What We Built, Why We Built It, and Where We're Going

We didn't start Neon & Slate because AI was trending. We started it because of something we'd been watching for years — from the inside.

After 14 years deploying technology in complex organisations — banks, corporates, mid-market companies across the US and Europe — we'd seen the same pattern repeat itself with every major technology wave. The curve of what was possible accelerated faster than organisations could absorb it. And the gap between the two kept getting wider.

AI made that gap structural.

Charlotte has been working on AI since 2009. Clément since 2012. In 2017, Charlotte built an AI-enabled fashion tech company and implemented AI systems for BNP Paribas. In 2018, Clément launched a SaaS platform applying AI to legal and paralegal workflows. At the time, most boards were still focused on dashboards and analytics. AI was a whisper in enterprise contexts — too futuristic, too far-fetched for most decision-makers to take seriously.

We took it seriously. Because we could see the curve steepening.

Neon & Slate was built to close that gap.

---

What we got right early

We made three bets before the market caught up. All three have compounded.

Bet 1: Builders over advisors.

The AI market is flooded with strategy decks. Roadmaps. Maturity assessments. Frameworks with impressive names and no production code behind them.

We went the other way. We built a network of 30 forward-deployed engineers — from UC Berkeley, MIT, Centrale, and Ecole Polytechnique — who ship systems that work in production. Not prototypes. Not demos. Working systems, inside messy enterprise environments, with real data, real constraints, and real users.

The reality is: AI only creates value when it runs. Everything before that is preparation.

Bet 2: The mid-market is where the value is.

Most AI firms chase enterprise logos or VC-backed startups. We focused on the middle: companies between $50M and $500M in revenue, often PE-backed, facing the same transformation pressure as Fortune 500 companies but with a fraction of the resources.

These companies are too complex for pre-built AI solutions. Too resource-constrained for custom development from scratch. And facing a critical AI talent shortage that makes internal hiring almost impossible.

This is the messy middle of AI adoption. And it's where the biggest ROI opportunities sit — because the baseline is so low and the operational leverage is so high.

Bet 3: Non-technical leaders are the real unlock.

Here's what no one talks about: the bottleneck to AI adoption is not technology. It's not budget. It's not even data quality — though that's usually a mess too.

The bottleneck is that the people who understand the business processes — the COOs, the heads of operations, the finance directors — don't know how to build with AI. And the people who know how to build with AI don't understand the business.

Our thesis from day one: the companies that will see real AI ROI are the ones where non-technical leaders become builders. Not coders — builders. People who understand their workflows deeply enough to redesign them with AI, given the right tools and the right guidance.

We designed everything around making that happen. Strategy sprints that end with working systems, not slide decks. 90-day implementations with executive-level ownership. Hands-on training where leaders build their own AI tools — and actually use them.

---

What we got wrong

We underestimated how long it would take to make companies care. In 2019, 2020, 2021 — AI was a tough sell to mid-market CEOs. The technology was real, but the urgency wasn't. We spent years educating a market that wasn't ready to listen.

We also overestimated how quickly the tooling would mature. The early years were painful. The infrastructure to deploy AI reliably in enterprise environments simply wasn't there. We built a lot of custom scaffolding that the industry has since productised.

And then ChatGPT launched. Suddenly, every company cared about AI. The urgency was there. The market caught up. And we were ready — because we'd spent years building the operational muscle that others were just starting to think about.

---

What we see now

The uncomfortable truth: 95% of generative AI pilots never make it past the experimental phase. 56% of CEOs report getting "nothing" from their AI investments. The market has moved from "we don't care about AI" to "we've spent money on AI and have nothing to show for it."

Most organisations fall into one of two traps:

→ Top-down enterprise rollouts — massive platform deals, months of integration, low adoption. The classic "we bought Copilot for 5,000 people and 200 use it."

→ Bottom-up experimentation — individual teams playing with ChatGPT, building clever hacks, generating impressive adoption numbers but zero measurable business outcomes.

What actually works is a guided middle path. Strategic direction from leadership. Hands-on empowerment of the people who run the business — in finance, operations, sales, customer service. Enterprise-level governance with individual-level building capability.

That intersection is where real ROI lives. It's where we operate.

---

Why we built a GEO practice

Something became clear as we worked deeper with Fortune 500 clients in banking, healthcare, and luxury: AI wasn't just changing how companies operate internally. It was changing how companies are perceived externally.

ChatGPT, Gemini, Perplexity, Claude — these systems are now the first touchpoint for millions of consumers, investors, and journalists. When someone asks an LLM about a brand, a financial product, or a medical treatment, the answer they get isn't pulled from a press release or a corporate website. It's synthesised from whatever the model considers authoritative: Reddit threads, outdated news articles, competitor content, forum posts from five years ago.

For industries where reputation is the product — luxury, banking, healthcare — this is an existential problem. And almost no one was measuring it.

We had the right profile to address it. Charlotte spent years inside BNP Paribas where reputation monitoring was the single biggest concern. Our work with luxury brands and financial institutions had given us deep relationships and a front-row seat to the pain: brands discovering that AI was surfacing outdated narratives, competitors being positioned more favourably in AI responses, and entire categories being reshaped by content that no PR team had ever seen or approved.

Traditional social monitoring and media tracking don't cover this. LLMs don't work like search engines. They cite different sources, behave differently model by model, and respond to different content signals. Measuring what AI says about your brand requires a fundamentally different approach — structured prompts across multiple models, per-model sentiment analysis, source chain mapping to understand which content is actually driving AI responses.

So we built a GEO practice — Generative Engine Optimisation — designed specifically for industries where perception, trust, and accuracy are non-negotiable. We deploy thousands of structured prompts across five major AI platforms, measure sentiment and visibility by model and by market, and work with industry coalitions to shift the narrative at the source level.

It's not SEO adapted for AI. It's a new discipline built from the ground up — and it sits naturally alongside our implementation work. We help companies build AI into their operations, and we help them control how AI represents them to the world.

---

Where we're going

AI is not another software wave. It's an operational transformation — and it's just getting started.

Over the next 18 months, we see three shifts that will reshape how companies work:

Agentic AI moves from demo to production. Autonomous workflows that run end-to-end — not chatbots, but systems that execute, verify, and iterate. The companies that figure out how to deploy these safely inside regulated environments will pull ahead fast.

The CAIO role becomes essential. Most companies don't have the internal expertise to navigate AI strategy, governance, and implementation simultaneously. The fractional CAIO model — senior AI operators embedded part-time in leadership teams — will become standard for mid-market companies.

AI reputation becomes a board-level concern. Every brand, institution, and regulated entity will need to monitor and manage how AI systems represent them. GEO will move from niche to essential — not just for luxury and banking, but for any organisation where trust, accuracy, and perception drive revenue.

We're building Neon & Slate for this moment. Two practices, one thesis: AI is reshaping both how companies work and how they're seen. Our implementation arm builds systems that hold up in production. Our GEO practice ensures that AI represents our clients accurately and favourably. Both are built around the same principles — specialist expertise, measurable outcomes, and a refusal to settle for strategy decks without production results.

We don't expect the market to stabilise. We expect it to accelerate. And we've built a team and a model that moves with it.