Keynote · BFI Southbank
Expanding the core: how enterprise products grow without fragmenting.
Product & Experience Leader · Enterprise platforms
Evi
Malisianou
Twenty years across energy, financial services, e-commerce and intelligence platforms, designing for expert users and the people they serve in regulated, high-stakes domains.
I work as a peer to product, engineering and commercial leadership, shaping operating models, org design and strategy as much as the roadmap.
I now work AI-native: building functional prototypes in code with AI tools, designing agentic products, and shifting my team's craft from screens toward systems thinking, research and orchestration.
Point of view
The decisions that shape an enterprise product happen beneath the interface.
The hardest decisions in enterprise products sit in the architecture: where a new capability belongs, what it means for the journey, what it implies for the commercial model, how it survives the regulation around it. That layer is usually owned by no one, which is exactly why products fragment.
I build the connective layer of complex enterprise products: the architecture, journeys, research and strategy that hold them together, so the experience stays simple and human for the people who use it.
AI-native products push more of the product into that same layer. How the model behaves, how you know whether it's behaving, where trust and human control sit, what the design system becomes once AI generates against it. Those are design decisions now, made in the same place as the rest, and they're where much of my recent work has gone.
The user experience
The connective layer
The craft
Surface
Product designInterfaces grounded in architecture.
Conversational designHow agentic products talk, listen and respond.
Structure
Information architectureWhere the next thing belongs, before it's built.
Journey managementOwned end to end, across product boundaries.
System
Service designThe system around the product.
Agent behaviour & evalsHow the model behaves, and how you know it's right.
Strategy
Mixed methods researchWired into roadmap and architecture.
Product & UX strategyWhere the product goes next, and why.
The foundations
Enterprise design systemOne shared language in code, design, voice and content, and the substrate AI generates against.
Quality guardrails & standardsAccessibility and design quality, kept to a measurable bar.
Front-end QADrift from the system, caught automatically before it ships.
AI-built research engineJourney intelligence and research synthesis, built with AI.
Trust, safety & human controlSafety-critical decisions kept out of the probabilistic layer.
Front-end architectureWhere the design system meets the engineering stack.
Enterprise architecture
How I think and lead, in two assessments.
HBDI shows a strong synthesiser with an analytical spine that holds its shape under pressure. I think in systems and futures, backed by structured reasoning. CliftonStrengths adds the leadership texture: Strategic, Analytical, Self-Assurance, Individualisation, Relator.
In practice, that means I move fast on patterns and slow on people: reading individuals, building teams around what each person uniquely does well. The operating models I design tend to reflect this: small high-performing groups, embedded generalists supported by deep specialists, decisions reviewed against evidence and a shared quality bar.
Selected roles
Where I've led design.
2022-present
Kaluza
Product & UX Director · Energy
UX as one craft. Joined to lead research, service design, journey management, product and content design as one function. Rebuilt it from siloed specialists into embedded generalists backed by senior specialists, 27 people, and own its operating model, budget and headcount. Accountable for UX quality across Kaluza globally.
Then the platform beneath it. I expanded my remit to lead the Global Enterprise Experiences Platform team, owning the foundational enablement layer across all Kaluza products. My scope spans Nebula, our enterprise design system, global information architecture, accessibility, and the shared front-end application shell. Beyond setting product roadmaps and measurable quality gates, I help architect Kaluza's operating model, transitioning our organisational structure from traditional value streams toward global product builders paired with forward-deployed engineering.
And the agentic programme. I now also lead the agentic customer care solution for energy retailers. I built the first prototype and took it to joint executive forums. It became the foundation of the programme. Its first release, the move-out journey, is now live with OVO customers, one of the first agentic journeys in UK energy to resolve a customer's request end to end.
OutcomeThe shift to value streams drove a step change in delivery across Kaluza.10×+ initiative throughput~150% faster delivery~40% faster discovery~90% fewer reassignments
2021-2022
Shopify
Senior UX Manager · E-commerce
Joined on maternity cover and used it to grow the function: sharpened how it operated, with a focus on upskilling, and established the data visualisation practice. Led the team behind ShopifyQL Notebooks, a data storytelling tool for Shopify Plus merchants, since retired. Reported to the Global VP of UX, whose organisation ran to around 500 people.
OutcomeSpeculative design practice introducedNext-horizon vision for data analytics and visualisation
2015-2021
Brandwatch
VP, User Experience · Consumer intelligence
Joined as lead designer on Vizia, an enterprise marketing reporting and command-centre platform, and led a full redesign, with new data visualisation at its heart.
As Head of UX Research and Content Design, I built the company's first research and content design function. I set the roadmap, grounded the research in contextual inquiry and mixed methods, and set the content design strategy alongside it, with a researcher embedded in every product team.
Promoted to VP of User Experience to lead a multi-discipline practice, connecting the dots across business silos to strengthen the end-to-end customer experience. Introduced UX outcomes as a shared measure of progress toward a better experience for product teams and squads.
OutcomeCSAT increasedTime-to-task reducedTraining and support load lowered by in-platform onboardingFeature adoption increasedVizia redesign lifted sales35+ feature-request channels in one tool
2011-2014
Deutsche Bank & RBS
Lead UX Consultant · Financial services
Deutsche Bank, AutobahnFX. The FX trading platform within the Autobahn suite, an architecture-level redesign that went well past interface polish. I ran the platform-wide user research and used it to move programme leadership off the cosmetic refresh they had scoped to a full rebuild.
RBS, MicroRates. A digital hedging and risk-management platform for specialist lenders, also an architecture-level redesign.
OutcomeAutobahnFX:Leading electronic FX serviceTop-tier industry awardsMicroRates:Adoption up sharplyTask completion higherUser errors near zero
Latest work
Strategy, AI tools and agentic products.
Agentic Customer Care for energy retail.
Product & design strategy and build · Energy platform
Customer
Agent
Supervisor
An AI assistant inside Kaluza Customer Care that handles real energy-retailer conversations, by voice or chat, and tries to resolve them rather than pass them on.
My roleEnd-to-end experiencePlatformDecagon, headlessInterfaceReact 19 SPA
What it does. It takes both simple and complicated contact through a single conversation and hands over to a person when that is the right thing to do. The bar is real resolution, as good as or better than a person, judged on how it feels as well as how fast it is.
How it is built. I write the assistant's behaviour directly in Decagon, so the configuration is the spec and nothing is lost passing it to someone else. The interface is a single-page app I build in code, Vite, React 19 and Tailwind v4, with shadcn components on Base UI, themed to OVO's brand and connected to Decagon through a headless integration. I design and test it with my own hands rather than handing over a brief.
Guardrails and evals. It discloses that it's AI every time and stays within its scope. It hands over to a person when it matters, especially in vulnerable moments, and each of these behaviours is measured: disclosure, scope adherence, missed escalations and vulnerability detection. Anything safety-critical stays out of the AI's hands, and when a proposal tried to soften one of those rules into a guideline, I kept it enforced. The evals check knowledge and logic separately, so a failure points straight to its cause. That caught a content gap before it reached customers.
OutcomeKaluza Customer Care, the product this sits within, is recognised at the 2026 Amazon Connect Customer Innovation Awards and publicly reports:32% lower handling time23% higher first-contact resolution45% lower cost to serve50% more customers per agentThe agentic layer is targeting 70% lower cost to serve and 50% higher CSAT.
Release 1 · Move-out journey · Live in production
The foundations beneath every product in the suite.
Design system Nebula, from a single-product library to an enterprise source of truth.
Kaluza's design system, inherited as a single-product library and rebuilt into the enterprise source of truth, in code and in design.
ScopeSeveral product linesMy roleGovernance and standardsParityReact and Figma, one to one
One source of truth. I own the governance, the React and Figma parity, the accessibility standards and the contribution model that lets teams extend it without fragmenting it. Each new surface tightens the shared standard, with no forks.
Now, encoding judgment. The work is less about adding components than writing down which pattern applies, and when, clearly enough that a person or an AI agent gets it right without asking. As more of the interface is generated against the system, that is what keeps human and machine output coherent and on brand.
OutcomeThe components teams reach for most stay on the shared standard, with detachment under 1%.
Quality guardrails Accessibility, design-system adherence and front-end QA, enforced automatically.
Standards only hold if something enforces them, so my team builds the tooling that does, with AI.
ChecksAccessibility, adherence, front-end QAStandardWCAG 2.2 AABuilt withAI
Accessibility, checked automatically. Structured WCAG 2.2 checks run against live applications, Figma files and component libraries, pairing automated scanning with AI analysis, and turn failures into drafted remediation tickets and exportable conformance statements.
Adherence and QA. A front-end QA check tests each build against Nebula and flags drift. Both started local; the accessibility tool is piloting in my team, and the adherence check is moving into the shared testing environment so every team can run it as they build.
OutcomeA full accessibility audit complete across the suite, with a defined path to full compliance by the end of 2026.
Front-end architecture A global core with local variation: one shared shell over isolated data.
I turned a fragmenting, per-client front end into one shared core with local variation, and I lead the team that builds it.
ScopeMulti-market platformMy roleArchitecture, standards and deliveryModelShared shell over isolated data
Resolving for the unhappy path. Exceptions where data will not reconcile, and payments that cannot move the normal way, where the real question is who is accountable for putting it right.
The bar has to be explicit. Many teams building local variations only stay coherent when it is, so I set the standards the core is tested against, including tenant data isolation as a UX standard in its own right: one client's context never leaks into another's.
The team shippedExceptions routed by who owns the fixA self-serve exception catalogueOne refunds design, three payback pathsPrevention over handling
Kept coherent. A measurement spine holds the core together across clients. The architecture and standards are set and in use; the broader shift they enable is still landing, by design.
Tier 3 Client & market, varies per client
Tier 2 Product core, configured per client
Tier 1 Shared shell, one, global, every client
One shared shell over isolated per-client data.
Journey & research A journey-intelligence and research engine, its tools built with AI.
The platform's journey-intelligence and research capability: one live taxonomy, a behavioural-analytics framework, and research tools built with AI.
TaxonomyFour altitudes, liveAnalyticsFramework is mineResearchAI-moderated
One live taxonomy. I own the global journey taxonomy, a four-altitude model that ladders every user flow up to an agreed set of lifecycle stages, kept as a single source of truth. Where the map is incomplete, the gaps are surfaced in the open where the team can see them, which is what the top-down review closes. The behavioural-analytics framework runs across every delivered flow.
A05Lifecycle stageThe end-to-end customer lifecycle.
A125Macro journeyCross-functional journeys across the lifecycle.
A233Micro journeyEnd-to-end experiences that carry the metrics.
A3206User flowThe task-level flows delivered in the product.
Tools built with AI. A product-insights agent codes insights against the product map and the journey taxonomy, so decisions run on a searchable body of evidence. An event-audit tool takes a flow's master design file, extracts the actions it should emit, audits the events in the codebase and returns exactly what is missing, turning instrumentation into an automated fitness test.
Discovery at a fraction of the cost. The discovery runs on an AI-moderated research method I introduced. One programme ran 135 depth interviews in weeks, work that used to take months, at roughly a tenth to a fifteenth of conventional cost, with every transcript analysed.
OutcomeLive in production since launch, with a zero error rate.~90% faster time-to-insight>75% of scoped H2 specs on validated insight
On stage and on the side
Speaking on design, AI and being a woman in leadership.
I present product and UX widely, from town halls and internal events I host to external conferences and panels, representing the organisation across these topics.
Keynote · Future of Utilities, Amsterdam
Testing solutions with users in the energy transition.
Talk · Kaluza annual conference · Bristol Beacon
Scaling up with information architecture and an enterprise design system.
Speaker & moderator · Panels
Design, AI, female leadership and regulated industries.
Mentorship, coaching, building on the side.
Formal mentorship inside the organisations I've led, and structured work with women from underrepresented backgrounds through a dedicated programme. I also coach independently on the side, with a view to formalising this further over time.
kallı: private rehearsal for high-stakes speaking.
Product, content & interaction design · Pre-launch
I built this tool to accompany my mentorship, specifically for professionals navigating high-stakes conversations: defending a budget, challenging senior peers, or delivering key updates in a second language. It's designed for both halves of communication: structuring your points upfront and delivering with pace, clarity and authority.
Join the waiting list, or try kallı. For the password, email me.
ORIO: the hidden cost of gender norms.
Product, content & interaction design · Pilot
A series of short, interactive data experiments for young adults exploring the hidden economic and social costs of gender norms using sourced evidence. I'm currently exploring its next chapter, including a standalone platform and partnerships with schools and youth organisations.
Try ORIO. For the password, email me.
Once a builder, always a builder
Interactive installations · Exhibited
Between training as an architect and moving into digital design, I built responsive installations with Max/MSP and sensors, testing how far a physical interface could listen back. Two were exhibited: Logtime Twister (2005) at Design in Delft, with an honourable mention from Microsoft Research, and Oracle (2009) at the 2nd Thessaloniki Biennale of Contemporary Art and at enTechno, an underground electronic music festival in Athens.
Contact
Get in touch.
The epitome of cross-functional thinking. Others look to you for direction and clarity as an informal leader across Product, Engineering and Design.
A design strategy that clearly ties into the wider company strategy. The senior team came to see the function as not just relevant but critical to the company's future.
The headline on Evi is, she will make a difference. Her resilience is not to be underestimated in a business which has undergone a fair amount of M&A with all the integration and transformation work this inevitably involves.
Evi has an ability to persuade sceptical stakeholders, to communicate methodologies in such a way that their value becomes clear, to instil confidence in the people around her, whether she's working in the design studio or the boardroom.