On 5 November I spent an evening on Zoom with translation students at Carnegie Mellon University. The class is called Translation as a Profession, and I’d been asked to talk about careers in the language industry. I got the time zone wrong, I had a sore throat, and the glue on a brand new tooth was still drying. It was great fun.
I opened with something Brian Cox once said to me over a drink: “A good conversation is better than a lecture.” Then I gave them the three ideas the whole evening hung on:
- Learn to sell.
- Adapt to the tech.
- Build your own brand.
These are my notes, for the students and for anyone else who hires, manages or works with translators.
Why stay in the language industry?
I’ve worked in it since the early 1990s, as a graphic designer, a project manager, in IT, in sales, as a general manager and as a CEO. The only job in a language company I haven’t done is the clever bit, the translating. I keep trying to leave. Like Michael Corleone in The Godfather, every time I think I’m out, it pulls me back in.
The reason is the people. The language industry is full of people with brains the size of planets: well educated, international and curious. I’ve done some work in construction recently, and it’s a very different crowd. In our industry you make friends all over the world, and you get both sides of every story, because you have friends on each side of every argument going.
I was in Edinburgh a couple of weeks ago and could hear 25 languages on one bus. London, New York and every other big city are the same. We’re all mixed up, a minestrone of nationalities, and people will always want to talk to each other. I’m bringing my own children up bilingual in French and English because I think it’s one of the best gifts you can give anyone.
The downside is the hours. It was past ten at night in the UK when I started talking.
Will there still be translators in five years?
I asked the room, and they’d just done their degrees, so I knew what they’d say. Yes, in some form, even if the work changes.
When I was their age, people in the pub or on the golf course would ask what I did. I’d say I ran a translation company. The first question was always “how many languages do you speak?” The second was “I suppose there’s an app on your phone for that now?” No. C-3PO doesn’t exist. There’s a human element to the work that people outside the industry don’t see.
For as long as I can remember, the industry has had a conference slide asking “is translation commoditised?” We’d all agree that it was, then go and sip champagne at the next conference in Hawaii and complain there was no money in it. There’s plenty of money in it.
I’ve watched the industry turned upside down four times, and each time someone said it was finished.
Design went digital (early 1990s). I’d just come out of university as a graphic designer, trained on Letraset and paste-up. Then the Apple Mac arrived, at something like £20,000 for a full set-up, and you had to buy every font separately. I thought my career was over before it had started. In fact there was more design work than ever.
The internet arrived (mid 1990s). The secret weapon of a language company was a Rolodex. If we needed Spanish, we phoned Miguel. Search looked as if it would cut out the middleman. Instead there was more content, more languages and more work.
Google and SEO (2000s). A whole business in multilingual SEO and web localisation grew up, and AI search is now changing it again.
AI (2022 onwards). The current monster. A lot of the fear comes from the way AI is marketed, by companies warning it might wipe us out while asking for trillions of dollars for data centres.
The tools inside translation tell the same story. In 1996 one of my companies won the translation contract for HMRC, the UK tax authority, and we went to Salford to meet their in-house translators. They wrote every translation out by hand, in pencil. It then went to the typing pool to be typed up and filed. It felt like going back to the 1940s. We brought in translation memory and a translation management system, and they looked at us like wizards from the future. When I first saw Google Translate, years later, I thought I might as well jump off the roof. A decade after that everyone was talking about DeepL. Today it’s large language models, and in five years it’ll be something none of us can imagine yet.
From pencil to AI
- Pre-1990sPencil and the typing poolEvery word by hand, typed up by someone else
- 1990sTranslation memoryNever translate the same sentence twice
- 2006Statistical MTGoogle Translate: free, rough translation for everyone
- 2016Neural MTGoogle, then DeepL: fluent enough to post-edit
- 2017The TransformerBuilt for translation, became the engine of AI
- 2022+LLMsAI dubbing, AI interpreting, translation inside everything
Every step, people said translation was finished.
Why does translation cope with change so well?
Because the industry works on the future. Translators spend their working lives on next year’s products: the next phone, the next laptop, the next drug launch. Translation brings the future to international markets, so people in it see change coming before everyone else and get used to it. It doesn’t feel special when you’re in it, but it is.
A few months ago I taught a group of accountants. I love accountants, but their world changed once, when the spreadsheet arrived, and has barely moved since. When the pandemic sent everyone home, the language industry shrugged and carried on, because it had been working remotely and across time zones for decades. Friends in other industries went through months of chaos. Resilient, adaptable and comfortable with technology: that’s a pretty good description of the employee every company wants.
The work stays human, though. I told the students to think about the human side of every project they take on. Communication is a deeply human thing, and a career gets built brick by brick, one project and one job at a time. We can get robots to do robot work. People are needed for the rest.
Did AI really grow out of translation?
Pretty much. The L in LLM stands for Language, and language is our home turf. The T in GPT stands for Transformer. The Transformer was introduced by Google researchers in 2017 in Attention Is All You Need, and they built it and tested it on machine translation, English to German and English to French. The engine behind today’s AI started life as a translation model.
So when large language models arrived, people in our industry understood them quickly, because we’d been living with translation memory and neural networks for years. For me it was a case of “oh, that’s what that is”, and getting on with it. My accountancy friends still don’t quite get it.
That’s an advantage, because most people aren’t using AI well yet. A National Bureau of Economic Research survey of nearly 9,000 senior executives in summer 2026 found that 80% use AI regularly, but for an average of only 2.1 hours a week. A McKinsey survey found just 1% of leaders called their company’s AI rollout mature. Plenty of people say they use AI. Not many use it to create real value, and most never get beyond rewriting LinkedIn posts. A young, multilingual person who really knows these tools walks into most companies already ahead of the boss. If you want somewhere to start, I’ve written about learning AI by playing with it.
Technology is moving as fast as it ever has, and it’s slower now than it will ever be again. My slide put the consequence simply: you can’t translate professionally today without technology. I told them an accountant should know every shortcut in Excel, and a translator should know the translation technologies inside out, whether that’s Trados, memoQ or whatever the client uses, plus the AI tools beside them. Different tools suit different jobs, and there’s a lot more to life than ChatGPT.
Why would a linguist want to sell?
Because if you don’t sell your work, someone else will, and they’ll keep most of the money. When I ran my language company, if we paid a translator 4p a word, we charged the client 10p and tried to get close to 15p. At best the translator saw 40% of what the client paid. Most of the students were planning to freelance through agencies, and I’m not sure many had done that maths.
Who keeps the money?
I know most linguists would sooner stick pins in their eyes than sell something. I resisted it for years myself, and only moved into sales about 20 years ago. So I put a question to the group: which is easier, teaching someone sales skills, or teaching a salesperson ethics?
The sales process can be taught in a day. Translators arrive with the ethics already: they come from the right place, wanting to help people understand each other. Sprinkle a little sales process on top and they make some of the best salespeople you’ll ever meet. All the really good ones I’ve met in my career came from linguistics or project management.
Why do two people get paid so differently for the same work?
Often the difference is the position. Water falls from the sky for free. It costs a fraction of a penny from the tap, 50p at Costco and £1.20 in the supermarket. In a nightclub it’s £8, and on a plane it’s £10.50. Same water.
Same water, different price
A friend of mine bought a £25,000 Rolex after his divorce, as men in their fifties do, along with the motorbike. There might be $100 of steel in it and 30 hours of skilled work. There’s probably more steel in a horseshoe, and a horseshoe costs about ten dollars. My own watch cost $120 and measures my heartbeat.
So if you’re not being paid what you’re worth, sometimes the work is fine and the position is wrong. Moving to where you’re valued is a perfectly good answer. If a large agency keeps squeezing your rates, that’s a signal.
What does Popeye have to do with selling translation?
The first page of the sales training book in all my companies was Popeye. Popeye eats spinach and has the best weekend of his life: he lifts the 500 pound kettlebell, knocks out Brutus and gets the date with Olive Oyl. If we sat on the board of Spinach Incorporated, our website wouldn’t say “it’s green, it comes in a can and your children will never eat it.” It would say “eat spinach and have a great weekend.”
Academics call this value in use. The value of a pen is created by me as I write with it over its life, rather than at the moment Amazon delivers it. Translation works the same way. Our industry often thinks of itself as a cog in the content supply chain, and that’s fair enough. But when you sell it, you talk about helping people communicate, helping a company sell in new regions, reaching a diverse home market, or saving lives in healthcare. It takes MBA students about six weeks to really get this idea. I suspect translators would get it faster. I’ve written more about it in value-based selling and SPIN Selling for language companies.
One of my companies did the Central European translation for Tesco, the UK’s biggest supermarket chain. I talked to them about how much business they were doing in those markets, rather than errors per million words or which translation memory software we used. What mattered was their customer, and their customer’s customer. In a negotiation, I want to be talking about other people’s fees, not my own. Procurement will always try to boil a service down to one number on a spreadsheet so that the lowest price wins. Talking about outcomes moves the conversation somewhere else.
When I hired young designers, they’d come out of university full of enthusiasm and wanting to design things. We taught them context: a chair sits in a room, the room in a house and the house in a street. Packaging is designed to sit on a shelf in front of a shopper. Translation has context too. Sometimes you’re saving lives, sometimes you’re helping stop a war, sometimes you’re helping a company sell more product. That’s the story worth telling a client.
It’s also the story worth telling at an interview. The person on the other side of the table is looking for patterns: what you did, and what happened because of it. Your company might not value that, and then you might have to move. Either way, I told the students to keep a journal of the outcomes, because in two years they’ll want to tell that story somewhere else.
Where did the best sales lesson of the evening come from?
A sauna, that morning. I knew I’d be working late, so I’d gone to the gym early. My gym has just put in a sauna, and at eleven in the morning I thought I’d have it to myself. There were two builders in there, enormous men, all muscle. One fitted windows and the other fitted floors, and they were comparing notes on what worked for their marketing: leaflets for one, Facebook adverts for the other.
When the floor man left, I said to the window fitter that I hoped he didn’t mind, but I’d enjoyed their conversation. He asked what I did. I said I lectured at Cranfield. He asked where I lived, then how old my house was. Then whether I had a conservatory. Then how old the conservatory was, and whether I’d thought about replacing it. Within about 45 seconds I was sitting there, naked and sweating, realising I was being qualified as a sales prospect.
I stopped him and asked where he’d learned to sell. He looked puzzled. He wasn’t selling, he said. This was just what he did. I left with his business card, and he’s coming round to quote for a new conservatory.
That’s lead qualification, done perfectly and without a script, in under a minute. Always be closing. So I set the students a challenge: talk to one new person every day, practise how they describe what they do, have a way to leave people with their details, and get used to rejection.
People will ask what you do, the way the King asks people in a line-up. “I help healthcare companies talk to patients in their own language” lands better than “I’m a translator.” You never know who you’ll meet in a sauna, on a bus or on a plane. And it takes something like a hundred noes to get to a yes, so every no gets you closer.
I also told them that LinkedIn is a cesspit we all have to use anyway, and that their alumni network is pure gold. I did my degree 25 years ago, and the people I studied with now run things. They still take my calls.
Arnold Schwarzenegger spent as much time promoting his films, on talk shows and at events, as making them. In his book Be Useful, in a chapter called Sell, Sell, Sell, he wrote: “You can have the most amazing idea, the most fantastic plan, the best in class of virtually anything. But if nobody knows it exists or what it is, then it’s a waste of time and effort.” Being excellent isn’t enough. You have to tell people you’re excellent, and never get bored of telling them. I know it goes against the grain for most linguists. If you don’t, someone else will.
For anyone who wants to study selling without reading a book, I use films in sales training, because salespeople have short attention spans. The coffee-is-for-closers speech in Glengarry Glen Ross is a famous HR nightmare, and “sell me this pen” in The Wolf of Wall Street shows how to create demand. If you’d rather read, there’s my B2B sales reading list.
Does it matter how you speak?
More than it should. Patrick Winston gave a lecture called How to Speak at MIT for decades, and I watch it at least once a year. It’s the best hour you’ll spend on public speaking, and he’s hilarious. The tragedy, as he says, is that you’re judged on how you speak, then how you write, and only then on the quality of your ideas. It’s perverse, but you could be Albert Einstein and still get nowhere if you can’t put your ideas across.
My own mentor at university teaches marketing with no slides at all. He walks in, draws diagrams on the whiteboard, and gets ten out of ten from every class. I’m not good enough to do that yet.
The best presentation I ever saw was at a board meeting 10 or 15 years ago. A general manager stood up with no slides, just a flip chart. He drew a perfect outline of Germany and wrote the sales figures and projections on it, then flipped the page and did France, then the UK. In the pub afterwards I asked how on earth he’d done it. He’d sneaked into the room the night before and drawn everything in faint pencil. In the meeting he just went over the lines. Anything that isn’t death by PowerPoint will stand out.
I’ve written much more about this in Tips to give a great presentation.
Is everyone else as confident as they look?
No. Everyone is faking it. I’ve coached plenty of senior people, and everyone is about six years old inside. We just pretend to be grown-up corporate people. So there’s nothing to feel bad about in starting out, or being young. I’d give my right arm to be their age. My advice was to jump in at the deep end and learn to swim.
What do you actually own?
Your name. You’ll come and go from companies. Companies will come and go too: they die, disappear and change. Technology moves on. The one constant is you, and your reputation.
I have a chip on my shoulder about marketing. I always say marketers are the voice of the customer inside the organisation, rather than the people who blow up the balloons for the board meeting. Branding deserves the same respect. I think of it as your way to build a small monopoly: something about you that is different enough that nobody else can quite replace you. Everything you do adds to it, brick by brick.
A bag of coffee from Costco costs me about $20 and lasts three months. Starbucks charges a fortune for a cup, and I’ve read that the paper cup can cost more than the coffee inside it. The farmer in Brazil or Guatemala doesn’t make the money. Starbucks does, and the reason is branding. If you want to earn a quarter of a million dollars a year, your personal brand is what gets you there. Seth Godin writes about this better than anyone, and This Is Marketing and Purple Cow are both short enough to read in an evening.
A reputation is fragile too. Warren Buffett is often quoted as saying it takes 20 years to build a reputation and five minutes to ruin it. Volkswagen is still paying for Dieselgate a decade on. It takes a lifetime to build, so it’s worth protecting. One practical point came up in the questions: employers will search for you online, and some recruitment systems now pull in what you’ve posted automatically. It’s worth making sure what they find is professional.
What does a freelancer need to know about business?
Quite a lot, because a freelancer is a business of one. I used to have 50 people reporting to me, which meant 50 different problems and 50 different mortgages to pay. Now it’s just me, and I like it. If I don’t want a busy day, I don’t have one. Some people are driven to build a company, and that’s great, but it can drive you crazy and turn your hair grey. I’ve written separately about the traps family businesses fall into.
Whichever way you go, you need the language of business: how the money flows, how tax works, what you really earn per hour, and how marketing and strategy work. If you don’t learn it, you’ll get fleeced. You don’t need an MBA. Mine mostly taught me which book to look in, and more importantly which person to phone when I’m in trouble.
At university we talk about T-shaped careers: deep expertise in one discipline, with enough breadth across finance, sales, the law and strategy to see the bigger picture. Look at the core subjects of any MBA and you have the top of the T. Specialising in linguistics is wonderful, just not at the cost of everything else. With the whole of human knowledge a search or an AI prompt away, breadth has never been easier to get.
How do you learn resilience?
Mostly by getting a few knives in your back, which isn’t much help when you’re 22. One thing I teach sales teams is to keep a win journal. Winning and losing both come in streaks, and it’s mostly psychology. So when you win a deal, you write it down: what you did, and the value it created. On a bad month, you dust off the journal and read it, and the feeling comes back. Once you’ve won one, you’re back.
I also gave the students a short toolkit:
| In a crisis | Reading | Practice |
|---|---|---|
| The Stockdale Paradox: face the brutal facts, and keep faith that you’ll come through | Man’s Search for Meaning, Viktor Frankl | Go looking for rejection, by asking for things you expect to be refused |
| The 3 Ps: it isn’t all personal, pervasive or permanent | The Obstacle Is the Way, Ryan Holiday | Keep a CV of failures. You survived every one |
| Control what you can, and let the rest go | Option B, Sheryl Sandberg and Adam Grant | Look after the basics: sleep, exercise and people you can call |
| Rejection Proof, Jia Jiang |
What did I leave them with?
Three last thoughts. The first was to keep learning. Most graduates finish their degree, close their folders and put them on the shelf. Taking even one class a year puts you in the top 5% of people. I’m practising what I preach: I’m going back to university in 2027, at 52, to study AI at Cambridge, which has been a lifelong dream.
The second was to look after the billion-dollar machine inside their heads: getting outside, gardening, yoga, playing an instrument, or sitting down to read a book with a cup of tea.
The third was that time goes quickly. I’d put Statler and Waldorf, the two old men from The Muppets, on one of my slides, because I kept saying things like “back in 1992”, which is probably before some of their parents were born. Thirty years goes in a flash. When I was their age I didn’t believe that. It’s worth stopping to smell the flowers along the way.
I finished on a line from Miles Kington: “Knowledge is knowing a tomato is a fruit. Wisdom is not putting it in a fruit salad.” AI will give everyone the knowledge. The careers will go to people with the wisdom, and the ability to sell it.
Thank you to Meredith and the students at Carnegie Mellon for a brilliant evening and great questions. I’m always available to talk to students, at Carnegie Mellon or anywhere else. If any of you think of something in a few weeks’ time, get in touch.
Sources: Vaswani, A. et al. (2017) Attention Is All You Need, Advances in Neural Information Processing Systems; Yotzov, I., Barrero, J.M., Bloom, N. et al. (2026) Firm Data on AI, NBER Working Paper 34836, revised October 2026; McKinsey & Company (2025) Superagency in the workplace; Winston, P. How to Speak, MIT OpenCourseWare; Gallo, C. Arnold Schwarzenegger reveals 4 communication tactics, Inc., on Schwarzenegger, A. (2023) Be Useful, Penguin Press; Godin, S. (2018) This Is Marketing, Portfolio, and (2003) Purple Cow, Portfolio; Collins, J. (2001) Good to Great, Random House, via Wikipedia, James Stockdale; Seligman, M. (1990) Learned Optimism, Knopf; Frankl, V. (1946) Man’s Search for Meaning; Wikipedia, Translation memory; Wikipedia, Volkswagen emissions scandal; Wikipedia, C-3PO; Wikipedia, Statler and Waldorf; Wikipedia, Miles Kington.