AI Drafts the RFP Now: 65% of Meeting Planners Use Generative AI, and DMCs Are Often the Last to Know

The short answer

Generative AI has moved to the front of the MICE procurement funnel, ahead of the RFP — and most destination management companies cannot see it happening. The Northstar Meetings Group/Cvent Meetings Industry PULSE Survey of more than 1,000 corporate, association and agency planners found 65% of planner organizations already using generative AI platforms, with another 13% saying they will be soon — yet only 16% report significant improvements to planning outcomes so far (Northstar/Cvent PULSE, December 2025 wave, published 2026). In Asia-Pacific, planner adoption jumped from 43% to 75% in a single year (Gevme, 2026), and the EMEA Q2 2026 PULSE wave reports adoption still accelerating, with planners citing measurable gains in efficiency, content creation and event planning support (Northstar/Cvent PULSE EMEA, Q2 2026).

Translation for a DMC or incoming operator: by the time a planner emails you an RFP, an AI assistant has most likely already researched the destination, drafted the program outline, suggested venue types and day counts, and — critically — assembled a shortlist of who to send it to. If your company’s destination expertise is not machine-readable, you were not on that list, and you never knew the opportunity existed. The leisure side of travel already crossed this threshold: 58% of Americans now say they use AI tools to plan summer trips, versus 55% who use traditional travel websites (Qlik summer travel survey, July 1, 2026). Group travel procurement is following the same path, with higher contract values attached.

This is not a story about robots replacing destination expertise. It is a story about where the shortlist is written — and it is increasingly written by a model, from sources a DMC does not control. The fix is the same discipline this blog applies to hotel AI visibility: publish your destination expertise in a form AI planners can parse, cite and recommend.

The scale of the shift

The business events industry is simultaneously growing and automating its research layer. The US MICE market was valued at USD 172 billion in 2025 and is projected to reach USD 186 billion in 2026 (Straits Research, 2026), while industry analysis projects the global MICE opportunity at roughly $103 billion by 2035 (Converve, 2026). Singapore alone logged S$2.3 billion in MICE tourism receipts in 2025, up 35% year over year (Gevme, 2026). Destination demand is not the problem.

The supply side is crowded. In Grand View Research’s MICE market survey, 27% of respondents were classified as DMCs — a quarter of the surveyed ecosystem competing on essentially the same pitch: local knowledge, venue access, operational reliability (Grand View Research, 2026). When planners research “DMC for 300-person incentive in Lisbon” or “destination management company Portugal corporate offsite,” generative AI now mediates what they find. A model that has never crawled your capabilities page, parsed your venue database or read your case studies will simply describe the market generically — or name the competitors whose content it could read.

The numbers that matter

The 16% figure deserves attention. Most planners are using AI for first-draft work — destination research, agendas, RFP criteria — not for final execution. That is precisely the stage where DMCs are selected out before they can compete on service, relationships or on-the-ground capability. You are not losing pitches; you are losing the chance to be pitched to.

How the machine-drafted shortlist actually forms

Ask ChatGPT, Gemini or Claude a realistic question: “We’re planning a 3-day leadership offsite for 60 people in Portugal in May, suggest the destination and 3 DMCs who could run it.” What comes back is constructed from what the model can access: association listings, trade-press mentions, published case studies, review platforms, and DMC websites structured well enough to parse. The output names the companies whose expertise is legible to machines — not necessarily the best ones.

Four failure modes quietly erase DMCs from that list:

1. Invisible capabilities. Services described in prose (“tailor-made programs, MICE and incentives”) with no numbers — group sizes handled, venues contracted, languages staffed, programs delivered per year — give a model nothing specific to cite.
2. Case studies behind PDFs. A beautiful 40-page proposal PDF is opaque to crawlers. The same program summarized as an HTML page — destination, group size, duration, objective, outcome — becomes citable evidence.
3. No destination-depth pages. A single “Destinations” dropdown listing 14 countries tells a model nothing. A page per destination — with named venues and capacities, seasonality, typical program structures and indicative budgets — is what an AI planner paraphrases into its recommendation.
4. Unstructured RFP responses. The 65% figure cuts both ways: planners increasingly use AI to summarize and compare the RFP responses they receive. Proposals written as narrative brochures get flattened; responses with clear structure — capacity tables, itemized staffing, explicit exclusions — survive machine comparison intact.

The leisure-side parallel is exact. When ChatGPT began linking hotels directly in recommendations in 2026, properties with structured, specific content captured the citations while generic brochure sites did not — the dynamic we documented in how AI planners pick the hotels they recommend. Group travel runs the same logic with five- and six-figure contracts.

What DMCs should publish now

The goal is not volume; it is recommendation readiness — the same standard our travel AI audit applies across a travel company’s footprint. For a DMC, prioritize:

1. One flagship page per destination covering named venues with capacities, transfer times, seasonality notes and program formats you actually operate. Structured data (Organization/TravelAgency, plus FAQ schema) makes the facts unambiguous.
2. Case studies with machine-readable specifics. Group size, nationality mix, duration, objective, one measurable outcome. Five good HTML case studies beat fifty PDFs.
3. A capabilities page with numbers. Programs per year, group-size range, staff languages, on-ground team size, response-time SLA. Numbers are what an AI comparing four DMCs can quote.
4. An AI-answerable FAQ. “What does a DMC cost?”, “DMC vs local event agency”, “How far in advance to contract a DMC for 200+ guests” — the questions planners actually paste into assistants, answered directly on your domain.
5. Consistent identity across the ecosystem. Trade-association directories, review platforms and your own site must state the same facts (founded, offices, destinations, specialty). Models reconcile sources; contradictions suppress confidence.

If you operate across many destinations and cannot prioritize, the sequence is: audit first (find out where you actually stand in AI answers today — a travel AI audit does this against real planner prompts), then build flagship pages for your two or three highest-value destinations, then case studies. An agency running this as a program should benchmark prompt-by-prompt, quarter over quarter — the methodology behind our travel GEO agency work, with pricing scaled to portfolio size.

The window is open — briefly

The PULSE data contains a strategic gift: only 16% of planner organizations see significant results from AI so far. Adoption is high, mastery is low, and the shortlisting behavior is already forming. That combination — heavy use, shallow sophistication — is the window where publishing better destination content yields outsized AI-visibility returns. When most planners’ AI workflows mature (the EMEA Q2 2026 wave suggests that is underway), the shortlists will ossify around whichever sources the models learned to trust during this period. DMCs that publish structured destination expertise in 2026 are teaching the shortlisting layer their name. DMCs that wait are teaching it their competitors'.

Corporate buyers are moving the same direction, as we documented in the corporate travel AI gap: 83% of buyers want conversational booking and 95% are comfortable with AI recommending negotiated-rate hotels (GBTA, May 2026). The group, incentive and meetings funnel sits directly in that path.

FAQ

Do AI planners actually recommend DMCs by name? Yes — when the model has citable material. Ask any assistant for DMC recommendations in a specific destination and it names companies drawn from readable sources: structured directories, trade press, case studies, well-built websites. The companies named are those whose expertise was machine-legible, not necessarily the largest or best.

Is this a problem only for big DMCs? No — smaller, single-destination DMCs are arguably better positioned, because deep, specific content about one destination is exactly what models reward. A boutique Lisbon DMC with ten detailed HTML case studies can outrank a 14-country operator whose site is a destinations dropdown.

What does GEO mean for a DMC specifically? Generative engine optimization for DMCs means making destination expertise parseable and citable: structured capability data, per-destination flagship pages, case studies with numbers, FAQ schema, and consistent facts across every listing an AI might read. It is the group-travel equivalent of hotel AI visibility work.

How do we measure whether this is working? Run a fixed set of realistic planner prompts monthly (“DMC for a 150-person incentive in [destination]”, “[destination] DMC for pharmaceutical offsite”) and track whether your company appears, how it is described, and which page the model draws from. That prompt-tracking discipline — not impressions — is the KPI. Our travel AI audit sets exactly this baseline.

How quickly does published content affect AI recommendations? Faster than search rankings, but not instantly. Models and their retrieval layers pick up well-structured pages within weeks; consistency across your site, directories and trade press is what converts a mention into a default recommendation. Plan on a quarter of sustained publishing before judging results.

The bottom line

The RFP still arrives in your inbox — but the shortlist was written earlier, by a model, from content you may not have known you were competing with. With 65% of planner organizations already running generative AI (Northstar/Cvent PULSE, December 2025 wave) and APAC adoption at 75% (Gevme, 2026), DMCs compete for AI citations before they compete for contracts. Publish the destination expertise you already have in a form machines can quote, measure your presence in real planner prompts, and the AI-drafted RFP starts arriving with your name already on the list it was sent to.

Palmtree.ai helps hotels, tour operators and DMCs win AI recommendations — from prompt-level audits to full travel GEO agency programs. Start with a travel AI audit or see pricing.