

“Get your hands dirty:” Safari365’s CEO shares why execs should roll up their sleeves with Agentforce.
Marcus Brain asked his team what they wanted to automate with AI agents – and boosted efficiency by 30%.
Marcus Brain asked his team what they wanted to automate with AI agents – and boosted efficiency by 30%.
Living in South Africa, I knew I wanted to do something meaningful here. Safari travel combined my personal love for the continent with the chance to create a business that shares Africa’s beauty with a global audience. When I launched Safari365 in 2006, online travel distribution was just coming into its own. Expedia was dominant, but no one was really catering to bespoke travel in markets like the U.K., Australia, or just outside the U.S. That made it the perfect time to start something disruptive.
What I didn’t anticipate was how the business would impact the remote communities here. By sending travelers deep into the bush, we’re creating new opportunities — in terms of employment or commerce — for the people living there. Helping uplift the remote communities we interact with has certainly been a motivating factor as we grow.
In the early days, we managed $30,000 custom trips with spreadsheets and Word docs. A single itinerary could involve 20 vendors across multiple destinations and our tech stack was totally fragmented. Leads came from one tool, emails were sent from another, and availability lived elsewhere. It was impossible to get a full picture of the customer journey. We had no way to track performance, spot bottlenecks, or understand where we were losing potential travelers. That just doesn’t work when you’re trying to grow and make the most of every opportunity.
We started with Salesforce about six years in. When adopting new technology, I look at the ability to expand it across my business. Like, if my 35-person staff can use it, great, but could a 350-person staff use it, too?
Salesforce became our system of record across departments. Sales consultants track every call and email in it. Our support team logs room preferences, accessibility needs, even if someone’s child gets carsick. Vendor requests, concierge preferences — it all lives there. The data richness lets us tailor trips down to the smallest detail.
Our biggest lift was rebuilding our pricing logic directly on Salesforce. It took a huge number of hours because it’s incredibly complex. We’re talking about multiday safaris across 3,000 suppliers, each with intricate rules. For example, pricing changes if an adult is sharing a room with a child versus a partner. But it was worth every bit of the effort. Now, we can price a full 12-day itinerary with all those variables in minutes, right inside Salesforce.
Honestly, the timing was perfect. We’d just completed a major data cleanup. Our biggest pain points were low-value, repetitive tasks that ate up hours — like reconfirming supplier bookings or chasing down availability across 20 properties. Agentforce could handle that. We started by asking our team, “What do you wish you didn’t have to do?”
We made a top 10 list, ran a complexity analysis, and began building agents to address those tasks. We prioritized internal use cases with high ROI and low risk. It was about freeing up human talent for the high-touch work only they can do.
For example, after we’ve priced an itinerary, Agentforce can instantly turn all the detailed trip info into a clear, easy-to-follow guide we call “day notes.” These day notes bring the safari to life for our customers by showing exactly what each day of their custom trip will look like. They help travelers know what to expect, pack the right gear, and — most importantly — get excited. This kind of repetitive task is a perfect fit for Agentforce. It saves my team hours and lets them focus on the high-touch, high-value conversations they’re best at.
Huge. Our internal efficiency is up 30%. Our teams used to manually generate two to three quotes a day and now we can deliver a first estimate in 30 minutes.
But the real magic is in the personalization. We’re really proud of our personalized introductory notes. Every quote goes out with one. These short, thoughtful messages show we’re paying attention.
Now? Agentforce writes the first draft. It pulls from everything a customer has ever shared with us, not just the most recent interaction. For example, say a customer comes back for a second trip, and says that this time, he’s treating his dad who has never seen Victoria Falls. Agentforce can go all the way back to his first conversation three years ago — where he casually mentioned, in a sentence about eight other things, that his son loves elephants — and say: “Hey, let’s get you to Victoria Falls. I also remember you have a son who loves elephants. Does he still? Want to add a close-up experience to really wow him?”
That level of memory and attention to detail is beyond what any human could do for this many customers per day.
We were very intentional about it. We started with internal agents so we could monitor everything closely. It was an iterative process. We’d test a conversation with Agentforce Testing Center, spot an issue, then tweak the instructions or adjust the topics. For example, early on it started pulling internal notes like “Customer isn’t ready to book” into messages meant for travelers. Because we built our internal agent with a human-in-the-loop, nothing went out — but it was a great reminder to pay close attention and make sure the proper guardrails were in place.
And because Agentforce understands natural language and is configured in natural language – versus being hard-coded – I could make changes myself without waiting on developer time. In fact, I built two of our agents myself. Being able to build agents and refine them on my own made Agentforce the fastest time-to-value of any Salesforce product we’ve deployed: just six weeks from signing to live.
What really made this feasible was the simplicity of the tools and a mindset of radical ownership. Agentforce is the most user-friendly Salesforce product I’ve ever worked with. I could sit down and create 10 iterations of a conversation in the time it would’ve taken to schedule one dev meeting. That kind of control keeps you close to both the customer experience and the risk.
We keep a tight feedback loop to this day. If someone catches a hallucination or issue, I can jump in and fix it right away.
I mean, I don’t want to be cavalier but I encourage everyone just to do it. Get your hands as dirty as possible. Find a good dev partner if things get tricky — we work with PS Advisory — but frankly, you can own it. The person who’s got the biggest vested interest should take ownership. That was obviously me, both as CEO, and as the person who saw an Agentforce demo, and immediately started thinking about what we could build.
Think quite deeply about your use case, consult with your team, and get their buy-in early. You need that support because even if you create something you think is amazing, you can shout about it all you want but your team isn’t going to use it if you don’t get them onboard first.
At the end of the day, you’re solving their problems. It has to matter to them.
When I first learned we could bring all of our data together in Salesforce, it was music to my ears. It turned our fragmented information into a single, incredibly rich customer view. All companies are collecting data. But organizing that data and making it accessible company-wide – that’s the unlock.
Because our data is so clean and structured, we were in a great position to launch Agentforce. We could immediately take advantage of the automation because all the inputs were already there, deeply integrated into our workflows. We sync our SQL Database to Salesforce using a connector from Synatic, and we pull in foreign exchange considerations through a company called Roam, which helps us manage global payments and exchange rates risk mitigation.
Our data is fully integrated, so Agentforce can respond with accuracy and nuance like a well-informed teammate. It’s what turns automation from a novelty into a real driver of impact.
We’re going all in on Data Cloud. Right now, so much of our staff’s time is spent calling small bush camps and digging through third-party sites. That’s not structured Salesforce data. But with Data Cloud, we’ll be able to marry external data — like TripAdvisor reviews or Booking.com content — with our internal records. It’ll help us vet new suppliers, enrich itineraries, and take our personalization to the next level.
Being able to build agents and refine them on my own made Agentforce the fastest time-to-value of any Salesforce product we’ve deployed: just six weeks from signing to live.
Marcus BrainCEO, Safari365