Tom Raftery

Tom Raftery

@tomraftery.com

Technology and sustainability analyst, keynote speaker and podcast host specialising in energy, climate and supply-chain resilience. Previously a Global VP at SAP and an analyst at RedMonk.

3K followers2.9K following1K posts

Enterprise AI is exposing a problem many companies have been quietly living with for years: data that works because humans know what it really means. Once machines start making decisions, that hidden context becomes operational risk. #ai #artificialintelligence #agenticai

View full thread

I’ve started thinking of this as “semantic debt”: years of undocumented definitions, exceptions and tribal knowledge that remained invisible because people compensated. Kismet: AI doesn’t create that debt. It starts charging interest on it.

As AI moves from analysis to action, competitive advantage may depend less on who has the smartest model and more on who can explain precisely what their own data means. I explore that argument in full here: tomraftery.com/2026/10/12/a... #AgenticAI #SupplyChain

tomraftery.comAI Is Exposing the Data Problems Humans Used to Hide - Tom Raftery.comEnterprise AI is exposing a hidden weakness: data systems that worked because people supplied missing context. As machines start making operational decisions, semantic debt becomes a business risk.

Enterprise AI is exposing a problem many companies have been quietly living with for years: data that works because humans know what it really means. Once machines start making decisions, that hidden context becomes operational risk. #ai #artificialintelligence #agenticai

View full thread

McKinsey reported in 2026 that just 7% of companies had fully scaled AI, while more than two-thirds of high-performing companies identified data as their primary obstacle. The bottleneck is shifting from model capability to organisational context.

That has practical consequences. An AI agent can spot procurement leakage across thousands of transactions. But if supplier identities, contract terms or definitions are inconsistent, it can also automate that leakage faster than a human team ever could.

Enterprise AI is exposing a problem many companies have been quietly living with for years: data that works because humans know what it really means. Once machines start making decisions, that hidden context becomes operational risk. #ai #artificialintelligence #agenticai

Take “order delivery date”. Warehouse may mean the truck has left. Finance may mean the customer has received the goods and revenue can be recognised. Both can be technically correct. An AI agent still has to know which meaning applies.

This is why “clean data” is no longer enough. Accuracy, completeness and consistency matter. But AI also needs semantics, provenance, freshness, ownership and intended use. A perfectly formatted field can still be wrong for the decision being made.

AI is exposing a dirty secret in enterprise data: many systems worked because humans quietly supplied missing context. As machines start making decisions, semantic debt, provenance and meaning may matter more than having the smartest model.

tomraftery.comAI Is Exposing the Data Problems Humans Used to HideEnterprise AI is exposing a hidden weakness: data systems that worked because people supplied missing context. As machines start making operational decisions, semantic debt becomes a business risk.

Tom Raftery reposted

USA is a belligerent petrostate China is the innovative electrostate Europe appears to be just in a State of Confusion My interview on Climate Confident with @tomraftery.com (and a call out to @mliebreich.bsky.social)

Prof Ray Wills@profraywills.futuresmart.com.au

We are living through the fastest energy transition in human history - we are still having difficulty accepting that it can be this fast We are at the pointy end now. 5 years ago, getting it wrong did not carry you terribly far into the future From here on, it will be scary if you get it wrong

Would you buy a car that disappeared when its manufacturer went bust? Wayne Scott uses that comparison to explain a risk with cloud software. I’ve put four moments from my Resilient Supply Chain podcast into this reel.

View full thread

Olaf van der Veen doesn’t mind whether a customer wants a more sustainable business or a bigger car. His argument: cutting food waste saves money either way. That leaves room for a fairly wide range of motivations.

Would you buy a car that disappeared when its manufacturer went bust? Wayne Scott uses that comparison to explain a risk with cloud software. I’ve put four moments from my Resilient Supply Chain podcast into this reel.

Procurement saves the business money. The response? “Why didn’t you find savings last year?” Spencer Penn on the peculiar experience of delivering a result and still having to defend yourself. Preventing a crisis can be an equally thankless job.

Terry Boyle says free returns incentivise bad behaviour. Then comes the difficult part: he’s not sure retailers can stay competitive without them. How do you change a practice your customers have come to expect?

Tom Raftery reposted

As a Type1DiabeticAdvocate and now ambassador heres an article i have just written ❤️ this issue is so important to me ❤️ Let me ask you can u imagine a world without Type1 diabetes Love you mary ❤️ you are my hero #Demcast #ProudBlue #Door #GTS ashleyloveswriting.substack.com/p/can-you-im...

ashleyloveswriting.substack.comCan You Imagine a World Without Type 1 Diabetes?By a Type1DiabeticAdvocate & Community Ambassador

Tom Raftery reposted

Solar, Batteries and EVs Are Turning Delay Into an Investment Risk The fastest shift in energy in human history 2026 the pointy bit If you got it wrong 5 years ago, get started now But don't, & from now on, you'll fall behind This freshly minted. Have a listen to these blokes youtu.be/W6TAoIpMPIg

youtu.beSolar, Batteries and EVs Are Turning Delay Into an Investment RiskYouTube video by Tom Raftery: Sustainability & Climate Talks

Your phone is a terrible crystal ball for an EV battery. Both bought in Dec 2024:
iPhone 16 Pro: 93% maximum capacity.
Kia EV3: 100% reported battery health. But my car isn't what settles the "new battery every three years" argument. The larger datasets are. #ElectricVehicles

View full thread

Next time someone says "new EV battery every three years", ask: Which model?
Which model year?
How many replacements, out of how many cars?
Source? Real-world data supports gradual wear and uncommon replacement, not a three-year countdown. Share the evidence, not the scare story.

Your phone is a terrible crystal ball for an EV battery. Both bought in Dec 2024:
iPhone 16 Pro: 93% maximum capacity.
Kia EV3: 100% reported battery health. But my car isn't what settles the "new battery every three years" argument. The larger datasets are. #ElectricVehicles

View full thread

Your phone is a terrible crystal ball for an EV battery. Both bought in Dec 2024:
iPhone 16 Pro: 93% maximum capacity.
Kia EV3: 100% reported battery health. But my car isn't what settles the "new battery every three years" argument. The larger datasets are. #ElectricVehicles

View full thread

Many EV batteries get serious protection: temperature management and buffers that limit how far cells charge and discharge. That doesn't make them indestructible. Heat, chemistry and charging habits still matter. www.geotab.com/blog/ev-batt...

geotab.comEV Battery Health: Key Findings from 22,700 Vehicle Data Analysis | GeotabAnalysis of electric vehicle battery health of 22K+ EVs reveals a 2.3% average battery degradation rate. High-power DC fast charging is the top risk factor.

Your phone is a terrible crystal ball for an EV battery. Both bought in Dec 2024:
iPhone 16 Pro: 93% maximum capacity.
Kia EV3: 100% reported battery health. But my car isn't what settles the "new battery every three years" argument. The larger datasets are. #ElectricVehicles

View full thread

Your phone is a terrible crystal ball for an EV battery. Both bought in Dec 2024:
iPhone 16 Pro: 93% maximum capacity.
Kia EV3: 100% reported battery health. But my car isn't what settles the "new battery every three years" argument. The larger datasets are. #ElectricVehicles

My screenshot starts the conversation; it doesn't prove the case. 100% reported health doesn't mean zero ageing. The phone and car have different usage and measurement methods. The "replace it every three years" claim needs to survive contact with much bigger datasets.

Long-duration energy storage is not waiting for one perfect invention. It needs markets that value time, contracts investors can finance and infrastructure that can deliver. My new evidence-led briefing examines what must change, and why grids are the next part of the…

A technically successful AI pilot is not necessarily a production capability. New research from the Resilient Supply Chain archive identifies four strongly supported reasons promising models struggle to create lasting operational value—and what leaders should test before scaling.

Tom Raftery reposted

Replying to @tomraftery.com

The world has lost a great light 😢 Dolly Parton donated her time, talent and treasure. To make making this a more equitable world. Her life is great example for all. Did you know ⬇️ bsky.app/profile/patr...

P Maroney@patrickmaroney.bsky.social

Did you know that Dolly Parton donated $1 million to fund pediatric infectious disease research? www.prevention.com/health/a4031... She donated 10s of millions of dollars & 100s of millions of books to various charities throughout her legendary career What an amazing human🫶 We will miss you💔

Can 2026 get any worse?

Dr. Samantha Montano@samlmontano.bsky.social

Dolly did what other disaster nonprofits are too chicken shit to do — she gave the families affected by a wildfire a universal basic income. www.bpr.org/news/2018-12...