How an AI-native agency turned ₹9 lakh of ads into ₹3.59 crore in safari bookings

Nature Safari India (NSI) runs conservation-led wildlife safaris across India's tiger reserves, plus its own lodge, Meru Vann, in Kanha. Before Marxx, it had never run a paid ad. Every booking came from organic content and inbound enquiries.

It sells two very different products:

	NSI safari	Meru Vann lodge
Buyer	International traveller (US, UK, Canada, Europe, Australia/NZ)	Affluent domestic traveller
Average trip	11 days	3 nights
Average booking	₹5,00,000	₹1,00,000
How they decide	Weeks of research on Google	Inspiration first, then intent

A small team can't research competitors, study travellers, write, edit and test dozens of ads a week, and manage two ad platforms every day, all by hand. That is the gap an AI-native agency closes.

Marxx is a performance agency built on Marxx AI, the creative intelligence platform we built and use every day. The platform does the research, creative volume and account monitoring that usually need a large team, and our strategists make the calls. Here's what that looked like inside NSI's workspaces.

1. Competitor research: we see the whole category

Before writing a single ad, we tracked the live ads of the safari operators and lodges NSI competes with for international travellers. Every ad is searchable by hook, offer and format, so we can see what the category keeps saying and find the space NSI can own.

2. Persona research: built from 229 real traveller comments

Marxx AI clusters real reviews and comments into personas, each with the pain points behind the booking decision. For NSI that surfaced groups a demographic brief would miss, such as serious wildlife photographers who want naturalists who can predict animal behaviour, and affluent milestone travellers who need to trust an operator before sending a large payment.

3. Brand research: NSI’s brand DNA in one place

Tone, values, voice and visual system are captured once and used for every brief and every ad. For NSI that means conservation, sustainability, expertise and responsible tourism, so each ad ends on a reason to book NSI, not just "a safari".

4. In-house video editing, 50+ ideas a week

Research turns into ideas, and ideas turn into edits. We generate and rank 50+ creative ideas every week, and edit the strongest into 9:16 videos from NSI’s own footage, so an idea can go live in days, not weeks.

5. Daily account intelligence, weekly reporting

Every account is checked every day: spend, search terms, wasted spend, creative efficiency and fatigue, plus scheduled audits. NSI gets a written report and a review meeting every week.

The result: a lean team with the research depth and testing speed of a much bigger one.

International travellers don't scroll into an 11-day India safari. They search for one. So NSI's international growth runs on Google Search.

• Structure by intent. Brand campaigns protect NSI's name. Non-brand campaigns capture travellers who haven't chosen an operator yet. Competitor campaigns reach travellers comparing operators.
• Markets ranked by search demand. The US comes first, then the UK, then Canada, with Europe and Australia/NZ alongside.
• Optimised for bookings, not clicks. The platform's daily search-term and wasted-spend checks kept moving budget toward queries that end in confirmed bookings.

Cost per lead fell from ₹1,500 at launch to ₹800.

A 3-night lodge stay is bought on desire, and on Meta you have to earn attention first. We ran Meta as a creative lab:

• 50+ videos tested, edited in-house from NSI's own footage
• A deliberately flat ₹50,000 a month budget, so money went into finding winning creative, not buying extra reach
• New ideas each week drawn from competitor and persona research. Weak videos were cut quickly and winners were scaled.

NSI safaris: Google Search

Stage	Volume	Cost each
Leads	500	₹1,200 average (₹800 now)
Qualified leads	300	₹2,000
Confirmed bookings	61	₹9,836
Revenue	₹3.06 Cr	51x ROAS

Meru Vann: Meta

Stage	Volume	Cost each
Leads	1,000	₹300
Qualified leads	200	₹1,500
Confirmed bookings	53	₹5,660
Revenue	₹53 lakh	17x ROAS

Combined: ₹9 lakh of ad spend became ₹3.59 crore in confirmed bookings, a return of about 40x, from a brand that had never run an ad before.

	NSI safari	Meru Vann
One booking pays for	~5 months of Google spend	2 months of Meta spend
Cost per booking as % of booking value	2%	5.7%

In high-ticket travel, the question isn't "how cheap can leads get?" It's "how many confirmed bookings does each rupee bring in?"

1. Research before you write. Competitor, persona and brand research come first, and the ads come out of them.
2. Match the channel to how the buyer decides. Researched trips belong on Search. Desire-led stays belong on Meta.
3. Volume wins on Meta. 50+ ideas a week beats a big budget behind a few ads.
4. Watch the account every day. Wasted spend and tired creative add up quietly.

Measure confirmed bookings. That one choice changes every optimisation decision.