Reflections from my conversation with Mohamed Soliman, Regional VP at Energy Exemplar, on The CX Algorithm Podcast.
I went into this conversation expecting to talk about energy. I left rethinking how I run customer experience.
Mohamed Soliman builds optimisation models for power systems — networks with thousands of nodes, dozens of competing objectives, and consequences measured in billions of dollars and millions of lives. The kind of complexity that should require a whiteboard, three engineers, and a very strong coffee to explain. He explained it to me in one sentence.
"Optimisation is finding the best path between two points in a network."
Two points become three. Three become a thousand. Add direction, cost, time, constraints, and a few competing objectives, and you are running the same problem an energy operator runs every minute of every day — except in CX, we tend to call it "Tuesday's roadmap meeting."
Here is what I took away from the conversation, and what I am taking back to my own work.
Mohamed's claim sounded sweeping at first — that the supply-and-demand problem in an energy grid is the same problem we are solving in FMCG, in logistics, in water, and in customer experience. I pushed back on it during the recording. The more I sat with it afterwards, the more I realised he was right.
In FMCG, your supply is your factories, your demand is your consumers, your channels are the constraints, and freshness is a hard variable that breaks the model if you ignore it. In logistics, you are routing nodes under capacity limits. In CX, your "supply" is the experience your organisation can actually deliver at any given moment — across people, process, technology, and data — and your "demand" is what your customer expects, contextualised by their channel, their lifetime value, and whether their flight just got cancelled.
The maths underneath is the same.
A 2025 systematic review of multi-objective optimisation across smart cities makes the point concretely. The same class of algorithms that improved post-earthquake evacuation efficiency in Tokyo by 20% is what Beijing used to reduce its urban heat island effect, and what Zurich used to balance density with green space. Three different cities, three different problems, one algorithm.
For CX practitioners, the implication is practical: stop treating your problem as fundamentally unique. The journey-mapping work you are doing has structural cousins in every industry that has been forced to model supply, demand, and constraints together. Borrow from them.
Picture your last week. You tried to maximise satisfaction while minimising cost-to-serve, while protecting agent wellbeing, while hitting SLAs, while respecting compliance, while not breaking the brand promise.
That is a multi-objective optimisation problem with competing constraints. We just call it "Monday."
We treat it as a judgment call because we lack the language to treat it as a model. nAG's research on multi-objective optimisation describes what we are losing when we do that — what they call the Pareto front. When you have competing goals, there is no single optimal answer. There is a curve of best-fit compromises, and the leader's job is to choose, deliberately, where on that curve you want to live.
What this looks like in practice is uncomfortable but useful: write down the four or five objectives your CX program is actually balancing — not the ones in the strategy deck, the real ones — and then write down the priority weighting between them. If you cannot do this on a single page, you do not have a strategy. You have a wishlist that gets re-prioritised every Monday by whoever spoke loudest in last week's exec meeting. Most teams I've worked with discover this the moment they try.
The part of the conversation I keep returning to is Mohamed's foundation layers of trust. Before you trust any number a model gives you — optimisation, AI, predictive, the dashboard your CFO reads on Monday morning — three layers need to be verified, in order.
First, the data. Garbage in, garbage out. Accuracy, freshness, trustworthiness. Most CX programs skip past this entirely and start arguing about which dashboard view to put on the boardroom screen.
Second, the methodology. What is the actual objective the model is solving for? What constraints did you bake in? Was the model designed for the question you are asking, or is it answering a different question and hoping you won't notice?
Third, the governance. Especially with AI in the mix — what are the guardrails? Who is allowed to challenge the output? What happens when the system says something that feels wrong?
This isn't theoretical for CX teams right now. CXToday published a piece in late 2025 with a headline I think every CX leader should pin to the top of their inbox — AI Hallucinations Start With Dirty Data — citing Gartner's view that automation without unified data pipelines is one of the leading causes of AI failure. Forrester's 2025 enterprise research puts the per-employee cost of cleaning up after AI hallucinations at roughly $14,200 per year. An MIT study found that language models use 34% more confident language when they are wrong than when they are right. The more wrong the AI is, the more certain it sounds.
For practitioners, this is where the AI conversation gets serious. If you are deploying AI into customer-facing journeys without a clear answer for which layer of the trust stack you have actually verified, you are not deploying an intelligent system. You are deploying a confident one. There is a meaningful difference, and your customers will find it before you do.
Here is the moment in our conversation where I stopped taking notes and just listened.
Mohamed described how the energy industry has been forced to redefine its customer. For a hundred years, the model was simple: producers generate, the grid distributes, and consumers consume. Energy flowed in one direction. Money flowed in the other. Then rooftop solar happened. Then home batteries. Then EVs that can sell power back into the grid at peak hours and earn money while their owners sleep. The "consumer" became a prosumer — someone who produces and consumes simultaneously. A single house on the network is now both a load and a supplier, depending on whether the sun is out.
Schneider Electric calls this "Grid to Prosumer." Salesforce describes it as a partnership replacing a transaction. And Power Magazine reported that California's main grid hub experienced over 1,100 hours of negative electricity prices in 2024, more than double the year before — not because the technology changed, but because the customer did.
CX is in the same moment. Your customers are no longer just on the receiving end of experience. They write the reviews that move purchase decisions before your marketing budget reaches the room. They generate the content that shapes your brand, whether you sanctioned it or not. They co-design products through community feedback, often more usefully than your innovation team. They are nodes in a two-way network — and increasingly, they are suppliers of value to your business.
McKinsey's 2025 research on customer-centric growth makes the implication hard to ignore — companies that put CX at the core grow revenue at twice the rate of their less customer-focused peers. The example that stuck with me is Starbucks: over 300 customer-driven ideas implemented through their My Starbucks Idea platform, and revenue grew to $26.5 billion in the process. That isn't a feedback program. That is a prosumer model dressed in a coffee apron.
Practically, this changes the questions we ask in CX design. Instead of "how do we deliver this experience to the customer," we should be asking "what does the customer produce, and how do we make space for it in the journey?" Reviews, advocacy, content, referrals, ideas, complaints that surface real product flaws — these are not by-products of CX. They are inputs into the next version of CX. Treat them that way and your experience economics starts looking very different.
Mohamed said something during the recording that I have quoted three times this week already: "Choosing between resilience and flexibility is, on its own, an optimisation problem."
A purely efficient grid would only generate when prices are highest. A purely resilient grid would overbuild redundancy at every node. Neither works. The optimum lives in the messy middle, and in energy, what holds the balance is regulation.
CX has the same pattern, and we almost never name it. Pure efficiency means automating everything, driving cost-to-serve to the floor, targeting 30-second handle times and a chatbot for every interaction. Pure resilience means human escalation paths everywhere, slack capacity at every level, policy exceptions for anyone who asks. Both extremes break, and both produce predictable forms of customer pain. The optimum lives in between, and in CX, what holds the balance is the brand promise.
For practitioners, this means making the brand promise the constraint, not the aspiration. When you write a journey design, the question is not "can we afford this experience" — it is "does this experience honour what we said the brand stands for?" If the answer is no, no efficiency saving is worth it. If the answer is yes, you have permission to invest. Constraint-led design is harder than aspiration-led design, but it is the only kind that holds up under cost pressure.
If I had to compress everything Mohamed said into one line, I can act on this quarter, it is the one he closed with: start from the vision, not the data.
Your organisation's vision should be the objective function of your CX model. Vision flows down to KPIs. KPIs flow down to inputs and constraints. Inputs flow down to metrics. Top-down, always — because the moment you start from the data, you optimise for what is measurable instead of what matters, and what is measurable is rarely what matters.
This is exactly what McKinsey's CX Building Blocks framework puts at the foundation of CX transformation: a clearly defined aspiration as the first of three pillars. It is also what Forrester singled out in McKinsey's CX Strategy Wave recognition — the ability to define the North Star, and then refuse to drift from it.
If you are looking for one thing to do this week as a result of all this, here it is. Sit down with your most recent CX strategy document, and ask three questions of it. What is the single objective function we are optimising for? What are the three to five constraints we have explicitly accepted? And what is the trade-off curve we are choosing to live on?
If you cannot answer all three on one page, the work isn't to do more CX. The work is to define the problem properly before the next quarter of effort goes into solving the wrong one.
Mohamed put it cleaner than I can:
"If your vision is to provide the best service for the customer, then it moves down to the KPIs of that service. Those become your optimisation objectives. Otherwise, there will always be a disconnect."
That disconnect is, I now believe, the real reason most CX programs underperform. Not bad data, not poor tools, not the wrong vendors — just an objective function that was never properly defined, optimising the wrong thing very efficiently.
Everything is an optimisation problem. Including your CX strategy. The only question left is whether you are running yours deliberately or by accident.
🎧 Listen to the Full Conversation
Listen to the full conversation with Mohamed Soliman on The CX Algorithm Podcast — episode title: "Everything Is an Optimisation Problem — Including Your CX Strategy."
https://open.spotify.com/episode/0VtRXkt8XE1emGCgqHAkb4?si=b76d9b20bb2d47fa
Mohamed Soliman is the Regional Vice President at Energy Exemplar, one of the world's leading platforms for energy market modelling and optimisation. Over the past two decades, Mohamed has operated at the intersection of strategic vision and commercial reality — scaling Energy Exemplar's MENAT and CIS operations, contributing to its A$1.6B acquisition, and driving a 12x revenue transformation in Siemens' steel sector. His foundation is rigorous: a B.Sc. in Electrical and Digital Control Engineering from Cairo University, an MBA, a DBA, and a Private Equity & VC certification from Wharton.
What makes Mohamed the right voice for this episode is his ability to take some of the most mathematically complex problems in industry — multi-objective optimisation across thousand-node energy networks — and explain them in language any leader can act on. He has led the GCCIA Interconnection Program underpinning regional grid stability across the Gulf, advised sovereign funds, regulators, and policymakers across 28+ countries, and built a reputation as "the quiet force" who unlocks deals where strategic insight has to meet commercial realism.
Dr. Eslam Afifi is a customer experience and data leader whose work sits at the intersection of human behaviour, technology, and decision-making. His career spans multiple industries and regions, where he has helped organisations navigate complexity, design better experiences, and make sense of the growing role of data and AI in everyday decisions.
Curious by nature and sceptical of one-size-fits-all thinking, Eslam created The CX Algorithm to explore how customer-centricity actually works in practice. Through conversations with leaders from different contexts, the podcast unpacks the choices, trade-offs, and design decisions that quietly shape customer experience over time.
Follow The CX Algorithm for thoughtful conversations on experience, intelligence, and leadership — and a reminder that great CX is designed, not declared.
The article draws on the following sources, in the order they appear:
1. Multi-Objective Optimization for Smart Cities: A Systematic Review of Algorithms, Challenges, and Future Directions (2025). PubMed Central / National Center for Biotechnology Information. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12453809/
2. nAG (Numerical Algorithms Group). Mastering Trade-Offs: Balancing Competing Objectives in Multi-Objective Optimization. https://nag.com/insights/balancing-competing-objectives-in-multi-objective-optimization/
3. CXToday (November 2025). AI Hallucinations Start With Dirty Data: Governing Knowledge for RAG Agents. https://www.cxtoday.com/customer-analytics-intelligence/ai-hallucinations-start-with-dirty-data-governing-knowledge-for-rag-agents/
4. Spivack, N. (2025). The Hidden Cost Crisis: Economic Impact of AI Content Reliability Issues — citing Forrester Research (2025), The Total Economic Impact of AI Hallucination Mitigation in Enterprise Settings. https://www.novaspivack.com/technology/the-hidden-cost-crisis
5. Four Dots (2026). Business Impact of AI Hallucinations — Rates & Ranks — citing MIT Research (January 2025), AI Confidence and Hallucination Correlation Study. https://fourdots.com/business-impact-of-ai-hallucinations-rates-and-ranks
6. Schneider Electric. Grid to Prosumer: Distributed Energy Resource Management. https://www.se.com/us/en/work/campaign/grid-to-prosumer/
7. Salesforce. New Opportunities for Utilities With Energy Prosumer. https://www.salesforce.com/blog/energy-prosumer/
8. Power Magazine (October 2025). Advanced Grid Solutions — Harnessing the Power of Prosumers. https://www.powermag.com/blog/advanced-grid-solutions-harnessing-the-power-of-prosumers/
9. CX Network (March 2025). Customer-Centric Companies Achieve 2x Revenue Growth, Says McKinsey. https://www.cxnetwork.com/cx-experience/news/customer-centric-companies-achieve-2x-revenue-growth-says-mckinsey
10. CustomerGauge. 6 Customer Experience Management Frameworks for 2025 — including McKinsey's CX Building Blocks. https://customergauge.com/blog/customer-experience-management-framework
11. McKinsey & Company (April 2025). An Inside Look at Why McKinsey Is a Leader in Forrester's Customer Experience Strategy Consulting Evaluation. https://www.mckinsey.com/about-us/new-at-mckinsey-blog/an-inside-look-at-why-mckinsey-is-a-leader-in-forresters-customer-experience-strategy-consulting-evaluation
There is a phrase I hear in boardrooms constantly. Every time I hear it, I tell myself I should reward myself with a treat. That phrase is "channel strategy." Not…
What does it take to survive — and thrive — through the internet boom, the booking.com wave, and now the rise of agentic AI? I recently sat down with Kate Couling…
Customer experience is ultimately delivered through thousands of small human decisions. A frontline employee deciding how far to go for a customer. A manager…
