Where should a business start with AI? Three things from the Ecosystem Alpha podcast
Three things from the Ecosystem Alpha podcast: start with what's blocking growth, not the tech; why bad data is more dangerous with AI; and why you only need to be 30-40% of the way there before you argue with it.
Subscribe to my newsletter to see content first:
I joined the Ecosystem Alpha podcast discussing everything from Behavioural Marketing to Agentic Orchestration, with Sebastian Hoelzl and Will Schnabel. You can see their write up here.
We covered a lot in just over an hour, so here are three things I think every CEO should hear before they start with AI.
Start with what’s in the way, not with the tech
Most businesses ask “how do we implement AI?” The better question is “what is getting in the way of our growth, and can AI help fix it?”
On one engagement we saved a company 42% on one of its big operational processes. We didn’t go in looking for places to put AI. We audited the business first.
Most companies do it the other way round. Someone well down the organisation is brilliant with AI and has automated their own job.
It moves the needle a tiny amount.
We also got into the people side, and where that 42% ended up. Hint: it wasn’t a cost saving.
Bad data, delivered with confidence
Garbage in, garbage out is an old phrase. With AI, the garbage comes out sounding completely sure of itself.
Someone I know was running analysis through AI on data where only about 40% had survived the import intact. For a couple of weeks it happily told them what to do and why.
The numbers for the business dropped.
It took someone challenging the AI before it owned up. I won’t spoil what it said.
Get 30-40% of the way, then argue with it
Ask an AI for a number between 1 and 10 and most of the time you’ll get 7. It gives you the most common answer, and it will happily agree that your bad idea is a great one.
That’s how everyone ends up sounding the same.
You only need to be 30-40% of the way to where you want to get. The difference is knowing that’s where you are, then going back and forth with the AI until you reach 90-95%.
Also in the hour
- Why I haven’t logged into HubSpot in a month, and I use it every day
- The EU’s new rules on labelling AI content. I had to stop myself ranting about the EU AI Act. Twice.
- Whether the MQL still exists
- How I follow people, not podcasts. Sebastian and Will are both on the list.
The full episode is on Ecosystem Alpha, Spotify and Apple Podcasts. Thanks to Sebastian and Will for having me.
If you’d like a conversation like this on your podcast or at your event, drop me a note.
About Riaz
I've spent over 25 years founding and scaling B2B companies - as COO, CEO, CMO and CTO, often across the same company's lifecycle. An engineering degree from UCL and an MBA from Bayes means I understand both the technology and how a business actually runs around it.
I co-founded Digital Oxygen, which was acquired by Silverpop and later by IBM. I scaled Profusion from 15 people to 100. I founded and led Radiate B2B for nine years, building the first MCP-based AI connector for LinkedIn Advertising, before closing it in 2026 rather than rebuild around a declining display-advertising market.
Today I run Connected Paths, an AI implementation consultancy, and work directly with a small number of B2B founders and CEOs through advisory, fractional COO, CMO and CTO work, mentoring and speaking. I've mentored 50+ founders through Techstars, UCL and Bayes, and chaired the Cass Entrepreneurs Network for its 3,000 alumni and investors.
Read the full storyRecommended posts
The three questions every CEO must answer before adopting AI across the business
Most AI conversations happen at the wrong level. Here are the three questions I ask every CEO before a single tool or roadmap gets discussed.
Read moreDoes employee monitoring improve behaviour? The research
Ellison says more cameras mean better behaviour. Workplace-monitoring research says visibility helps in narrow, bounded cases - and erodes trust and performance when used to control people over time.
Read moreWhat does a fractional COO do in a B2B startup?
A fractional COO works in your business one or two days a week to put the systems, processes and structure in place so it can scale - then hands the day to day back to your team. Here's what that looks like in practice.
Read more