces an optimal solution unusually quickly but can’t walk through even one worked example, or fails when asked pointed questions about specific lines of their own submitted code. Several interviewers have said plainly that when a candidate can’t explain their own logic, it doesn’t matter whether AI was involved or not: the explanation gap alone is the disqualifying factor. How to Actually Close the Gap None of these point toward using AI less. It points toward a specific habit that’s easy to skip under deadline pressure: before submitting anything AI-assisted, being able to state in one or two sentences why it’s structured the way it is, what it assumed, and what would break if a key input changed. That single check is close to what Canva’s interviewers are actually testing for, and it’s the difference between directing a tool and simply forwarding its output. Teams that build

Most new exporters find the product, find a buyer, and only then discover how much paperwork sits between those two things and an actual bank credit. Export compliance in India isn’t a single registration or a single authority. It’s a stack of requirements from DGFT, customs, the GST department, and the Reserve Bank of India that all have to align correctly, and missing any one of them can hold a shipment at port or block payment from reaching an exporter’s account. IEC Registration: The Non-Negotiable Starting Point Nothing else on this list matters until this one is done. The Importer Exporter Code, a 10-digit number issued by the Directorate General of Foreign Trade, is mandatory for any commercial export from India, costs a one-time fee of ₹500, and is typically approved within 1 to 3 working days through the DGFT portal. A sole proprietor can apply using personal PAN and Aadhaar without setting up a separate legal entity, and the code itself has lifetime validity. The part that trips up busy exporters is what comes after approval: DGFT requires an annual profile update between April and June every year, and skipping it gets the IEC deactivated, which blocks customs clearance and any RoDTEP incentive claims until it’s fixed. AD Code and ICEGATE: The Operational Layer This is where export compliance in India moves from paperwork to logistics. An IEC alone doesn’t let a shipment move. Every exporter also needs to register their bank’s Authorised Dealer Code, commonly called AD Code registration, separately at each port they plan to ship from, since it’s a one-time registration per port rather than a single nationwide filing. Shipping bills and bills of entry then get filed through ICEGATE, the Indian Customs EDI Gateway, which is where customs actually processes and clears an export consignment. Without both the AD Code on file at the relevant port and an active ICEGATE registration, a shipping bill simply cannot be filed, regardless of how complete the IEC and product documentation already are. GST and the LUT That Most New Exporters Miss GST is another piece of export compliance in India that trips up new exporters through omission rather than a wrong filing. Exports of goods and services are treated as zero-rated supplies under GST, which means no GST is charged on the export invoice itself. The step that gets skipped is the paperwork that makes this automatic: filing a Letter of Undertaking, or LUT, once at the start of each financial year, which allows an exporter to ship goods or services without paying IGST upfront and then claiming it back later. An exporter who never files the LUT either has to pay IGST on every export and claim a refund afterwards, tying up working capital for weeks, or risks GST department queries over an export invoice with no LUT reference on record. The FEMA Deadline That Keeps Moving, and Why the Date Matters This is the part of export compliance in India that’s genuinely in flux right now, and getting the timeline wrong has real financial consequences. As of a Reserve Bank of India amendment dated 5 June 2026, the deadline to realise and repatriate export proceeds currently stands at 9 months from the date of shipment, reverting from the more relaxed 15-month window that had applied through most of late 2025 and early 2026. That 9-month rule is not permanent either: a fully new, consolidated FEMA (Export and Import of Goods and Services) Regulations, 2026 takes effect from 1 October 2026, restoring a 15-month standard window, extended to 18 months where the export is invoiced and settled in Indian rupees. Which rule applies depends on the shipment date, not the date an exporter happens to be checking the rule, so a shipment made in August 2026 sits on the tighter 9-month clock even though the friendlier 15-month rule is only weeks away. Penalties for Missing the FEMA Realisation Window Missing this deadline isn’t a paperwork inconvenience. Failing to realise and repatriate export proceeds within the applicable window is legally classified as a FEMA contravention, and penalties can run up to three times the unrepatriated amount. The e-BRC, or electronic Bank Realisation Certificate, issued once payment is confirmed, is the document that proves compliance and closes out the shipment in the bank’s and DGFT’s tracking systems, and it also directly affects whether a GST refund claimed under LUT stays valid or gets reversed if proceeds arrive late. RCMC and RoDTEP: The Incentive Side of Compliance Not every registration inside export compliance in India is about avoiding a penalty. A Registration cum Membership Certificate, issued by the relevant Export Promotion Council or Commodity Board for a given product category, isn’t mandatory to export, but it becomes essential the moment an exporter wants to claim benefits under DGFT schemes, most notably RoDTEP, which refunds embedded duties and taxes that aren’t already refunded through other channels. Exporters who skip RCMC registration because their first few shipments went out fine often only discover the gap when they try to claim an incentive months later and find they were never eligible to begin with. Product-Specific Rules Layered on Top Every product category can add its own compliance layer on top of the general export requirements above. Food products need FSSAI licensing before export, and every packaged product still has to meet Legal Metrology labelling rules even when it’s headed out of the country rather than sold domestically. Certain categories, chemicals, pharmaceuticals, defence-adjacent items, and a list DGFT maintains and updates periodically, need specific export licenses or fall under restricted or prohibited categories entirely. Checking a product’s ITC-HS classification against DGFT’s current restricted list before the first shipment, rather than after a container gets held, is one of the cheapest compliance steps available and one of the most commonly skipped. Building a Checklist That Actually Gets Followed For a new exporter starting from zero, the practical order is IEC first, since nothing else can proceed without it, followed by AD Code registration at

Every founder has heard some version of the same compliment early on: this is a genuinely good idea. Few hear the follow-up that actually matters more. The business idea vs business execution debate isn’t really a debate once the data is laid out. An idea decides what gets built. Execution determines whether it survives contact with real customers, real cash flow, and real competitors, and the numbers on business failure make it clear which one carries the most weight. Business Idea vs Business Execution: Why the Idea Gets All the Attention Ideas are the part of a business that’s fun to talk about at a dinner table. They’re also, in almost every practical sense, cheap. Anyone can have a good idea. Building a working product, pricing it correctly, finding the first hundred paying customers, keeping cash flowing while revenue is still unpredictable- none of that shows up in the initial pitch, and none of it is optional. The gap between having a good idea and running a business built on that idea is where most of the actual work, and most of the actual failure, happens. What the Failure Data Shows About Business Idea vs Business Execution Product-Market Fit and Cash: The Execution Side of Failure CB Insights has tracked why startups fail for over a decade, and its most recent analysis widened the picture considerably. A 2024 update analysing 431 failed venture-backed companies found that 43 per cent failed because of poor product-market fit, essentially an execution failure in understanding and serving the market, while running out of cash was cited in about 70 per cent of post-mortems but flagged explicitly as a symptom rather than a root cause. In other words, the company didn’t die because the idea was bad. It died because nobody executed the process of figuring out, early and cheaply, whether the market actually wanted what was being built. Team and Competition: Where Execution Gaps Show Up Next Why Business Execution Failures Outnumber Idea Failures The same dataset breaks down the reasons further, all pointing in the same direction: 23 per cent of failures were attributed to not having the right team, and 19 per cent to being outcompeted by a rival that executed better. None of these is an idea problem. They’re operating problems, hiring problems, and speed problems, the parts of running a business that a great concept does nothing to solve on its own. What Investors Actually Bet On: Business Idea vs Business Execution The 2020 VC Survey Behind the Numbers If anyone had an incentive to bet purely on ideas, it would be venture capitalists, whose entire business model depends on picking winners early. A widely cited 2020 academic survey of 885 institutional venture capitalists across 681 firms found that when asked to name the single most important factor in an investment decision, 47 per cent chose the founding team, compared with just 13 per cent for the product itself, 10 per cent for the business model, and 8 per cent for the target market. Ninety-five per cent of respondents said the team was essential regardless of what else they weighed. Kleiner Perkins, one of Silicon Valley’s oldest venture firms, has summarised this instinct plainly for years: they’d rather back a strong team with an average idea than a weak team with a great one. This isn’t investors undervaluing ideas. It’s investors who have watched enough companies rise and fall to know that a mediocre idea in the hands of a team that executes well gets fixed, refined, and pivoted into something that works. A great idea in the hands of a team that can’t execute usually just runs out of runway more slowly than expected. Executing the Wrong Things Well Is Its Own Trap Execution isn’t automatically the antidote to a weak idea, either, and one specific failure pattern proves it. Startup Genome’s research has found that 74 per cent of high-growth startups fail due to premature scaling, spending aggressively on hiring, marketing, or geographic expansion before the core business model was actually validated, while companies that scale in step with validated demand grow roughly 20 times faster than those that don’t. These are founders executing hard, hiring fast, spending fast, moving fast, on a plan that hadn’t earned that pace yet. Effort and execution only compound in the right direction once there’s something real underneath them to scale. Small Businesses Aren’t Exempt From This Either It’s tempting to treat all of this as a venture-backed tech problem, something that only applies to startups chasing a billion-dollar exit. The underlying pattern shows up just as clearly in ordinary small business survival data. Bureau of Labour Statistics figures show that about 20 per cent of new US businesses close within their first year, 49 per cent within five years, and 65 per cent within ten, across every industry, not just high-growth tech. A local service business, a small manufacturer, or a neighbourhood retailer faces the same fundamental test as a funded startup: whether the person running it can manage cash, pricing, hiring, and customer acquisition well enough to survive the ordinary friction of operating, regardless of how good the original concept was. Survival rates do vary meaningfully by sector, which itself says something about execution. Agriculture, forestry, and fishing businesses post the lowest first-year failure rate of any BLS-tracked category, largely because those industries run on decades of established, well-understood operating practices. The information sector, where business models change constantly and the ground never stays still long enough to master, shows the highest ten-year failure rate. The idea matters less here than the discipline required to keep executing correctly as conditions shift. What ‘How Business Really Works’ Actually Means Underneath the statistics is a fairly unglamorous list of skills that idea-stage excitement tends to skip entirely: reading a cash flow statement well enough to know how many months of runway remain, pricing a product so margin actually survives contact with real costs, hiring the third and fourth employee correctly instead

Type a question into any AI tool today, and an answer appears in seconds: a diagnosis to consider, a stock to research, a hiring candidate to shortlist, a strategy to weigh. AI can give you answers faster and more fluently than almost any resource that came before it. What it can’t do is decide for you, not really, and a growing body of 2026 research suggests that the more we let it try, the weaker our own decision-making muscles get. An Answer Is Not a Decision It’s worth separating these two things clearly, because AI tools are built to blur the line between them. An answer is information: a prediction, a summary, a recommendation generated from patterns in data. A decision involves weighing that information against context the system doesn’t have, your risk tolerance, your specific constraints, the parts of the situation that never made it into the prompt, and then owning the outcome. AI is extraordinarily good at the first part. It has no stake in, and often no visibility into, the second. This distinction sounds obvious stated plainly, but it disappears fast in practice. A confident, well-formatted answer feels like a conclusion, not an input, and that feeling is exactly where most of the risk in this conversation actually lives. What Happens When Judgment Stops Getting Exercised A 2025 MIT Media Lab study, Your Brain on ChatGPT, used EEG to track 54 participants writing essays with either an AI assistant, a search engine, or no tool at all, and found that the AI-assisted group showed the weakest neural connectivity of the three, plus the poorest recall of their own writing 24 hours later, remembering roughly a third as much of their own argument as the unaided group. The researchers called this pattern cognitive debt: borrowing mental effort from a task now, and paying for it later in weaker recall and weaker independent reasoning. It’s worth noting this is a preprint with a modest sample size, but it lines up with a broader pattern showing up across other studies. That broader pattern includes a 2025 study of 666 participants that found a statistically significant link between frequent AI tool use and lower scores on standardised critical thinking tests. Separate systematic reviews of AI use in education report the same direction of effect: heavier reliance on conversational AI correlates with reduced ability to analyse information independently, question assumptions, and construct a reasoned argument without assistance. None of this means AI makes people incapable of thinking. It means thinking, like any skill, atrophies with disuse, and outsourcing it by default is a real cost, not a neutral convenience. Automation Bias: Trusting the Machine Even When It’s Wrong There’s a well-documented psychological pattern called automation bias, the tendency to favour an automated system’s output over contradictory evidence, including your own correct judgment. A 2025 study of pathology experts found that out of 560 AI-assisted diagnostic estimates, practitioners abandoned their own initially correct assessment in favour of incorrect AI advice in 38 cases, about 7 per cent of the time, even though these were trained specialists evaluating their own area of expertise. Time pressure didn’t make this more frequent, but it did make it more severe when it happened. This isn’t limited to medicine. Cockpit simulator studies have found that more than half of professional pilots either missed important information or made dangerous errors when automated systems failed to flag a problem or actively fed them wrong data, a factor investigators cited in real crashes including Eastern Air Lines Flight 401 and Air France Flight 447. In finance, an automated trading algorithm at Knight Capital executed roughly 440 million dollars in unintended trades in about 45 minutes in 2012 before anyone intervened. And in the justice system, an independent investigation into the COMPAS risk-assessment algorithm, used by courts to help judges gauge reoffending risk, found it was wrong 44 per cent of the time when flagging Black defendants as high risk, a statistic that stayed embedded in real sentencing decisions until it became public. The Order You Consult AI In Actually Changes the Outcome None of this means AI assistance makes decisions worse by default. The research on physicians using large language models for diagnosis points to something more specific: one study found diagnostic accuracy improved by 18 per cent when physicians formed their own initial assessment first and only then reviewed an AI system’s suggestion, compared to letting the AI weigh in before the physician had reasoned through the case independently. The AI wasn’t the problem or the solution on its own. The sequence was. Forming a view first turns the AI’s answer into a check on your reasoning. Consulting it first turns your reasoning into a rubber stamp on its answer. Where This Shows Up Outside the Lab The same pattern plays out in ordinary business and personal decisions, just with less dramatic stakes and far less scrutiny. A manager asks an AI tool to evaluate two vendor proposals and takes the recommendation without reading either proposal closely themselves. A job seeker lets an AI-optimized resume and cover letter represent them in an interview they haven’t actually prepared for. A small business owner accepts a pricing strategy an AI generated from a handful of prompts, without checking it against their actual margins or their customers’ actual price sensitivity. In each case, the AI answer was probably reasonable. Whether it was right for that specific situation was never actually tested, because nobody checked. These situations rarely announce themselves as risky. Nobody sits down and decides to skip due diligence. The AI answer simply arrives fast enough, and sounds finished enough, that the step of checking it starts to feel redundant. That’s precisely the moment automation bias takes hold, not as a dramatic lapse in judgment, but as a quiet decision to not bother verifying something that already looks correct. Why Confidence in the Answer Isn’t Evidence It’s Correct Part of what makes this hard is that AI-generated answers tend to read

Most new sellers open a store on Amazon, Flipkart, Meesho, or their own Shopify site the same way they’d open a lemonade stand: list the product, set a price, wait for the first order. The GST FSSAI legal checklist that actually governs online selling in India rarely gets read until a platform account gets suspended, a shipment gets flagged, or a notice shows up in the mail. None of these requirements is hidden. They’re just spread across different departments, and almost nobody explains them together before a seller needs them. GST Registration Is Not Optional, Even for Tiny Sellers For an offline business, GST registration only becomes mandatory once turnover crosses ₹40 lakh for goods or ₹20 lakh for services in most states. Online selling doesn’t get that runway. Under Section 24 of the CGST Act, anyone selling goods through an e-commerce operator like Amazon, Flipkart, or Meesho must register for GST regardless of turnover, even if total annual sales are a few thousand rupees. Selling through your own website counts too, since a self-run online store still falls under the legal definition of e-commerce. There’s a second detail that trips up a lot of new sellers: GST composition scheme taxpayers, who pay a flat rate and file simplified returns, are barred under Section 10(2)(d) from selling through an e-commerce operator at all. A seller who registered under the composition scheme thinking it would simplify things often finds out only after their platform onboarding gets rejected. TCS: The Deduction That Confuses Almost Every First-Time Seller Every e-commerce operator is required to collect Tax Collected at Source on the net value of taxable supplies made through its platform. This TCS rate was halved from 1 per cent to 0.5 per cent effective July 2024, and it still applies in 2026. Amazon, Flipkart, or Meesho deduct this amount before paying a seller out and deposit it with the government on the seller’s behalf, filing GSTR-8 to report it. The part sellers usually miss isn’t the deduction itself; it’s the credit. The TCS collected shows up in the seller’s GSTR-2B only after the platform files its return, and it has to be separately claimed as a credit in GSTR-3B to actually offset GST liability. Sellers who don’t reconcile this every month often overpay GST without realising there was already a credit sitting unclaimed in their account. FSSAI Isn’t Only for Restaurants and Food Manufacturers Anyone selling packaged snacks, spices, pickles, sweets, dry fruits, tea, coffee, or home-baked goods online needs an FSSAI registration or license before listing a single product, and this is one of the most commonly skipped steps by home-based and D2C sellers. FSSAI operates on a three-tier system: Basic Registration for businesses with turnover up to ₹12 lakh, a State License for turnover between ₹12 lakh and ₹20 crore, and a Central License above that, or for specific categories like food importers, exporters, and nationwide e-commerce food businesses regardless of turnover. Applications go through the FoSCoS portal, which replaced the older licensing system and now handles the entire process digitally. Marketplaces increasingly enforce this at the listing stage rather than waiting for a complaint. Swiggy, Zomato, and Amazon’s grocery categories generally won’t activate a food listing without a valid, current FSSAI number tied to it, which means the registration has to happen before launch, not after the first sale. Legal Metrology: The Labelling Rules Nobody Reads Until a Product Gets Delisted Every packaged product sold online in India, not just food, has to comply with the Legal Metrology (Packaged Commodities) Rules, 2011. Mandatory declarations include the manufacturer or packer’s name and address, net quantity, manufacturing date, MRP inclusive of all taxes, and consumer care details, and both the seller and the platform share legal responsibility for getting these right. Selling above the declared MRP, or altering or smudging a printed price, is a punishable offence under the Act, not just a platform policy violation. This area is getting stricter rather than looser. The Legal Metrology (Packaged Commodities) Amendment Rules, 2026, notified in February 2026 and effective from 1 July 2026, require every e-commerce entity selling imported products to make the country of origin searchable and sortable in product listings, not just printed somewhere on the packaging. Any seller sourcing inventory from overseas, even through a domestic wholesaler, needs to know exactly where each product originates well before this obligation kicks in. Consumer Protection Rules Put Real Obligations on the Seller, Not Just the Platform It’s easy to assume Amazon or Flipkart absorbs all the consumer-facing legal risk. They don’t. Under the Consumer Protection (E-Commerce) Rules, 2020, sellers themselves must display accurate pricing with a full breakup, expiry dates, country of origin, warranty and guarantee terms, and clear return and refund policies for every listing, and platforms are required to maintain a written contract with each seller confirming this. Refusing a legitimate return, posting fake reviews, or misrepresenting a product’s condition exposes the seller directly to liability under the Consumer Protection Act, 2019, separate from whatever the marketplace’s internal policy says. Marketplaces are also required to appoint a grievance officer who acknowledges complaints within 48 hours and resolves them within a month. New sellers sometimes assume slow customer response is a minor inconvenience. Under these rules, it’s a compliance gap that can trigger a formal consumer complaint. Income Tax and Bookkeeping Obligations Sellers Tend to Postpone GST compliance gets most of the attention, but it isn’t the only tax obligation that comes with online selling. Marketplace payouts are business income, and they need to be reported under the correct head when filing income tax returns, regardless of whether the seller operates as a sole proprietor, a partnership, or a private limited company. E-commerce operators are also required to deduct TDS at 0.1 per cent on gross sales made through their platform, a provision formerly under Section 194-O and now folded into Section 393 of the Income Tax Act 2025, separate from the GST-related TCS deduction. Individual and HUF sellers get

The conversation around AI agents’ future of work has moved past speculation. An AI agent, unlike a chatbot that answers a single question, can hold a goal, plan a sequence of steps, use tools, and carry a task through to completion with minimal human input. That shift, from answering to acting, is what’s actually restructuring how work gets assigned, measured, and paid for in 2026. The honest answer to what changes isn’t simple optimism or simple alarm. It’s a mix of real disruption, real new work, and a fair amount of hype that hasn’t been tested yet. The Headline Numbers Are Bigger Than They Sound The most cited figure in this conversation comes from the World Economic Forum’s Future of Jobs Report. It projects that AI and related technologies will displace roughly 92 million jobs by 2030, while creating about 170 million new ones, a net gain of close to 78 million positions globally. That net number sounds reassuring on its own, but labour economists are quick to point out the catch: the workers losing roles are rarely the same people filling the new ones. A net gain at the global level can still mean real, painful churn at the individual and industry level. That churn is already visible in hiring data. Q1 2026 saw more than 78,000 tech layoffs, with nearly half explicitly attributed to AI adoption. At the same time, LinkedIn’s 2026 Labour Market Report found that employers have created at least 1.3 million AI-related job opportunities in the past two years alone, roles like AI engineers, data annotators, and forward-deployed engineers that barely existed five years ago. Both things are true at once, which is exactly why the picture feels confusing from the inside of any single company. Middle Management Is the Layer Under the Most Pressure If there’s one part of the org chart taking the most direct hit, it’s the middle. Gartner has predicted that a fifth of organisations will use AI to flatten their structure, cutting more than half of current middle management roles as a result. The logic is straightforward: a large share of middle management work is status reporting, scheduling, data aggregation, and progress tracking, and those are exactly the tasks agents are already good at. When an agent can compile the weekly report and flag what’s off track, the manager whose main job was compiling that report has a harder case to make for their seat. This doesn’t mean management disappears. It means the parts of management that were really administration are the parts at risk, while the parts built on judgment, coaching, and difficult conversations are not something current agents can do. The Adoption Story Is Messier Than the Marketing Suggests It’s worth being sceptical of how fast this is actually going in practice. Gartner has also predicted that more than 40 per cent of agentic AI projects will be cancelled by the end of 2027, citing escalating costs, unclear business value, and inadequate risk controls. Notably, the reason isn’t that the underlying models don’t work. Gartner’s own analysts point to weak governance, unclear ownership, and a rush to deploy pilots that were never designed to reach production. A lot of what gets marketed as an agent is still closer to a chatbot with an ambitious job title, a pattern the industry has started calling agent washing. That gap between hype and deployment matters for anyone trying to plan a career or a hiring strategy around this. The technology is real and improving quickly, but the organisational discipline needed to use it well- clear goals, defined ownership, and a human who can override the agent when it’s wrong- is still catching up in most companies. What Agents Can Actually Do Right Now As of 2026, agents are reliably good at a fairly specific set of tasks: status reporting, meeting summaries, scheduling, data aggregation, first-draft writing, and multi-step research that follows a clear, repeatable process. What they’re not reliably good at is judgment calls with ambiguous inputs, reading a room, or taking accountability when something goes wrong. That distinction is doing most of the work in deciding which jobs shift and which stay largely intact. There’s also a cost showing up on the human side of this that doesn’t get discussed enough. A February 2026 Harvard Business Review study found that 62 per cent of associates and 61 per cent of entry-level workers reported AI-related burnout, compared with just 38 per cent of C-suite leaders. The study followed about 200 employees at a single U.S. tech company over eight months, so it’s a close-up look rather than an industry-wide survey, but the pattern it found is simple: when AI cuts the time it takes to produce a report from eight hours to two, the expectation often becomes four reports instead of one, not more free time. Agents are raising the baseline of what a role requires, not necessarily lightening the load. The Skill That’s Becoming More Valuable, Not Less Microsoft’s 2026 Work Trend Index makes a point that’s easy to miss in the noise of layoff headlines: as AI use matures, the most effective workers won’t be the ones producing things faster. They’ll be the ones setting clear intent, judging the quality of what the agent produces, and designing how work flows between humans and AI. The question for a lot of roles is shifting from what tasks define this job to what outcomes this person is now positioned to drive. That shows up concretely in creative and analytical work already. A marketer who can generate fifty ad variations with an agent and select the three that will actually perform has a real advantage over both the marketer producing one ad by hand and the one who ships all fifty without curating. The value moved from production to judgment, and judgment is exactly what current agents can’t reliably supply on their own. Which Industries Are Feeling It First The disruption isn’t spread evenly across the economy, and that unevenness is

A new graduate walks in with a strong GPA, a finished thesis, and four years of coursework behind them. Within weeks, the job asks for something different: running a negotiation call, fixing a spreadsheet nobody else wants to touch, explaining a delay to a client without sounding unsure. None of that came up in class. The degree vs job skills gap behind that moment isn’t a personal failing, and it isn’t limited to one school or one student. It shows up in survey after survey, and 2026 has produced some of the clearest numbers yet on how wide it actually is. What the Data Actually Says The National Association of Colleges and Employers runs one of the most closely watched surveys on this exact question every year, and its 2026 Job Outlook report puts a real figure on the disconnect. Employers rated communication skills as important 98.7 per cent of the time, but only 55.4 per cent of them said recent graduates were actually proficient at it. That is a 43-point gap. Critical thinking showed almost the same pattern: 93.5 per cent important, only 50 per cent proficient. Professionalism came in close behind, with a 39-point gap between what employers wanted and what they saw. A separate 2026 study from Gallup and the Lumina Foundation found something almost as telling on the student side. Ninety-three per cent of current bachelor’s degree students said they were confident their school was teaching them the skills they needed to get the job they wanted. Only 54 per cent of employers agreed that colleges are producing graduates with the competencies their businesses actually need. Students and employers are, in a real sense, describing two different educations. A Hult International Business School survey of 1,600 people, split evenly between HR leaders and recent graduates, added a detail that stings a bit more. Seventy-seven per cent of recent graduates said they learned more in their first six months on the job than during their entire undergraduate experience. Ninety-six per cent of HR leaders said schools need to take more responsibility for preparing students for actual work. Those aren’t fringe opinions. They are close to consensus. Why the Mismatch Exists in the First Place Part of the answer is structural, not anyone’s fault in particular. Universities are built to teach transferable, general knowledge that holds up across industries and over decades. A finance professor teaching capital structure theory is preparing a student to think about any company’s balance sheet, not the specific reporting software one employer happens to use this year. Dick Startz, an economist at UC Santa Barbara, has made this point plainly: colleges teach general skills by design, and employers naturally want skills built around their specific needs. Both sides have a reasonable case, and the mismatch is baked into what each side is actually trying to do. But there’s a second, more uncomfortable finding buried in a 2026 report from Cengage Group, which surveyed educators, employers, and graduates together for the first time. Employers ranked job-specific technical ability as their top priority in a new hire. Educators ranked it dead last, focusing instead on soft skills like critical thinking and problem-solving. Eighty-nine per cent of educators still believed their students were workforce-ready, while graduates themselves told a noticeably less confident story. That is not just two groups teaching different things. It is two groups that don’t fully agree on what preparation even means. The One Variable That Actually Moves the Needle Buried inside NACE’s spring 2026 data is a number that matters more than almost anything else in this conversation. Graduates who completed at least one internship or co-op were hired within three months of graduation at a rate of 81.6 per cent. Graduates with no work experience at all were hired at 40.7 per cent. That’s a 40-point gap, and unlike the abstract skills numbers above, it’s something a student can actually act on well before graduation. This lines up with what employers keep saying when asked what they actually look for. NACE also found that 70 per cent of employers now use skill-based hiring in some form, up from 65 per cent the year before. And 71 per cent of those employers use it for at least half of their open roles. Grades and school prestige still matter, but demonstrated experience has become the thing that separates candidates who get hired quickly from candidates who don’t. What Students, Schools, and Employers Can Each Do About It None of this means a degree has stopped mattering. Employers still say they value one: 48 per cent report that most roles at their organisation require a college degree, and three quarters expect a degree to be as important or more important five years from now. The issue isn’t whether to get a degree. It’s what happens, or doesn’t happen, around it. For students, the clearest lever is direct, unglamorous experience. An internship, a part-time job with real responsibility, a student-run project with an actual client, anything that forces contact with ambiguity and other people’s expectations closes more of the gap than an extra semester of coursework tends to. For schools, the data points toward documenting skills in ways an employer can verify, rather than assuming a transcript speaks for itself. One researcher quoted in coverage of the Gallup-Lumina study put it well: a transcript is not a form employers can actually read for competency. For employers, the more useful move is being specific about what a role actually requires on day one, since a lot of the disappointment on both sides traces back to expectations that were never clearly stated to begin with. Most new hires figure out the gap the same way: on the job, not in a classroom. The skills that close it- pushing back on a client politely, admitting a mistake without panicking, asking a question in a meeting without it reading as unprepared- aren’t a different degree. They’re learnable, but they usually aren’t taught, and until that changes, the degree

Shipping costs are rising again in 2026, and this time the pain is landing hardest on exporters who thought the worst of the freight crisis was behind them. If you run an export business out of India, you’ve probably already felt it: a quote that was fine last month suddenly no longer covers your freight line, and your buyer isn’t interested in hearing why. That mismatch between what you promised and what it now costs to deliver is exactly where margins get quietly destroyed. This isn’t a vague “logistics is expensive” complaint. There’s a specific, traceable chain of events behind it, and understanding that chain is the first step to protecting what’s left of your margin. What’s Actually Driving the Rise in 2026 The core problem goes back to the Red Sea. Since late 2023, Houthi attacks on commercial vessels have forced most carriers to avoid the Suez Canal entirely and reroute around Africa’s Cape of Good Hope instead. As of January 2026, only about 26 containerships were sailing through the Suez Canal, compared with 175 taking the longer Cape route, well below the roughly 80 weekly Suez transits that were normal before the crisis. That detour alone adds 7 to 14 days to a typical Asia-Europe voyage and eats up an estimated 2.5 million TEU of global shipping capacity just to keep existing schedules running. Carriers briefly tested a return to the Red Sea in late 2025 and early 2026. Then in late February 2026, US and Israeli military strikes against Iran ended any hope of a large-scale return for the year, and vessels were redirected back to the Cape route to protect crews. What analysts now describe as a temporary detour has effectively become a semi-permanent pricing factor for the year. For Indian exporters specifically, the picture got worse fast. Freight rates on Asia-West Asia routes reportedly jumped from $1,200 to $1,800 per FEU, up to $3,500 to $4,500 per FEU, and some routes saw increases of 250 to 300% once war-risk surcharges were added on top. Add to that Brent crude oil pushing past $105 a barrel and maritime war-risk insurance premiums surging by over 1,000% on affected routes, and container shipping rates on some corridors were set to climb by as much as 40% starting April 1, 2026. India’s merchandise trade deficit widened to $27.10 billion in February 2026, nearly double the figure from a year earlier, with rising freight costs cited as a direct contributor. Why This Hits Indian Exporters Harder Than Most Freight cost increases don’t land evenly across every exporter. Labour-intensive, thin-margin sectors like textiles, garments, and leather goods absorb the biggest hit, because they typically operate on margins too tight to simply pass the increase on without losing the order to a competitor. Industry bodies have repeatedly warned that even a small cost disadvantage can push buyers to shift orders to competitors like Bangladesh or Vietnam, which is exactly the kind of order diversion India can’t afford while chasing its stated goal of $2 trillion in exports by 2030. There’s also a compliance and forecasting problem layered on top of the pure cost problem. Roughly 70% of Indian decision-makers say cross-border trade has become more complicated, and freight forwarders themselves aren’t optimistic either, with 92% expecting tighter margins in 2026 due to geopolitical risk and surcharges. When the people moving the cargo expect it to get harder, exporters planning shipments six months out are effectively pricing in the dark. The Government Response: RoDTEP and RELIEF To its credit, the government hasn’t ignored this. Two schemes matter right now if you’re exporting from India. RoDTEP (Remission of Duties and Taxes on Exported Products) refunds embedded taxes and duties that would otherwise sit buried in your export cost, and it’s confirmed to continue at existing rates at least through 30 September 2026. Rates generally range between 0.3% and 4.3% of export value depending on the HS code, and on a large turnover that’s not trivial: even a 1% swing in RoDTEP on a ₹10 crore export turnover works out to roughly ₹10 lakh in annual incentive value. In late February 2026, the government actually cut RoDTEP rates by 50%, then reversed that decision within a month and fully restored rates from 23 February 2026, specifically citing the pressure of West Asia freight disruptions on exporters. The second, newer scheme is RELIEF (Resilience & Logistics Intervention for Export Facilitation), approved on 19 March 2026 with an outlay of roughly ₹497 crore. It’s built specifically around the Gulf and West Asia shipping crisis. It covers three groups: exporters already insured by ECGC get compensation of up to 100% of eligible losses for shipments made between 14 February and 15 March 2026, exporters shipping between 16 March and 15 June 2026 can access up to 95% risk coverage through ECGC, and MSME exporters who weren’t insured at all during the disruption window can claim a partial reimbursement of up to 50% of incremental freight and insurance costs, capped at ₹50 lakh per exporter. It’s worth being clear-eyed about this scheme too: it’s explicitly time-bound and not designed as a permanent fix, so exporters shouldn’t build long-term pricing models around it being there indefinitely. How Exporters Can Actually Protect Their Margins Government relief helps, but it isn’t a substitute for how you run pricing and freight decisions inside your own business. A few practical shifts make the biggest difference. Stop pricing freight as a fixed line item. Treating freight as a controllable, variable cost rather than a fixed external expense is what separates exporters who protect margins from those who absorb every shock. Build your quotes with a freight buffer instead of locking in today’s rate for a shipment three months out. Use contract rates as your baseline, but keep spot access open. Analysts recommend working with a freight forwarder that gives you both contract and spot rate access, so you get a stable floor while still being able to grab a spot rate when it dips

How you explain what you know decides whether that knowledge is worth anything to the person standing in front of you. I’ve sat in meetings where the smartest person in the room said absolutely nothing useful, not because they didn’t understand the problem, but because they explained it in a way only they could follow. And I’ve watched a junior employee, who knew a fraction of what the expert knew, win the whole room over in five minutes flat because they knew how to make the idea land. That gap between knowing and explaining is where most good ideas quietly die. We’re taught from school onward that knowledge is the prize. Get the degree, read the book, collect the certification, and the rest takes care of itself. Nobody really teaches us the second half of that equation: knowing something and making someone else understand it are two completely different skills, and one doesn’t automatically produce the other. Why Smart People Struggle to Explain Simple Things There’s a well-documented psychological pattern behind this, and it isn’t a personality flaw; it’s just how the brain works once it learns something. Psychologists call it the curse of knowledge. In 1990, a Stanford psychology graduate student named Elizabeth Newton ran a now-famous experiment to study exactly this. She split participants into two groups: “tappers” and “listeners.” Tappers picked a well-known song from a list of 25 tunes, things like “Happy Birthday” and “Twinkle Twinkle Little Star,” and tapped its rhythm out on a table. Listeners had to guess the song from the taps alone. Before the listeners guessed, Newton asked the tappers one question: What percentage of listeners do you think will get it right? The tappers predicted around 50%. The actual result, across 120 songs tapped out, was just 2.5%, only three correct guesses in the entire experiment. Here’s what makes the study so revealing: while a tapper taps, they can’t help but hear the full melody in their head, complete with instruments and lyrics. The listener, meanwhile, hears nothing but disconnected knocks on a table. The tapper’s knowledge of the song made it almost impossible for them to imagine what it sounded like to someone who didn’t already know it. That’s the trap. Once you know something, you genuinely cannot remember what it felt like not to know it. This is why a doctor rattles off medical terminology to a worried patient without realising they’ve lost them at the second sentence, why an engineer explains a bug fix to a client using acronyms nobody outside the team recognises, and why a manager assumes their team understood instructions that were, in the manager’s own head, clear. Knowing Something Doesn’t Mean You Can Teach It There’s a popular quote floating around the internet that goes something like, “If you can’t explain it simply, you don’t understand it well enough,” usually attributed to Albert Einstein. Worth knowing: there’s no real record of Einstein ever saying it, and it’s often mistakenly credited to physicist Richard Feynman as well. The line makes a tidy motivational poster, but it actually oversimplifies something more interesting. Feynman himself, famous for his gift for teaching complex physics in plain language, once admitted something far more honest. When a colleague asked him to explain, at a beginner level, why a certain class of particles behaves the way it does, Feynman said he’d prepare a freshman lecture on it. A few days later, he came back and said he couldn’t do it. He couldn’t bring the explanation down to that level, and he took that as a sign that even physicists at the highest level hadn’t fully cracked the concept themselves. That’s a more honest picture than the fake Einstein quote. Understanding and explaining are related, but they’re not the same skill, and even brilliant people sometimes hit a wall between the two. The upside is that explaining forces a kind of understanding you can’t get any other way. Trying to teach a concept to someone else, or even to a blank page, exposes exactly where your own grasp of it gets fuzzy. That’s the entire logic behind study techniques where you explain a topic out loud in plain language and circle back to the parts where you stumble. What Actually Makes an Explanation Land If the curse of knowledge is the disease, the cure isn’t complicated, but it does take deliberate effort. A few things consistently separate explanations that land from ones that don’t: Start from where your listener actually stands, not from where you stand. The tapper’s mistake wasn’t a lack of knowledge; it was forgetting that the listener didn’t share it. Before explaining anything, it helps to ask what the other person already knows and build from there instead of your own starting point. Concrete language beats abstract language every time. “Our churn rate increased” is technically accurate and instantly forgettable. “We’re losing one in five customers within their first three months” makes people sit up. Numbers, comparisons, and real examples stick where abstractions slide right off. Stories and analogies do the heavy lifting that jargon can’t. When you connect a new idea to something the listener already understands, you’re not dumbing it down; you’re giving their brain a hook to hang the new information on. Short sentences and plain words aren’t a sign of weak thinking. If anything, the ability to compress a complicated idea into a few clear words is usually a sign you understand it deeply enough to strip away everything that isn’t essential. Check for understanding instead of assuming it. A tapper never got feedback mid-tap about whether the listener was following along. In real conversations, you can. Asking “does that make sense so far?” or watching someone’s face for confusion costs you nothing and saves everyone the trouble of a conversation that goes nowhere. Why This Matters More Than Ever We live in a time where information is genuinely cheap. Anyone can look up a fact, a formula, or a definition in seconds. What’s

Export Payment Methods are one part of an export deal that decides whether you actually get paid or spend months chasing money across borders. Choosing the right export payment method matters more than most first-time exporters realise. If you’ve ever sat across the table from a new overseas buyer and heard the words “let’s talk payment terms,” you know that moment changes the whole mood of a deal. Everything else- price, quantity, delivery schedule- suddenly feels secondary. Because in export trade, getting paid isn’t automatic. You’ve shipped your goods across an ocean; they’re sitting in a container somewhere, and the only thing standing between you and your money is a piece of paper (or these days, a SWIFT message) and the promise behind it. I’ve seen exporters lose sleep, and lose money, over choosing the wrong payment method. So let’s go through the five real export payment methods on the table: Letter of Credit (LC), Advance Payment, Documents against Payment (DP), Documents against Acceptance (DA), and Open Account. No jargon soup, just what each one actually means for your cash flow and your risk. 1. Letter of Credit (LC): The Bank Steps In Between An LC is basically the buyer’s bank telling you, “Ship the goods, present the right paperwork, and we’ll pay you, even if our customer doesn’t.” That’s the whole appeal. Instead of trusting a stranger on the other side of the world, you’re trusting his bank. Of all the export payment methods available, this is the one built specifically to remove buyer risk from the equation. This isn’t some informal handshake arrangement either. LCs are governed by a real rulebook: the Uniform Customs and Practice for Documentary Credits (UCP 600), published by the International Chamber of Commerce. It’s been through several revisions since the first version in 1933, and the current UCP 600 came into effect in 2007 and is made up of 39 articles. What’s notable is the scale: these rules apply in around 175 countries and govern roughly $1 trillion worth of trade every year. That’s not a niche instrument; it’s the backbone of global trade finance. Here’s the catch people don’t always appreciate: the bank doesn’t care whether the goods are good or bad, whether they arrived on time, or whether the buyer is happy. The bank only checks documents. If your invoice, bill of lading, packing list, and certificate of origin match the LC’s conditions exactly, down to spelling and dates, you get paid. If there’s a mismatch (a “discrepancy,” in trade-finance speak), the bank can refuse to honour it, and now you’re negotiating with your buyer to waive the discrepancy. One important detail: under UCP 600, banks have a hard maximum of five banking days to examine documents, whereas before that it was a vague “reasonable time,” which used to drag disputes out for weeks. LCs also come with a real cost: issuance fees, confirmation fees (if you want a bank in your own country to add its guarantee too), and sometimes discounting charges if you need the cash before the credit matures. For a first-time buyer, a big-ticket order, or a country where you’re not sure about political or banking stability, an LC is still the gold standard of security. For a small, repeat order to a buyer you trust, it can feel like using a sledgehammer to crack a nut. 2. Advance Payment: Zero Risk, One Problem This is the exporter’s dream: the buyer wires money before you even start production, sometimes 100% upfront, sometimes a partial advance with the balance on shipment. You carry zero credit risk. You know your money is in the bank before the goods leave the factory floor. The problem? Almost no buyer wants to pay 100% in advance to a supplier they’ve never dealt with. Would you? Advance payment tends to work only when you have serious leverage: a scarce product, a strong brand, or an existing relationship built over years. It’s common for custom-made goods, small sample orders, or when the buyer is equally new and cautious and prefers to test the relationship with a small deposit first. As a full payment method for large first-time export orders, it’s honestly rare in practice. 3. Documents against Payment (DP): Pay First, Then Get the Papers DP, also called “cash against documents,” sits in the middle ground of the documentary collection world. Here’s how it actually works: you ship the goods, then hand your shipping documents, including the bill of lading, which is what lets someone claim the cargo, to your bank. Your bank forwards them to the buyer’s bank with instructions: release these documents only once the buyer pays. This whole process runs on a different rulebook than LCs: the Uniform Rules for Collections, ICC Publication No. 522 (URC 522), which has been in force since 1996 and is built around 26 articles across seven sections covering the roles of exporter, importer, and the banks involved. One thing that trips people up: under URC 522, banks are explicit that they will not examine the documents for accuracy or completeness; they’re just following instructions, not underwriting the deal the way an LC-issuing bank does. So the safety here comes from control over the documents, not from a bank guarantee. The real risk in DP: your buyer might simply refuse to pay once the goods arrive, especially if market prices have dropped or he’s changed his mind. Now your cargo is sitting at a foreign port, you’re paying demurrage charges, and you’re scrambling to find another buyer or ship it back. It’s cheaper and faster to arrange than an LC, but it depends heavily on trust and on your buyer actually wanting the goods badly enough to pay. 4. Documents against Acceptance (DA): Pay Later, On Trust DA works almost identically to DP, except for one crucial difference: instead of paying immediately, the buyer only needs to accept a bill of exchange, essentially signing a promise to pay at a future date, say 60 or