Take Control of M&E: Why being AI Smart beats being AI First

Fast forward five years, and in January 2025 we began working on a new framework around what it means to be AI Smart as opposed to AI First. Two things pushed us there.

The first was the Salesforce advert at the 2025 Super Bowl, which saw Matthew McConaughey dining in a downpour, because his AI agent didn’t have enough context to properly book. It captured how throwing AI at a problem without purpose or meaning is worse than insufficient; it's absurd. It was obvious to us that applying AI First thinking to meetings and events, without the domain intelligence to deploy it meaningfully, wasn't going to add value for our customers. Speed without expertise just delivers faster mistakes.

The second was that, against this backdrop of transformation, the meetings and events industry had a big problem hiding in plain sight. 91% of travel managers are now also responsible for M&E. An entire category of spend has been added to their remit, with no additional resource, no specialist technology and no reliable data to work from. The challenges this creates are consistent and painful. For small and medium events, which represent the vast majority of M&E activity, the major legacy players who hold dominant market share have deployed almost no meaningful technology. They still operate through predominantly human-led, manual processes. Shadow spend is everywhere, and when a budget is scattered across dozens of unconnected tools, direct venue relationships and offline negotiations, it’s invisible. It’s important to note that venue supply in meetings and events is fundamentally different from the supply travel managers are used to. Travel sits with 50 to 100 large airline and hotel groups, manageable and data-rich; M&E venue supply is tens of thousands of independently owned bars, restaurants, private dining spaces and experience providers, disaggregated and impossible to negotiate with at scale without the right intelligence layer. On top of all this, corporate buyers are no longer willing to accept a simple  hotel conference room. They want distinctive spaces that reflect their brand and give people something worth remembering.

The old model has quietly broken. Sourcing is still manual and approvals still sequential in a world where speed is now competitive currency. There is no intelligence layer sitting behind the process, so nothing learns from the last decision. Spend is fragmented, reporting is retrospective, and procurement is blind in real time. Growth requires proportionally more headcount because there is no flywheel and every event starts from scratch. And when employees source venues and sign contracts independently, each interaction becomes a compliance and legal exposure, with sensitive event and attendee data passing through uncontrolled channels and no audit trail. The model is simply too slow, too expensive and too human-dependent for the AI era.

Why being AI Smart in the AI Era matters

To meet this challenge, we framed our strategy around what it would take to be a truly AI Smart platform: one that combines AI infrastructure with ten years of domain knowledge and proprietary data, so that we deploy AI solutions intelligently instead of indiscriminately. At the core of this is what we call our system of intelligence.

Here's what most organisations misunderstand about AI competitive advantage: the model is not the differentiator. GPT, Claude, Gemini are utilities, like electricity. Every competitor has access to the same power source. What cannot be replicated is the data you plug into it. Netflix's recommendation engine isn't powerful because of its model; every streamer uses similar models. It is powerful because of decades of irreplaceable behavioural data. HeadBox has spent over a decade building exactly this kind of asset: 100,000+ venues, 50,000+ events, tens of thousands of enquiries, and a decade of corporate buyer behaviour across six continents, all structured and queryable by AI.

This is the shift from system of record to system of intelligence. The system of record for meetings and events captures briefs, bookings and invoices. It's useful, but it forgets. Our system of intelligence remembers, because supply and demand live together on the same platform, and every brief, sourcing decision and completed event adds to a memory that compounds. Platforms that keep supply and demand apart have no memory. With our latest release of HeadBox for Business, we have a platform that never forgets.

That system of intelligence rests on three core data assets: the company and its meetings and events programme, so we understand each organisation at a structural level and what matters to them; the booker and their key data points, built from years of booking history, knowledge and preferences; and the full picture of venues, hotels, suppliers and accommodation. It is the interaction of supply and demand across all three that creates uniquely rich, compounding data, a flywheel that drives smarter outcomes for every next brief and every next booking.

Human, AI, and Risk Synthesis

Our system of intelligence for meetings and events only matters if it enables the right calibration across three pillars: how much to dial the human role up or down; how far to push AI automation, from assistive, to copilot, to fully agentic; and how deliberately we manage risk as automation increases. HeadBox for Business is built around getting this synthesis right.

Why does this matter so much to our customers? Because there's a tension sitting at the heart of the AI era that cuts against the instinct to automate everything. As AI becomes ubiquitous, trust is becoming the scarcest commodity of all. Deepfakes have enabled multi-million-dollar wire fraud through faked video calls. Political leaders now dismiss real evidence simply by claiming it's synthetic, what researchers call the Liar's Dividend. The result is a genuine cultural counter-reaction: people spending more time in face-to-face interaction, brands discovering "authenticity fatigue" as audiences reject content that feels too smooth and too perfect.

In a world where you cannot fully trust a video call, a voice, or a written communication, a room full of real people is the one thing that cannot be generated. A shared meal, a conference, a team offsite become epistemically safe spaces in a world of synthetic uncertainty. For meetings and events, this is a tailwind of historic proportions. And precisely because it is, more automation demands more deliberate control, not less. The more of the workflow AI handles, the more disciplined we have to be about where human judgement, governance and accountability sit.

Three HeadBox Deployment Areas Bringing ‘Being AI Smart to Life’

1. Service as software, rather than service replaced by software.

In our HeadBox for Business platform, we are scaling our human service so it is faster, smarter, and higher quality. AI is embedded across the whole workflow, from qualify, source and propose through to negotiate, contract and close, while human expertise stays in charge of the judgement calls. We have also built the agentic layer,  an agent that reads the brief, invites venues, chases responses, shapes proposals and flags watchouts, with the team supervising every step. Alongside it, we're running Agent Shadow, an agent that makes its own call on every decision the team faces, learns from what the team actually chooses, and earns the right to act alone one decision at a time. This is not about removing the concierge. It's about making the concierge faster and freer to focus on the parts of the job that need a human.

2. Scaling supply: launching a new city in hours rather than months.

An agent discovers venues in a new market, scores them against our ideal customer profile, researches the ones worth pursuing in real depth, writes the listing itself, and manages the venue through claiming. A human is able to step in at any point, whether that means choosing tasteful photos or simply approving listings before they go live. A further ten US cities are already staged and going live now, and we are well on our way to adding another 100 cities to our existing global footprint in the next 12 months. This is what an AI Smart supply engine looks like in practice: a compounding, self-improving pipeline that a human still controls, in place of a team manually onboarding venues one at a time.

3. Personalisation and curation: elevating the role of the event planner.

Booker personalities are built from years of M&E team notes and real booking behaviour, summarising likes, dislikes and patterns for whoever picks up the next brief. Bookers can ask AI anything about a venue and get answers grounded in deep, personalised research, with their preferences and collections carrying across every future brief. The effect, for our mandated corporate bookers, is that the administrative and operational burden lifts. They get to spend their time on what they're actually good at: relationship building, creativity and curation, instead of chasing venues and reconciling spreadsheets.

Conclusion

Put these three together and you can see what an AI Smart platform actually looks like. It is not AI bolted onto an old process, and it is not a bet that AI can replace the people who make events work. It is a decade of proprietary domain knowledge, compounding with every transaction, deployed through AI infrastructure that scales human service, scales supply, and puts the booker in a more valuable role. Human judgement is dialled up precisely where creativity, risk and trust require it. That combination is what a legacy incumbent cannot retrofit and what an AI-native entrant cannot buy.

It's also the only version of this future worth building towards: one where AI does the heavy lifting on speed, scale and the intelligence layer, from sourcing and compliance to data, procurement and reporting, while  humans bring the creativity, relationship building and judgement to design an experience that moves people. The role shifts from administrator to lead architect.

This is  how the AI era itself has to be built, with humans in the lead and AI in tandem. Humans belong not just in the concierge team but in the boardroom, and, above all, at the level of the governments and institutions that will decide how this technology is allowed to reshape our economies, our democracies and our security. If we cannot get that calibration right on the smaller stage, we have little chance of getting it right on the biggest one.