The business lens on AI. What it is worth, what it takes, and what to do first.
About Perform Advisory
THE FIRM
Built for the investment decision, not the experiment
Perform Advisory is a specialist AI transformation advisory firm. We help organisations move from AI ambition to AI capability through structured diagnostic work, strategic advisory, proof-of-concept design, operating model design, and change management support.
Delivery is founder-led and engagements are taken selectively. The team that scopes the work runs the interviews, builds the value case and presents it to your board. There is no handover to a delivery team you have not met.
Trust and ownership. You leave with numbers your board will accept, a roadmap your teams can run, and an honest price on what production costs before it reaches your budget.
Founder and Principal Advisor
ILIJA SOVILJ
Over a decade leading transformation in complex, regulated organisations, and the business architecture of enterprise AI: strategy, prioritisation, operating model, governance, and the adoption that turns pilots into run-rate value. The discipline is commercial first and technical second. The question is never what the model can do, but what it is worth to this organisation and what has to change to capture it.
Most recent work
Built the enterprise AI advisory practice of a venture-backed AI firm from a standing start, as Head of AI Strategy and later Acting COO: the delivery approach, the operating cadence, a team of ten across consulting, technology and delivery, and the full commercial cycle from first conversation to signed contract. Designed and oversaw live AI pilots in client environments spanning analytics, document intelligence, process automation, predictive maintenance and customer-facing AI, sequenced by business impact rather than technical interest. More than 150 AI conversations at C-suite level.
Selected engagements
Infrastructure · UK
Technology strategy for one of Europe's largest infrastructure programmes. A roadmap sequencing digital twins, industrial sensing and machine learning into delivery waves, each with the investment case that justified its place in the order. Target operating model, governance and decision rights designed alongside it. Reported to the CIO.
Power and utilities · UK
Digitising customer operations at a national electricity market body. The business case, platform selection and fit-gap analysis, then paperwork-heavy processes redesigned into automated flows and proven on a working demonstration rather than a deck. Reported to the CEO.
Technology and ICT · Middle East
A three-year technology strategy that unlocked government procurement. A quarter-by-quarter roadmap across AI, connected devices and emerging technology, built to the national digital regulator's submission requirements. Accreditation achieved, opening a market the client could not previously bid into.
Food and beverage · Africa
Enterprise digital transformation for a leading regional group. A technology-enabled target operating model across supply chain, manufacturing, distribution and customer service, with the organisation-wide upskilling programme that made it stick.
Earlier career
Digital and business transformation across infrastructure, energy and utilities, telecommunications, healthcare and government, in Big Four consulting. Product leadership for a greenfield digital bank in a regulated Southeast Asian market, directing two delivery teams of twenty across product, design and engineering. Earlier still, advisory and audit for banks and insurers.
How we think about this
Why AI programmes stall, and what we do differently
01
Most AI budgets are spent on the model and lost in the organisation
02
A business case built on belief will not survive its first budget review
What decides whether AI reaches production is everything around it: who owns the data, whether the process was ever designed to be automated, whether anyone's targets changed to match. Get that wrong and the model performs perfectly while the business case never lands.
Vendor benchmarks describe someone else's operations. We build the case in your volumes, your costs and your margins, and mark each figure as measured or estimated, so your CFO knows exactly what they are signing off before they put their name to it.
03
Sequencing is
strategy
04
The fastest way to lose money on AI is to keep funding the wrong thing
Everything cannot be first, and the wrong order is expensive. A use case that depends on data foundations you have not built will spend a year proving exactly that. We map what depends on what, so the roadmap matches the order your organisation can actually deliver in.
Every engagement ends with a straight answer: build it, widen it, or stop. Where the numbers do not support the spend, we say so. That is worth considerably more than a green light from someone who was never going to give you anything else.
CONTACT
The first step is a conversation.
30 minutes. No pitch.
Bring the AI question you are weighing. You will leave knowing whether it is worth pursuing and what it would take, whether or not we work together.