
Sign up to save your podcasts
Or


Why have only 10% of companies redesigned the workplace for AI? Because Lenovo's Work Reborn study of 1,200 IT leaders found that 80% say the digital workplace is critical to successful AI-led transformation — yet only 10% have fundamentally redesigned it for AI, and only 27% have scaled AI across business functions. The priority is declared, the money is moving, and the workplace itself is still the old workplace with new software on top of it. In this episode of CHANGEReadyX™ Insights, Ram Narayanan and Priya unpack the ten-percent gap, the confidence performance Avanade found on the other side of it, and the discipline that turns a workplace into an AI-ready one.
Lenovo published the sixth Work Reborn report, "The AI-ready workplace", in August 2026: 1,200 IT leaders at organisations with more than a thousand employees, across twelve countries. They measured readiness across five dimensions — Architecture, Operations, Trust, Workforce, Value — and on every one the numbers are thin. Only 7% operate a fully hybrid AI architecture. Only 15% have predictive, largely autonomous IT support. Only 10% have comprehensive AI governance. Only 12% measure AI adoption and effectiveness at the individual and team level. Only 10% have a continuous, workflow-level return-on-investment framework. Avanade's AI Reality Report, released 4 November 2026 and researched by Vanson Bourne among 3,600 business decision-makers across eleven markets (1,400 in Asia Pacific), is the mirror image. In Asia Pacific, 76% describe their teams as Established, Advanced or AI-first, and 73% have high or complete confidence their teams use AI effectively — yet only 17% have fundamentally reshaped how their entire teams operate around AI, 68% have yet to formally calculate the financial return on their AI investment, 79% say too much AI investment goes to incremental improvements rather than transformational opportunities, and 76% say their own organisation publicly portrays itself as more AI-mature than it actually is. Globally it is the same shape: 77% Established or better, 72% highly confident, only 18% reshaped. The debate bank explains why: MIT NANDA found 95% of generative AI pilots show no measurable return on the P&L; S&P Global found 42% of companies abandoned most of their AI initiatives in 2025, up from 17% the year before; Gartner says over 40% of agentic AI projects will be cancelled by the end of 2027; BCG's 10-20-70 rule says 70% of AI success is people and processes. The workplace is where that 70% lives.
Three takeaways: (1) Only 10% have fundamentally redesigned the workplace for AI — Lenovo, 1,200 IT leaders — while 80% say it is critical and 76% rank AI-led transformation a top IT priority. The gap is the story. (2) Confidence is a performance: Avanade found 76% of APAC teams call themselves Established or better and 73% are highly confident, yet only 17% reshaped how teams operate, 68% never calculated the return, and 76% admit their organisation portrays itself as more mature than it is. (3) Redesign the work, not just the tools — diagnose readiness across Architecture, Operations, Trust, Workforce and Value before the next purchase, redesign the workflow with named owners and a workflow-level ROI framework, and measure adoption to value, not licences bought.
Chapters: 00:00 Cold Open — Ten Percent Redesigned · 02:10 What Lenovo Actually Found · 04:16 The Avanade Mirror — Confident, But Not Reshaped · 06:37 The Pattern — Installed Software Is Not Redesigned Work · 08:54 The Four-Part Discipline — Redesign the Work, Not Just the Tools · 11:34 The Diagnostic Question — What Have You Actually Redesigned?
Run the diagnostic: https://tapasadvisory.com/diagnostic
Substack: https://open.substack.com/pub/ramchangereadyx
Apple Podcasts: https://podcasts.apple.com/us/podcast/rams-changereadyx-from-tapas-advisory/id6831708010
Ram Narayanan – Transformation Leader | Architect of CHANGEReadyX™, TAPAS Advisory
Why does $1 in change management return $4 in AI value? Because Deloitte's study of 1,000 change leaders found that every dollar invested in change management was associated with four dollars of realised business value at the median — and still, almost nobody funds it. Deloitte and Workplace Intelligence (fielded by Walr, June–July 2026) surveyed 1,000 change-management leaders, 51% in the C-suite, across the US, UK, Australia and Canada. In this episode of CHANGEReadyX™ Insights, Ram Narayanan and Priya unpack the four-to-one return, what it costs when the people work is skipped, and the discipline that turns the headline into a result.
Deloitte published "The Change Paradox" on 18 October 2026: every $1 in change management was associated with $4 in realised business value at the median. The study spoke to 1,000 change leaders, half of them C-suite, each of whom had led a large-scale initiative of at least $1M in the past two years. The paradox is that the same leaders report the return and refuse the funding. Seventy percent say their latest transformation failed to deliver its intended value. Only 41% report a strong positive ROI on their AI investments. Forty-three percent do not know how AI is actually affecting their employees, 61% feel AI transformation is moving faster than employees can handle, and 59% say AI overload is negatively affecting their employees. The cost of skipping the people work is measured in people and money: 66% say employees left as a direct result of their most recent initiative, at an estimated average of $20.5M in turnover-related costs — $25.5M among AI-implementation leaders. Forty-five percent say poorly managed change contributed to the departure of high performers; 69% report increased workload and burnout; 68% report at least one form of employee resistance; 83% say poor change management contributed to unrealised business value. Half of leaders say their company has gaps in change-management execution. Why does it keep happening? Sixty percent treat change management as a project cost, not a business capability — so it is cut first, and the cut is invisible until it fails. The debate bank explains why: MIT NANDA (95% of GenAI pilots show no measurable P&L return), BCG's 10-20-70 rule (70% of AI success is people and processes), McKinsey (only 37% report any EBIT impact from AI). The fix is the four-part discipline: diagnose readiness before a single dollar of AI spend is approved, design the change as a capability, equip managers to carry it, and measure adoption to value. The episode closes with the diagnostic question: before you approve the next dollar of AI spend, how much of it is funding the people who have to carry it — and what is your four-to-one?
Three takeaways:
1. The median return on change management is four to one — Deloitte's study of 1,000 change leaders — and it is the highest-return line in the AI budget, not a project cost to trim.
2. Skipping the people work is the most expensive saving a company makes: 66% of leaders saw employees leave as a direct result of their last initiative, at an average of $20.5M in turnover costs ($25.5M for AI programmes), with high performers leaving first.
3. Fund the people first — diagnose readiness before a single dollar of AI spend is approved, design change as a capability with the four-to-one in the business case, and measure adoption to value, not licences bought.
Chapters:
00:00 Cold Open — The Four-Dollar Claim
02:10 What Deloitte Actually Found
04:12 The Cost of Getting It Wrong
06:09 The Pattern — The Invisible Line Item
08:30 The Four-Part Discipline — Fund the People First
11:01 The Diagnostic Question — What Is Your Four-to-One?
Run the diagnostic: https://tapasadvisory.com/diagnostic
Substack: https://open.substack.com/pub/ramchangereadyx
Apple Podcasts: https://podcasts.apple.com/us/podcast/rams-changereadyx-from-tapas-advisory/id6831708010
Ram Narayanan – Transformation Leader | Architect of CHANGEReadyX™, TAPAS Advisory
Why is Qantas cutting 850 jobs in the name of AI? Because the announcement braids three strands together — AI and automation, outsourcing to Accenture teams mostly in India, and a redesign of more than 1,000 legacy systems — and Qantas has not said how many of the 850 roles AI will actually replace and how many will simply move. Qantas Group announced on 10 November 2026 a multi-year Accenture partnership to modernise its back-end systems and step up investment in technology and AI; around 850 Australian corporate roles are expected to be affected, first changes by mid-2027. In this episode of CHANGEReadyX™ Insights, Ram Narayanan and Priya pull the strands apart.
Qantas announced the Accenture partnership on 10 November 2026: around 850 Australian corporate roles affected. The headline was simple: Qantas cuts jobs for AI. But only one of the announcement's three strands is AI. Qantas spends more than A$1 billion a year on technology and runs more than 1,000 older applications. Some services go to Accenture, delivered mostly by its teams in India, and in time transformed further with automation and AI. The roles, about 3% of the workforce, sit mainly at Mascot HQ in Sydney, starting with finance, procurement, people and marketing. The contract value is undisclosed, and the automated-versus-relocated split is missing. The Australian Services Union disputes that AI is the real driver: Scott Cowen told ABC News the first step is moving Australian work to India, and the union says Qantas had previously assured staff there was no plan to offshore jobs. The day before, the Reserve Bank of Australia published analysis finding little evidence so far that AI had materially disrupted Australia's labour market. That is the pattern this episode names AI washing: a layoff wearing an AI costume, which makes a cost decision sound like strategy, a redundancy sound inevitable, and moves the conversation from the people leaving to the technology arriving. The debate bank explains why the costume fits: MIT NANDA (95% of GenAI pilots show no measurable P&L return), BCG's 10-20-70 rule (70% of AI success is people and processes), McKinsey (only 37% report any EBIT impact from AI) and Kyndryl (70% of Indian companies have more AI pilots than they can scale). The fix is the four-part discipline: diagnose readiness before a single role is cut, redesign the work before removing the people, and measure adoption to value. The episode closes with the diagnostic question: of the roles you are about to cut in the name of AI, how many has AI actually replaced, with a number and a date?
Three takeaways: (1) A layoff announced in AI's name is not proof that AI caused it — Qantas has not disclosed how many of the 850 roles will be automated versus relocated; attribute the cause to the evidence, not the headline. (2) The costume does three jobs at once — it makes a cost decision sound like strategy, a redundancy sound inevitable, and moves the conversation from the people leaving to the technology arriving. (3) Diagnose before you cut — which roles are truly automatable today, and which carry knowledge no model holds; redesign the work before removing the people, and measure adoption to value, not roles removed.
Chapters:
00:00 Cold Open — The Announcement and the Question
02:07 What Qantas Actually Announced vs What the Headline Says
04:04 The Union's Challenge and the Reserve Bank Evidence
05:55 The Pattern — AI Washing as a Management Move
08:15 The Four-Part Discipline — Diagnose Before You Cut
10:37 The Diagnostic Question — Has AI Replaced Them, or Just Renamed Them?
Links:
Diagnostic: https://tapasadvisory.com/diagnostic
Substack: https://open.substack.com/pub/ramchangereadyx
Spotify: https://open.spotify.com/show/1P4bNrFJ2SvYwGMMYHtNL7
Apple Podcasts: https://podcasts.apple.com/us/podcast/rams-changereadyx-from-tapas-advisory/id6831708010
Ram Narayanan – Transformation Leader | Architect of CHANGEReadyX™, TAPAS Advisory
Why do Indian companies have more AI pilots than they can scale? Because a pilot is the easiest yes in enterprise — cheap, contained, visible — and nobody owns the decision about which ones scale, which ones merge, and which ones die. Kyndryl's Readiness Report 2026 (3,700 leaders, 21 countries, 200 in India) found 70% of Indian organisations have more AI pilots than they can realistically scale, and 58% say innovation frequently stalls after the proof of concept. The pilots are not failing. The pilots are piling up. In this episode of CHANGEReadyX™ Insights, Ram Narayanan and Priya show why the backlog builds, why returns look better than readiness, and how CHANGEReadyX™ turns a backlog of demos into a pipeline of production systems: Sense & Diagnose · Design & Align · Enable & Activate · Adopt & Measure.
Seventy percent of Indian companies are running more AI pilots than they can ever scale — and every single demo worked beautifully. That is the opening fact of Kyndryl's Readiness Report 2026, released 10 November 2026. The backlog builds because a pilot is the easiest yes in enterprise: customer service wants a chatbot pilot, finance wants an invoice pilot, HR wants a screening pilot, and the CIO says yes to all of them. Twelve months later the organisation runs a dozen pilots and nobody owns the decision about which ones scale, which ones merge, and which ones die. The report holds a contradiction: 85% say they are deploying AI in workflows beyond pilot projects, yet 56% still describe themselves as experimenting — deployment and transformation are not the same thing. Then the paradox: 60% of Indian organisations investing in AI and machine learning report positive returns, up from 57% last year — but only 27% say they are fully ready to manage external risks, down from 35%, and only 35% say they are completely ready to manage cyberattacks, down ten points. The returns are self-reported, not independently verified; McKinsey found only 37% report any EBIT impact from AI. The stall has a shape: a proof of concept proves the model can do the task in a clean room, and production asks four harder questions — is the data good enough every day, who owns the decision when the machine is wrong, who checks the output, and who changes their job. MIT NANDA (2025) found 95% of generative AI pilots show no measurable P&L return, and BCG's 10-20-70 rule puts 70% of AI success in people and processes — while the pilot budget goes into the 10 and the 20. The fix is the four-part discipline: diagnose readiness before another rupee is spent, redesign the work around the pilot, build role-based capability through managers, and measure adoption to value. Every pilot gets a kill-or-scale gate with a date.
Three takeaways: (1) A pilot without an owner is a hobby — diagnose readiness before another rupee is spent, and give every pilot a kill-or-scale gate with a date. (2) Returns up, readiness down is a fragile return — 60% report positive returns while only 27% are fully ready to manage external risks; a return built on falling readiness can reverse. (3) The proof of concept proves the wrong thing — the model in a clean room, not the organisation around it; 95% of GenAI pilots show no measurable P&L return because the 70% (people and processes) was left to sort itself out.
Chapters: 00:00 Cold Open — Pilot Purgatory · 02:03 The Backlog — More Pilots Than Hands · 04:40 The Returns Paradox — Sixty Percent See Returns, Twenty Seven Percent Are Ready · 07:08 Why Pilots Die After Proof of Concept · 09:48 The Four-Part Discipline — From Pilot Purgatory to Production · 12:29 The Diagnostic Question — Which Pilots Earn Their Scale?
Run the diagnostic: https://tapasadvisory.com/diagnostic
Substack: https://open.substack.com/pub/ramchangereadyx
Apple Podcasts: https://podcasts.apple.com/us/podcast/rams-changereadyx-from-tapas-advisory/id6831708010
— Ram Narayanan, Transformation Leader | Architect of CHANGEReadyX™, TAPAS Advisory
Why is AI spending up but profits still flat? Because organisations bought the tool without redesigning the work. McKinsey's global survey on the state of AI in 2026 found only 37% of respondents report any impact from AI on EBIT — essentially unchanged from a year earlier — while 89% use AI in at least one business function and 44% are scaling it enterprise-wide. The 6% high performers are the exception: nearly three-quarters of them fundamentally redesigned their workflows. In this episode of CHANGEReadyX™ Insights, Ram Narayanan and Priya show why personal productivity is a mirage, why the return sits in the workflow, and how CHANGEReadyX™ helps leaders start from people, not technology: Sense & Diagnose · Design & Align · Enable & Activate · Adopt & Measure.
A chief financial officer signs a bigger AI licence in January. By March every team has access; by June the dashboard shows adoption — seats active, prompts run, documents drafted. The vendor calls it success. In September the chief financial officer looks at the profit and loss account and asks the obvious question: where is it? (Illustrative composite, not a real client.) McKinsey's 2026 global survey — 1,719 respondents in 97 countries — gives the answer. Only 37% report any impact from AI on EBIT, essentially unchanged from a year earlier, even though 89% of organisations use AI somewhere and 44% are scaling enterprise-wide, up from 38%. Eighty percent of respondents say AI improved their individual productivity — yet the profit line is flat, because saved minutes evaporate inside workflows that waste hours. Only 6% are AI high performers, attributing at least 5% of EBIT to AI, and nearly three-quarters of them fundamentally redesigned their workflows, against about one-quarter of everyone else. MIT NANDA (2025) found 95% of generative AI pilots show no measurable P&L return; BCG's 10-20-70 rule puts 70% of AI success in people and processes. Gartner expects over 40% of agentic AI projects to be cancelled by the end of 2027, and S&P Global found 42% of companies abandoned most of their AI initiatives in 2025, up from 17% the year before. The episode lands on the standing thesis: redesign the work, equip the people, measure adoption to value — CHANGEReadyX™ starts from people, not technology.
Three takeaways: (1) Redesign the workflow before renewing the licence — the 6% high performers rebuilt who decides, who checks, and where freed hours go; everyone else installed software and kept the old work. (2) Personal speed is not enterprise value — 80% feel individually more productive, yet the EBIT line is flat, because saved minutes evaporate inside workflows that waste hours. (3) Measure adoption to value, not seats active — seats active is a procurement metric; the measure must land on the P&L, where only 37% saw any impact.
Chapters: 00:00 Cold Open — The Thirty Seven Percent Problem · 01:59 The Spending Paradox — More Money, Same Return · 04:32 The Productivity Mirage — Eighty Percent Feel Faster, Nothing Moves · 07:00 The Thirty Seven Percent — What McKinsey Actually Found · 09:38 The Four-Part Discipline — Redesign the Work, Not the Licence · 12:18 The Diagnostic Question — What Did We Redesign?
Run the diagnostic: https://tapasadvisory.com/diagnostic
Substack: https://open.substack.com/pub/ramchangereadyx
Apple Podcasts: https://podcasts.apple.com/us/podcast/rams-changereadyx-from-tapas-advisory/id6831708010
Ram Narayanan – Transformation Leader | Architect of CHANGEReadyX™, TAPAS Advisory
Why do AI pilots fail on data, not on models? Because the model is almost never the weakest link — the data underneath it is. In this episode of CHANGEReadyX™ Insights, Ram Narayanan and Priya unpack healthcare technology chief David Jarmoluk's trending argument, his two field stories — customer identifiers that were not as unique as everyone claimed, and a model that earned trust by being tested rather than by being bigger — and show how CHANGEReadyX™ helps leaders start from people, not technology: Sense & Diagnose · Design & Align · Enable & Activate · Adopt & Measure.
A healthcare team builds a model on customer records, believing the customer identifier is unique — one customer, one number, one record. It is wrong: the same identifier points at different people, different identifiers at the same person, and the model learns from a fiction, confidently and wrongly (Jarmoluk's field account). In the second story, a model earns clinicians' trust not by being bigger but by being tested — shown where it is right, where it is wrong, and how it behaves at the edges. MIT NANDA (2025) found 95% of generative AI pilots show no measurable return on the P&L; BCG's 10-20-70 rule puts 70% of AI success in people and processes and only 20% in technology and data — the very 20% these stories starve. Gartner expects over 40% of agentic AI projects to be cancelled by the end of 2027, and S&P Global found 42% of companies abandoned most AI initiatives in 2025, up from 17% the year before. Klarna went AI-first in customer service, quality fell, and it began hiring humans again. The episode lands on the standing thesis: test the data, test the model with the people who must trust it, measure adoption to value — CHANGEReadyX™ starts from people, not technology.
Three takeaways: (1) Test the identifier before the model — one question, asked before a single line of model code, saves the programme: is this field actually unique, complete, and owned by someone? (2) Trust is a record, not a feeling — a smaller model, honestly tested with the people who must use it, beats a larger model nobody believes, because the larger one sits on a shelf. (3) Measure adoption to value, not model accuracy — accuracy on untested data is a beautiful number that means nothing; 95% of pilots showed no P&L return because nobody measured adoption to value, only deployment.
Chapters: 00:00 Cold Open — The Data Under the Pilot · 01:49 The Duplicate ID — A Number That Was Not Unique · 04:28 Trust Is Tested, Not Bought — The Model That Earned Belief · 06:54 The Ninety Five Percent — Pilots Die Before Production · 09:55 The Four-Part Discipline — Diagnose the Data, Equip the People · 12:27 The Diagnostic Question — What Does Your Data Actually Say.
Links: Diagnostic https://tapasadvisory.com/diagnostic · Substack https://open.substack.com/pub/ramchangereadyx · Apple Podcasts https://podcasts.apple.com/us/podcast/rams-changereadyx-from-tapas-advisory/id6831708010
Ram Narayanan – Transformation Leader | Architect of CHANGEReadyX™, TAPAS Advisory
Why does AI adoption fail on behaviour, not technology? Because no rollout has ever failed for want of a stronger model — it fails because nobody named the behaviour that must be different on day ninety. In this episode of CHANGEReadyX™ Insights, Ram Narayanan and Priya unpack Ayo Adebayo's trending argument, the MIT NANDA finding that 95% of GenAI pilots show no P&L return, and the Slack research showing one in five workers hides their AI use — and show how CHANGEReadyX™ helps leaders adopt from people, not technology: Sense & Diagnose · Design & Align · Enable & Activate · Adopt & Measure.
A global capability centre rolls out an AI assistant to three thousand analysts; licences live on day one, logins at ninety percent — and on day ninety nothing has changed, because the behaviour never arrived (illustrative scenario). MIT NANDA (2025) found 95% of generative AI pilots show no measurable return on the P&L, and BCG's 10-20-70 rule puts 70% of AI success in people and processes. The episode also applies evidence discipline to the profession's most repeated number: "70% of change fails" has no citable primary source — Mark Hughes traced it in a 2011 peer-reviewed paper and found no valid and reliable empirical evidence; it began as Hammer & Champy's 1993 "unscientific estimate". Slack's Workforce Lab survey of more than 5,000 desk workers shows why behaviour is the gap: about one in five workers uses AI but hides it, and nearly one in five avoids it entirely.
Three takeaways: (1) Logins are not adoption — adoption is the named behaviour that is different on day ninety, with an owner and a measure; without that sentence the plan is a shopping list. (2) The evidence discipline cuts both ways — the "70% of change fails" number has no primary source (Hughes, 2011), so name the source or do not use the number, just as you name the behaviour or do not claim the adoption. (3) Hidden use is the signal — one in five workers hides AI use and nearly one in five avoids it (Slack Workforce Lab); permission without a named behaviour is a slogan, and managers are the people who turn a licence into a behaviour.
Chapters:
00:00 Cold Open — Name the Behaviour by Day Ninety
01:58 The Licence Trap — Bought the Tool, Kept the Habit
04:09 The Seventy Percent Myth — A Number With No Source
06:34 The Hidden Workforce — One in Five Hides Their A I Use
09:04 The Four-Part Discipline — CHANGEReadyX Starts From People
11:42 The Diagnostic Question — What Changes on Day Ninety
Free readiness diagnostic: https://tapasadvisory.com/diagnostic
Substack: https://open.substack.com/pub/ramchangereadyx
Apple Podcasts: https://podcasts.apple.com/us/podcast/rams-changereadyx-from-tapas-advisory/id6831708010
Ram Narayanan – Transformation Leader | Architect of CHANGEReadyX™, TAPAS Advisory
Why AI Investment Is Not The Problem — Underinvestment In Capability - what actually decides the outcome.
Ram Narayanan, Transformation Leader | Architect of CHANGEReadyX™, TAPAS Advisory
Free readiness diagnostic: https://tapasadvisory.com/diagnostic
Is cutting AI investment the real mistake? Yes — when the cut is indiscriminate. Capital investment only pays when it is matched by organisational readiness and workforce capability.
Alphabet's 13F filing of 6 November 2026 (Reuters) disclosed 2,976,982 Qualcomm shares worth $547.9M at 30 September 2026 — but the holder is GV, Alphabet's venture arm, which received the shares when Qualcomm acquired AI software company Modular in a ~$3.9B all-stock deal: not a fresh cash purchase. The real story is the infrastructure shift: Qualcomm targets $5B of data-centre revenue in FY2027 and $15B+ by FY2029, Amazon has a multi-generation custom-silicon deal with warrants tied to up to $60B of purchases, and Meta and Microsoft are also Qualcomm data-centre customers. MIT NANDA (2025) found 95% of organisations saw no measurable P&L return from GenAI pilots, and BCG's 10-20-70 rule puts 70% of AI success in people, process and change.
Three takeaways: (1) The Alphabet holding is a signal, not a transaction. (2) Discipline beats both extremes — investors are right to demand returns, but indiscriminate cuts starve the 70% that matters most. (3) Readiness is strategy — diagnose readiness before deploying a single tool: Sense & Diagnose · Design & Align · Enable & Activate · Adopt & Measure.
Chapters: 00:00 Cold Open — Is Underinvesting the Real Mistake? · 00:57 The News — Alphabet, Qualcomm and the Nuance · 03:11 Why Investment Matters — The Infrastructure Shift · 04:37 The Other Side — Discipline, Not Indiscriminate Cuts · 08:06 The Transformation Challenge — Readiness Before Capital · 11:17 The Strategic Question — Diagnose First.
Run the free readiness diagnostic: https://tapasadvisory.com/diagnostic
Substack: https://open.substack.com/pub/ramchangereadyx
Apple Podcasts: https://podcasts.apple.com/us/podcast/rams-changereadyx-from-tapas-advisory/id6831708010
Ram Narayanan – Transformation Leader | Architect of CHANGEReadyX™, TAPAS Advisory
Most companies are not transforming. They are optimising — improving what already exists, faster and cheaper — and calling it transformation. In this episode of CHANGEReadyX™ Insights, Ram Narayanan and Priya draw the line between business change (reference point: the past) and business transformation (reference point: the future), apply the same test to enterprise AI — AI optimisation inside an old operating model vs. AI transformation that redesigns how work happens — and show how CHANGEReadyX™ helps leaders move from optimisation to transformation: Sense & Diagnose · Design & Align · Enable & Activate · Adopt & Measure.The episode draws on a framing Rob Llewellyn at CXO Transform puts brilliantly in his Boardroom Brief work, and on the MIT NANDA finding that 95% of GenAI pilots show no P&L return — evidence that the bottleneck is organisational readiness, not the technology.Three takeaways:1. Change and transformation have different reference points — change looks backwards and improves the existing story; transformation looks to the future and writes a different one. A faster caterpillar is still a caterpillar.2. The questions in the room give the programme away — 'how do we get better?' is a change question; 'what must we become?' is a transformation question.3. The same test applies to AI — AI optimisation makes the existing workflow quicker inside an old operating model, while AI transformation redesigns how work itself happens; 95% of GenAI pilots show no P&L return (MIT NANDA, 2025) because readiness, not technology, is the bottleneck — and CHANGEReadyX™ diagnoses readiness first.Chapters:00:00 Cold Open — A Faster Caterpillar Is Not a Butterfly01:07 Two Reference Points — Past vs. Future03:35 The Questions Give You Away05:21 AI Optimisation vs. AI Transformation08:35 From Optimisation to Transformation12:04 The Closing QuestionRun the CHANGEReadyX™ diagnostic: https://tapasadvisory.com/diagnosticSubstack: https://open.substack.com/pub/ramchangereadyxApple Podcasts: https://podcasts.apple.com/us/podcast/rams-changereadyx-from-tapas-advisory/id6831708010— Ram Narayanan, Transformation Leader | Architect of CHANGEReadyX™, TAPAS Advisory
Gartner predicts that more than forty percent of agentic AI projects will be cancelled by the end of 2027 — and the cancellations will not happen because the technology does not work. Gartner names three reasons: escalating costs, unclear business value, and inadequate risk controls. The agent works in the demo; the project dies in the budget review. Diagnose readiness before deploying a single agent.
In this episode: the AI Failure Handbook — hype first, no real business problem, no baseline, a rush to agentic, a shiny solution with no buyer, and a sixty-slide deck claiming success (Dr David Hardoon, November 2026); why skipping process mapping is how agentic AI dies (Yvan AGNES, November 2026); the readiness gap — 77% of CEOs expect AI industrialisation to have a medium or high impact through 2027, only 20% report high readiness, and 44% struggle to demonstrate AI financial returns (Gartner 2027 CEO Survey, published 20 October 2026); and the people-first discipline that breaks the script.
Three takeaways:
1. The cancellation is written in advance — Gartner (June 2025) predicts over 40% of agentic AI projects will be cancelled by the end of 2027, for escalating costs, unclear business value and inadequate risk controls; most current projects are early experiments driven by hype rather than a defined business case.
2. The failure follows a handbook — hype first, no real business problem, no baseline, a rush to agentic, a shiny solution with no buyer, and a sixty-slide deck claiming success (Dr David Hardoon, "AI Failure Handbook", November 2026); skipping process mapping is how agentic AI dies (Yvan AGNES, LinkedIn, November 2026).
3. The readiness gap is the story — 77% of CEOs expect AI industrialisation to have a medium or high impact through 2027, only 20% report high readiness, and 44% struggle to demonstrate AI financial returns (Gartner 2027 CEO Survey, 203 CEOs, published 20 October 2026); 42% of companies abandoned most of their AI initiatives in 2025, up from 17% the year before (S&P Global Market Intelligence, 2025); and 95% of organisations get no measurable return on the profit and loss from generative AI pilots (MIT NANDA, July 2025).
Chapters:
00:00 The Number Nobody Can Ignore
02:40 The Failure Handbook
05:50 The Readiness Gap
08:23 Why the Script Repeats
10:30 The People-First Discipline
Run the free readiness diagnostic: https://tapasadvisory.com/diagnostic
More from Ram: https://open.substack.com/pub/ramchangereadyx
Ram Narayanan – Transformation Leader | Architect of CHANGEReadyX™, TAPAS Advisory