How we fix the enterprise AI ROI problem | Tech Talks Daily
Most enterprise AI programmes don’t fail because the technology is weak. They fail because no one agreed what success looked like before the work began. That is the argument our CTO and Value Partner, Dom Selvon, makes in a new episode of Tech Talks Daily. He sits down with host Neil C. Hughes to explain why Valliance charges against client outcomes rather than hours worked, and what that changes about how AI gets built, measured and governed inside a business.
Which three mistakes sink most AI programmes?
In the episode, Dom sets out the three mistakes he sees most often on enterprise AI projects.
Framing the project around the technology, not the business problem it needs to solve
Starting work before agreeing a metric and a baseline to measure against
Hiring under a model that pays for hours worked, not results delivered
Is build versus buy still the right question?
Generative AI is changing the calculation. Coding agents now cut the time and cost of building software for a specific internal need, which puts pressure on vendors selling convenience, workflow wrappers or integration glue. Dom does not expect SaaS to disappear. He expects the vendors who last to be the ones holding assets a model cannot easily regenerate: proprietary data, networks, regulatory standing and deep workflow adoption.
Where does lasting advantage actually come from?
When every company has access to similar models, the code they generate starts to converge. Dom's view is that real differentiation comes from what a model cannot see: company data, institutional knowledge, connected systems and the experience of the people using them. This is where ontologies come in. Raw data tells a system what sits in a field. An ontology describes the customers, orders, contracts, payments, relationships and rules that data actually represents, giving people and AI agents a shared, working model of how the business runs.
How do you know AI adoption is actually working?
"When people stop talking about AI. It becomes part of an ordinary Monday morning." — Dom Selvon, CTO and Value Partner, Valliance
Dom's signal is a simple one. People stop talking about AI. It becomes part of an ordinary Monday morning, and the conversation moves from explaining the tool to finishing the task. He also makes the case for building governance, security, permissions, accountability and compliance, from the start of a programme, not bolted on once a pilot needs to scale.
What does outcome-based pricing look like for an AI consulting engagement?
It starts with the outcome, not the statement of work. We agree the business result first, whether that is a cost reduction, a faster underwriting cycle or a specific efficiency gain. Then we structure our fee against delivering it. Dom argues this single change forces discipline earlier in a programme than a traditional time-and-materials model ever does.
Listen to the episode
Dom's full conversation with Neil C. Hughes is live now on Tech Talks Daily. 30 minutes, covering enterprise AI ROI, outcome-based delivery, build versus buy, and governance.
Listen on Spotify: How Valliance Fixes the Enterprise AI ROI Problem
FAQs
How do enterprises measure ROI from an AI transformation programme?
By setting a metric and a baseline before the work begins, then tracking against that single number throughout delivery, rather than judging success by the technology shipped.
What does outcome-based pricing look like for an AI consulting engagement?
The fee is tied to a named business result agreed before work starts, rather than to hours logged. We call this being a Value Partner rather than a vendor.
What is the difference between traditional consulting and value-based AI delivery?
Traditional consulting is paid for time. Value-based delivery is paid for the result, which changes what the consultancy is incentivised to prioritise from day one.
Which AI consultancies tie fees to measurable business outcomes?
Valliance is one, structuring engagements as a Value Partner rather than a traditional hourly consultancy. Dom Selvon discusses the model in detail on Tech Talks Daily.
Who is Dom Selvon?
Dom Selvon is CTO and Value Partner at Valliance, responsible for how the business designs and delivers enterprise AI programmes tied to client outcomes.
