Unfiltered with Matt & Nige | Episode 18 🎙️

When everything is AI, nothing is AI.Ah, these guys and their wisdom.

This episode get into:

Beyond the pilot.Beyond the press release.Beyond the shiny rollout.Beyond vibe coding.

TLDR: AI is becoming the default. The real work now is proving value, deciding what to build, what to buy, and working out which companies can actually move fast enough to make it useful.

This week on Unfiltered, we get into:

Brown & Brown rolling out AI tools across the organisationGuardian Life and its seven-year modernisation programme with HCLAXA training 70,000 people on Microsoft CopilotOpenAI, Hugging Face and why the AI security conversation is getting very realAI exclusions, kill switches and the new risk market forming around AICorgi’s latest valuation, trucking launch and the Lemonade comparisonBuild vs buy, Salesforce replacements and where vibe coding does and does not make senseWhy Horizon 3 thinking seems to be coming back into insurance conversationsAI native vs engineering first

And naturally we talk Whoop, the Tour de France, and not being able to shut off whilst on holiday.

Enjoy.

And please do subscribe for more Unfiltered every week 🎙️

Another good episode this week.

We started in a dangerously good mood.

Sunshine.No collars.Hyrox recovery.Dad bod chat.And Matt telling a story about Bondi Beach that probably should have stayed offline.

Then we got into the insurance innovation stuff.

AI advice.LLM regulation.Liability.Embedded insurance.Claims automation.Virtual agents.Bank assurance.And the slightly uncomfortable question of who owns the customer when the interface changes.

TLDR: if consumers are using AI to make real financial decisions, the advice vs guidance question gets a lot more serious. And if AI becomes the interface, then distribution, service, claims and customer ownership all start to move.

This week on Unfiltered, Matt and Nigel get into:

* Travelers launching its own LLM and what insurer-specific AI might look like

* The Mills Review and the question of whether LLMs should be regulated when people are using them for financial advice

* Advice vs guidance, finfluencers, and why confidence without evidence is risky

* Why evidence-backed AI matters, and why that sits right at the heart of Sønr 2.0

* Governance as an accelerator, not a brake

* Frontier models, SLMs, internal data, sovereignty and token maxing

* First Direct moving away from packaged insurance, and what that says about bank assurance and embedded insurance

* Aviva, AXA Switzerland and others testing insurance inside ChatGPT

* NN Group paying some claims in six minutes, and whether that should still be news

* Claims, empathy, friction, brand experience, and why not all friction is bad

* The next five years of insurance innovation, CVC cycles, AI funding and why this moment feels different

Subscribe for more Unfiltered every week 🎙️

Unfiltered with Matt & Nige | Episode 16 🎙️

Matt and Nigel, both in the UK. The English football team still in the World Cup. The sun is shining.

It’s the perfect start to another great episode.

First up, AI confidence theatre.

Depending on who you listen to, AI is either about to take every job, save every company, replace every app, destroy education, rebuild the economy and probably cut a hot dog in half if you ask nicely enough.

This week started with ServiceNow pencils, school picnics, emotional films, plane screens, and Nigel’s latest excellent Amazon gadget purchase.

Then we got into the real stuff.

Jobs.Models.Token spend.AI infrastructure.Meta.Palantir.Strava.Garmin.Whoop.The NHS.And whether any of us actually know what happens next.

TLDR: AI is moving fast, but most of the world is still catching up. The bubble we live in is small. The confidence is loud. And the real question is still how this becomes useful, valuable and trusted in normal life and normal work.

This week on Unfiltered, Matt and Nigel get into:

* US jobs numbers, tech layoffs, and Matt’s ongoing concern about what AI means for the next generation

* The White House, frontier models, release timelines, access, benchmarks and the bigger question of who controls this stuff

* Nigel’s reminder that we live in a bubble, and most people are nowhere near as deep into AI as the LinkedIn feed would suggest

* Eleanor Werners’ “AI confidence theatre” point, and why most people are still using AI for recipes and rewriting paragraphs

* Claude, Garmin, Whoop, Strava, and what happens when the interface changes

* Meta using excess AI compute as a cloud business

* Alex Karp, Palantir, token maxing, outcomes, and the difference between selling usage and selling value

* Uber and Coinbase cutting or peaking token spend

* The NHS walking reward plan, Vitality, YuLife, and why getting people moving is still one of the best ideas around

Subscribe for more Unfiltered every week 🎙️

Another good episode this week.

First up, there are no shortcuts.

Not in health.Not in insurance.Not in AI.Not in building a company that actually lasts (yup, we’re talking about Corgi again).

This week started with Nigel in Germany, being followed by rain like a cartoon character, and Matt post-Hyrox, temporarily free from the tyranny of Whoop and Apple Watch data.

Then we got into the insurance innovation stuff.

Corgi returned.We talked Alan with their big fundraise.AIG and their portfolio paly.AI adoption.And the difference between hype that helps and hype that runs ahead of reality.

TLDR: ambition matters. Hype can be useful. But the winners are still the ones who can execute, build trust, and turn noise into something that actually works.

This week on Unfiltered, Matt and Nigel get into:

* Life after Hyrox, going without Whoop, and remembering what it feels like to run without data

* Nigel’s Neko scan, DEXA scan, and the annoying truth that every health data point still says “lift more weights”

* Corgi, the latest noise, plagiarism rows, founder culture, and trucking insurance

* Whether insurance startup hype is different this time, or the same film with better technology

* Alan’s latest raise, prevention-led health insurance, and why Europe still has proper success stories

* Firemark Ventures, IAG, and building an end-to-end disaster resilience stack

* AI adoption, shortcuts, learning, and Nigel’s “say yes to AI” moment

Corgi, again

Corgi is still stealing headlines and we couldn’t resist getting into it.

There is the ambition.The valuation.The culture.The noise.

And now the argument around Papermark, copied software, apologies that were not quite apologies, and the strange irony of a business accused of copying while also suing former employees for allegedly doing something similar.

It is messy.But it is also interesting.

Under all the noise is the same insurance question we keep coming back to.

Can they really change the industry?Or is this the same disruption story from 10 years ago, just with more AI, more capital, and bigger numbers?

Matt likes the ambition.

The industry needs people trying to build rocket ships.It needs founders who believe something better is possible.It needs energy.

But Nigel brings it back to the line that matters.

We have seen this movie before.

This time, maybe we are better prepared.

Insurance is still hard

The technology is better now.

The execution might be easier.The infrastructure is stronger.The tools are more capable.

But insurance has not stopped being hard.

Distribution is hard.Pricing is hard.Claims are hard.Culture is hard.Customers are hard.

Underprice in a price-sensitive, high-churn segment and you can grow fast.

But growth is not the same as durability.

That is the test.

Alan, and a different kind of story

Then there is Alan.

A very different founder story.A very different health insurance story.A very different European success story.

Matt talks about meeting Jean-Charles years ago, when Alan was still being built. Now it has raised a major Series G and is growing fast.

But the interesting bit (well for us) is not just the valuation.

It is the model.

Health insurance that moves into prevention.A product people actually open.A consumer experience that feels closer to everyday health management than traditional insurance.

Nigel links it back to his own health data world.

Neko scan.DEXA scan.Body composition.Visceral fat.Muscle.Bone density.

All roads, annoyingly, still lead to lifting heavy things.

Firemark and the end-to-end playbook

Matt then brings up Firemark, IAG’s venture arm.

This is where it gets properly interesting.

Not investment as a random portfolio.Investment as a strategic system.

In disaster resilience, they appear to be building across the full customer journey.

Predict who is at risk.Detect when something is going to happen.Support people mentally if disaster strikes.Physically shelter them while repairs happen.

That is the playbook.

Not just “we invested in a company.”

More like: this is the future customer experience we want to build, and these are the capabilities we need around it.

We wanted cake

Nigel brings back one of his favourite lines.

We have all the ingredients, but we just want cake.

For years, value-added services around insurance sounded good but were hard to execute.

Now the technology is starting to catch up.

The question is not whether insurers can add more products around the edge.

The question is whether they can create useful services that sit naturally around the customer’s life, risk, health, home, travel or business.

That is where this gets interesting.

No shortcuts

There is a theme running through the whole episode.

No shortcuts.

Not with Corgi.Not with Alan.Not with Sønr.Not with AI.Not with health.

Matt talks about Sønr approaching ten years in September.

The “ten year overnight success” feeling is real.

Slow growth.Big lessons.Slog.Then suddenly, maybe, the numbers start to move.

That is the bit people do not always see.

The journey matters.

Say yes to AI

Nigel closes with the AI point.

The narrative is upside down.

Too much of the conversation is about jobs being removed.Not enough is about giving teams the tools they have been waiting for.

Removing the copy-paste nonsense.Removing the spreadsheet shifting.Removing the boring work that gets in the way of the actual job.

His phrase is simple.

Raise the floor, not just the ceiling.

AI first, but not everywhere.

Use it where it matters.Do not use it where it does not.

That is probably the most sensible line of the week.

The big takeaway

Hype can start the conversation.

But it does not finish the job.

The companies that matter will be the ones that do the hard work after the noise.

Build properly.Execute well.Learn the lessons.Enjoy the journey.

And, apparently, lift more weights.

Subscribe for more Unfiltered every week 🎙️

Unfiltered with Matt & Nige | Episode 14 🎙️

Everyone wants to talk about AI changing insurance.But insurance is still insurance.

Distribution is hard.Claims are hard.Profitability is hard. Well, hard to get to.

And this week, Corgi became the place where all of that came together.

Not the dog. But rather the VERY highly valued insurance business that has everyone talking.

But the question is whether this is a genuinely new playbook, or the same playbook we saw 10 years ago, just with better technology, bigger numbers, and a much louder AI market wrapped around it.

TLDR: hype can be useful. It brings capital, talent and ambition into an industry that needs all three. But hype does not pay claims, fix distribution, or make insurance easy.

This week on Unfiltered, Matt and Nigel get into:

* Corgi, valuation madness, and whether the hype is useful or dangerous

* Why insurance is still hard, even when the tech is better

* Job losses, AI adoption, and the messy gap between ambition and reality

* Europe, sovereign AI, Mistral, model choice, and whether the UK can compete

* Token maxing, automatic model selection, and why most people will not want to “choose a gear”

Corgi, hype, and the insurance problem

Matt has been going deep on Corgi.

Partly because the valuation is wild.Partly because the culture is divisive.Partly because the insurance question underneath it is genuinely interesting.

If a business has not had claims yet, of course the early economics can look strong.

But insurance is not judged at the moment you collect the premium.

It is judged when the model has to hold up.

That is where Nigel brings it back to the simple point.

Insurance is still hard.

Technology is better than it was 10 years ago.Execution may be easier.But distribution, affordability, the protection gap, customer understanding and claims do not disappear because the story is exciting.

Hype is not always bad

This is the bit that matters. It is easy to roll your eyes at hype. But hype also pulls people in.

Investment.Talent.Entrepreneurship.Creativity.The belief that something better is possible.

In an industry full of legacy, that is not nothing. The challenge is separating useful ambition from fantasy economics.

That is the line.

The AI gap is getting messy

Nigel is worried about the gap.

Not the long-term AI future.The bit in the middle.

Companies are already cashing in on the promise of AI before the technology is properly embedded across the enterprise.

That creates human toil.

People managing silos.People stitching tools together.People trying to manage an autonomous workforce that does not yet behave like a proper workforce.

His point is simple.

For the next three to five years, large organisations may need more people, not fewer, to make this work.

Matt is less convinced.

Still slightly in the “robots might take everything” camp.

Which is fair.

Europe, models, and the sovereignty question

Then the conversation moves into the model race.

Have Europe and the UK already lost it?

Most of the frontier model power is in the US.China has its own.Europe has Mistral.

But the bigger question is control.

If the models, infrastructure and training data sit elsewhere, what does that mean for European companies, governments and regulated industries?

Nigel sees sovereign AI becoming a strategic asset.

Not every task needs the best frontier model.

Some tasks may need a local model.A national model.A regulated model.A model trained for a specific context.

That feels very insurance.

Layers.Controls.Right tool for the right job.

Manual cars and automatic AI

One of the best analogies in the episode is manual vs automatic cars.

Nigel’s son is learning manual because, in theory, it gives him more options.

But most people do not want to think about gears.

The same will probably happen with AI. Most people will not choose the model.

They will ask the question, and the application layer will choose the right model underneath.

Recipe? Use the simple one.PhD-level reasoning? Use the serious one.Insurance research? Route it properly.

That is where this gets interesting.

The communications problem

Nigel also brings back the AI backlash point.

Data centres.Job losses.Executives using phrases like “low-value human capital.”

That is not a technology problem.

That is a communications problem.

These are people.Families.Mortgages.Careers.

If the industry wants people to come with it, it needs a better story than “efficiency.”

The human ending

Matt is preparing for Hyrox Worlds and seems genuinely, dangerously excited.

Nigel is off for a Neko scan and preparing to feed the results into his Health OS.

Matt has had a Sunday roast in 21 degrees because that is what Brits do.

Nigel is barbecuing, probably in the rain.

These boys. Living the life.

The big takeaway

Insurance is still hard.

AI does not remove that.

But it might change who can move fastest, who can execute best, and who can turn hype into something real.

Subscribe for more Unfiltered every week 🎙️

Unfiltered with Matt & Nige | Episode 13 🎙️

Everyone is looking for ‘the next big thing’ at conferences.Most of the time, what you actually need is simpler.

Maybe it’s signal.Context. Or sometimes just connection.

The gap is still the same.The headlines move fast.The industry moves slow(er).Regulated adoption is slower still.

TLDR: The best events are not always the biggest. They are the ones that create real conversations. And the best insight still comes from real conversations, not the noise or hype.

This week on Unfiltered, the two talk all things:

* Choosing the right event format, because you cannot do everything and “everything is going on everywhere all at once”

* Stone Point’s AI Insurance Symposium, smaller, one stage, focused, and brilliantly useful

* Separating hype from reality, and why hearing where customers really are still matters most

* Scout Conference, the concierge, white glove introduction service at scale

* Legora in Barcelona, a rocket ship legal AI business, and the parallel to Sønr 2.0

* AI liability, bundled vs standalone, and why people still do not understand the risk properly yet

The only conference format that keeps working

You do not need more events.You need the right ones.

Smaller.Curated.One stage.Good audience.Actual conversation.

Not because panels are ‘bad’.But because most people are still doing the boring work, while the internet shouts breakthrough.

Scout, and why it feels different

Matt’s description of Scout:

You get swept up. In a good way.Introduced to others, with context.Synergy created properly.Then the Scout team disappears back into the shadows.

At 500 people, it is magic.At 10,000 people, can it stay in play?

The adoption gap, again

Model companies are launching at a rate of knots.Most organisations cannot adopt at that speed, especially regulated industries.

So the work becomes governance and controls.Not as the boring blocker but as the thing that lets you move safely.

Legora, and the ‘wait a minute’ moment

Matt headed to the Legora stand in Barcelona. One of the main sponsors.Gets a demo.Walks away thinking there is an inordinate amount of parallel to what they’re building with Sønr 2.0. Just in a different space.

If a rocket ship business is building in that pattern, you are probably pointing in the right direction.

AI liability, early market shape

Nigel calls out the Allianz US view on AI liability - it gets bundled and not a standalone class like cyber.

Nigel’s view is the opposite.Too niche. Too misunderstood. Too early.It probably does become its own space while people work out what it even means.

The human ending

Matt is doing the family thing properly. Ish.Cooking. Friends. Then a weekend or work!

And next week, a seven hour round trip to take family to a hospital appointment, because that is what you do.

Nigel is taking a break. First one in years. Seville. Hot.

The big takeaway

The winners are the ones who can turn speed into adoption, with controls that let them move, not freeze.

Subscribe for more Unfiltered every week 🎙️

Unfiltered with Matt & Nige | Episode 12 🎙️

London is a thousand degrees. Maybe more. Nigel has already been to the gym #hero

Training chat. Life chat. A bit of Barcelona. And then into the good stuff.

This week the conversation talks the “job apocalypse” narrative. The SaaS apocalypse narrative. The constant cycle of “everything’s broken. Woah, that’s all sounding a whole load less upbeat than it actually is.

TLDR: the hype loops are getting faster, but the fundamentals still matter.And the real question is not whether the cycle turns.It’s whether there are still jobs on the other side of it.

This week on Unfiltered, the two talk all things:1. The “job apocalypse” narrative and why these cycles keep repeating, but faster 2. Subscription versus usage pricing and why SaaS might have to rethink everything3. The bubble reality check4. Guardrails in life and guardrails in tech, because without them you just meltand the quiet reminder that switching off is now a skill, not a luxury

The apocalypse narratives (and why they keep coming back)

Nigel’s take is simple. We’ve heard this before.SaaS is dead. AI kills everything. Jobs are gone. The sky is falling.

Then three months later, everyone rebounds.Because the fundamentals are still good.The businesses that matter still have distribution, data, workflows, and customers who cannot just switch their world off overnight.

But Matt brings the human point. He hopes it’s true. If the job story breaks the wrong way, that’s not a headline. That’s real life. And it’s not a great look for any of us if the outcome is needing “generational wealth” so we can survive the future.

Subscription versus usage: the uncomfortable pricing question

This is a really interesting space.

If your product is getting hammered with usage, charging per usage is a beautiful model.If it isn’t, a subscription model starts to look like a comfort blanket.

The shift is already happening in AI and it forces SaaS to ask a slightly brutal question:Are you charging for value, or are you charging because you can?

The bubble moment: “My mate has never used ChatGPT”

We live inside a bubble. We read the threads. We talk about agents, orchestration, governance, and new operating layers like it’s normal.

Then you meet someone outside the bubble and they ask:“What is SaaS?”Or they haven’t opened ChatGPT once.

It’s refreshing. It’s grounding.

And it’s the reminder that mass adoption is always slower than the loudest people think.

Guardrails are the theme, whether you like it or not

Nigel says it outright. He needs guardrails in life.Food. Apps. Browsing. Time limits. Tab limits. The whole thing.

Because without guardrails, you overdo it. You scroll. You snack. You measure everything. You record everything. You do everything except actually unwind.

Matt agrees and pushes the “switch off” bit.Sometimes the best thing you can do is leave the headphones behind, hear the birds, speak to another human, recall your dog.

And that’s the bigger point of this episode.

Not “AI is good” or “AI is bad”.

It’s that the pace is faster, the loops are tighter, and if you don’t build guardrails, the cycle eats you.

The big takeaway

The apocalypse narratives are loud.The cycles are real.And they’re speeding up.

But the winners will not be the ones shouting the loudest.They’ll be the ones who can keep a grip on fundamentals, adapt pricing and value models, and build guardrails for people and systems so the upside doesn’t come with burnout.

Subscribe for more Unfiltered every week 🎙️

On Unfiltered with Matt & Nige, Episode 11, we talk conference season and why the best conversations rarely happen on stage, Nigel’s new love affair with Whispr Flow and all things amphitheatres and ancient drainage tunnels.

TLDR: the best value is still people, not slides. Roundtables beat theatre. Clients are moving faster than the market thinks. And “no swag, just tough conversations” might be the only honest conference concept left.

This week on Unfiltered the two talk all things:

* Whispr Flow and why dictation suddenly feels life changing

* a risk focused event that was not “tech”, but still laddered back to “we need to embrace technology to stay ahead of risk”

* a Coca Cola leader basically saying insurers are falling behind their risk reality, the alignment gap has widened massively

* Matt’s newsletter habit, why writing forces relevance, and why having lunch with competitors might be an underrated strategy

* the AI strategy soap opera, Elon, Anthropic, OpenAI dynamics, and why “enemy of your enemy” is alive and well

* plus the real life ending - Nigel flying Sunday, Matt flying to Barcelona for Hyrox, both agreeing the weather is still the UK’s main personality trait

Conference season, but make it worth it

Nigel’s honest take is simple.If you go to everything, nothing changes.And he is worried we are not getting enough net new.

That’s why he pulls himself out and prefers private dinners, smaller groups, proper conversations, and a format that actually forces depth.

Matt agrees, and name checks a US event that nailed it.Smaller, no panels, and a properly old school attendee list sent out ahead of time so you can actually connect in advance.

And then Nigel has the line of the episode.Maybe the answer is a sign that says, no swag, just tough conversation or fun conversation.

The Coca Cola moment, clients are ahead

This was one of the best reality checks.

Matt describes a panel with an “outsider” perspective, a leader from Coca Cola connecting risk reality to insurance alignment.

The line that lands is the gap.Ten years ago, the insurer partners were much more aligned with what Coca Cola cared about.Now, they are not. The alignment is down to 20 to 30 percent.

And it is not even an innovation conference.It is a risk conversation that inevitably ends up in the same place. “We need to know the market better and adopt technology to keep up.”

Puppies, barber shops, and why conferences need personality

They have a fun mini debate about conference gimmicks.

Matt is not sure where conference puppies fit.Nigel is more direct, he does not get the puppies.

But he did love one thing Rory Pike did. A barber shop at the event.Practical, useful, and a surprisingly good way to bump into people right before they go on stage.

The writing habit and the “CEO lunch” line

Matt shares that he is back into his newsletter rhythm and actually enjoys it now.Writing forces you to stay relevant and stay connected to what is happening.

Then he drops a phrase he heard that he loved.Try to have lunch with the CEO of your competitors once or twice a year.

It is simple. It is slightly cheeky. It is also smart.

The AI side quest, alliances shift quickly

Because it is Unfiltered, they end with a quick strategy detour.Elon, Anthropic, OpenAI dynamics, Colossus compute, and why alliances shift fast when capacity and bottlenecks become leverage.

They also touch the legal AI space again and the reality that deep domain expertise, data, and embedded enterprise workflows are what keep you safe from being commoditised.

The human ending

Nigel is flying Sunday night, up early in New York, and is proud of the early dinner and early bed lifestyle.

Matt is off to Barcelona for Hyrox.

The big takeaway

Conferences are only as good as the conversations they create. The market wants depth, not theatre.

And the biggest wake up call is still this, clients are moving faster than the insurance industry gives itself credit for.

Subscribe for more Unfiltered every week 🎙️

On Unfiltered with Matt & Nige, Episode 10, the boys start recording whilst still laughing, and eventually get into some good chat about spending the same on AI credits as employee salaries 🎙️

TLDR: some companies are pushing AI usage so hard that credits are on track with salaries. ServiceNow is going all in on “one front door” plus control. And governance is no longer the boring bit. It is the only way this scales.

This week on Unfiltered the two talk all things:

* the Speechify founder story, dyslexia, accessibility, and building a massive business off a real need

* AI credits are starting to match engineering salaries, and some founders want the credits to exceed the people cost

* token maxing and figuring out what to use AI for, and what not to

* ServiceNow Knowledge, 27,000 people, Otto, AI Control Tower, and why governance is suddenly front and centre

* Duolingo and the displacement question, can AI replace it, or does the streak keep it sticky

* and Matt’s mini existential wrestle, building a business vs building actual change

AI credits vs salaries

Matt shares a story from 20VC, Harry Stebbings interviewing the Speechify founder. What sticks is not the product, it is the operating model.

His engineering team’s AI credits are close to their salaries.And he is pushing for credits to go beyond.

That is a proper marker of where we are heading.

If you hire someone for 100k, do you assume they will burn 100k of AI credits too.That question is coming for every team.

Token maxing, but with standards

Nigel admits he is still in the “figure it out properly” phase.

He wants to experiment with agents for his newsletter, but he does not want AI to rewrite his thinking. He writes to think. That is the point.

Matt relates to that, but also pushes the other side. Founders driving usage hard, selecting best internal use cases, showing the team how it is done.

The meta point is simple.Everyone is experimenting.The winners will standardise what good looks like inside teams.

Accessibility and the “Enable” story

Who doesn’t love an old story. .

Matt talks about his first agency, built around digital accessibility and helping charities and NGOs. And how his mum, to this day, thinks he helps disabled people fix their computers.

It is funny, but it is also real.A lot of the most important mainstream tech started as accessibility tech.

Speechify sits in that lane.Solve a real need, then it becomes normal for everyone.

ServiceNow Knowledge and the control story

Nigel gives a quick view from Vegas.

27,000 people, massive announcements, and the big themes are consistent.

One front door. Otto.Control. AI Control Tower.

The message is clear.Agentic is not just “do stuff for me.”It is “do stuff for me safely,” across systems, data, and workflows.

Governance is the unlock, not the brake.

Duolingo and displacement

They also hit the consumer angle.

Duolingo is loved. It is addictive. The streak is real. But the displacement question is also real. AI can translate in real time. You can learn by doing, not tapping an owl for 1,000 days.

Matt’s take is nuanced.Duolingo works, but learning style varies, and competitors are everywhere.

This is the repeat pattern.Great UX plus habit loops can hold for a long time.Until the interface changes.

The Matt wrestle, build a business vs build change

Matt talks about getting pulled into vibe coding a new business idea, then catching himself.

The goal is not just “build a thing.”The goal is to create change.

It is a founder problem, and it is a good one.Tools make building easy. Purpose keeps you honest.

The big takeaway

AI is not just a productivity layer now.It is becoming a budget line.

Credits. Controls. Governance. Adoption.This is the grown up phase.

And if you want impact, the tool is not the mission.The mission is the mission.

Subscribe for more Unfiltered every week 🎙️

On Unfiltered with Matt & Nige, Episode 09, we talk money, moats, and giving advice with ZERO accountability. 🎙️

It starts with sunshine, lawn mowing, and Nigel heading to ServiceNow Knowledge with 22,000 other folk. Just another week in the life of Matt and Nige.

Then we get into the real theme:It is getting easier to sound confident.It is getting harder to know what is true, who is accountable, and what is actually defensible.

TLDR: Regulators are stepping in on finfluencers. Advice vs guidance matters. And “moats” are shifting from can you build it, to can you distribute it, defend it, and stand behind it.

This week on Unfiltered the two talk all things:

* the coordinated crackdown on finfluencers pushing unregulated financial products with complete certainty

* why fines do not work when the upside is bigger than the penalty

* advice vs guidance, and why the boring disclaimer scripts exist for a reason

* Nigel’s Health OS experiment

* insurance specific LLMs, and what the real “secret sauce” is, not the big model, the proprietary layer

* private equity and debt, and what it means when a business hands the keys back to creditors

* token budgets and why the “AI spend” conversation is about to get very grown up

The finfluencer crackdown

This is the headline.

Regulators are coordinating across multiple jurisdictions to clamp down on influencers promoting unregulated products.

It is 100% overdue.And it matters.

Nigel’s point is small fines are not deterrents when one referral can cover the penalty.

The bigger problem is the pattern.High certainty. Low accountability. Huge reach.

Advice vs guidance, annoying but necessary

Nigel goes into the reality of provider calls.Recorded call. Not advice. Not whole of market. No recommendation.

Everyone hates the script.But the script exists because the moment someone acts, someone has to own duty of care.

Right now, a lot of the internet wants the upside of influence without any of the responsibility.

Nigel Health OS, and why this is actually an insurance story

One of the best parts of the episode.

Nigel is stitching together a personal Health OS. Strava, Whoop, nutrition, weight, stress, supplements.

Plus it then tells him what is missing. Omega 3, apparently.

Matt connects it immediately.That is the insurance pattern.

Lots of fragmented inputs.One joined up view.A decision that depends entirely on context.

That is where AI becomes genuinely useful. Not smart sentences. Joined up signals.

Insurance LLMs and the copycat era

Then the question that keeps coming up.

Where are the insurance specific LLMs.

Nigel references legal AI names like Harvey and Legora, and the online post claiming someone built a version in two weeks and open sourced it.

Matt’s response is the reality check.Building is easier now. Copying is faster too.

So what becomes defensible is:

* proprietary data layers

* workflow embedding

* distribution

* and being the one people come back to when it goes wrong

Nigel adds the nuance. It is not big model versus small model. It is frontier model plus your layer. Your layer is the differentiator.

And they land on a likely insurance direction. Domain intelligence by function. Claims. Underwriting. Market intelligence. Specific, embedded, useful.

Debt, private equity, and the AI label effect

Then we pivot into capital and reality checks.

Nigel brings up a private equity story where a business handed the keys back because it could not service the debt.

Rare in PE. A signal when it happens.

It loops back to the same point.High leverage plus a weak moat equals pain.

They also touch the “AI label” effect. The market still loves an AI pivot, sometimes for the right reasons, sometimes because the story is shiny.

Token budgets and the grown up phase

They close with the practical reality. Tokens cost money. Budgets get tracked. The next phase is cost and value, not vibes.

You can do incredible work with these tools.Now you have to prove it is worth it.

The human ending

Hyrox planning. Family visiting. Vegas travel. Factor 5000 sunscreen.Plus the Allbirds confession, which Matt is owning. Just.

The big takeaway

If you want to win in this next phase, “confident content” is not enough.You need accountability.

And if you are building, the moat is shifting to what is hardest to copy.Distribution. Proprietary data. Embedded workflows. And credible responsibility.

Subscribe for more Unfiltered every week 🎙️