I run the sales and marketing department here at Cintra. Over the last 24 months we have taken a new product to market—an AI-first expense management platform built by our CTO, James Rowell. It has taken us from a new market entrant in the expense management space to having the best product in the UK in a very short time. The speed and quality have been remarkable to witness, and it has opened my mind to the potential of AI and the future that is in front of all of us.
It also forced me to prioritise the application of AI inside sales and marketing. AI is here now, and from a pragmatic business perspective it must be embraced fully.
The problem we're solving first
I want on-demand access to the data that drives commercial decisions: pipeline, win rates, conversion, sales cycle length, channel performance, deal size trends.
For example, I might need a table showing closed deals over the last twelve months for a given product line and deal size band, with a pie chart next to it breaking those deals down by industry. I want to request that in natural language and get it back immediately, in that format.
CRM tools offer custom reporting, but it is rarely quite what you need, at the time you need it, in the format you need it. It’s fiddly, often requires manual configuration, and when you want something bespoke, someone must build it.
We have a strong Sales Operations Manager who does a lot of this work. The issue is not his ability. The volume and variety of reporting we want, at the pace we want it, is not sustainable as a manual process. AI changes that.
How we’re bringing AI into the department
Our Sales Operations Manager is moving into a Go to Market Engineer role—something we have seen emerging in companies leading on AI adoption. The purpose is to build AI capability directly into our commercial workflows, starting with reporting.
His first brief is to build an agent that takes a natural language request for data and returns it in whatever format is needed. Tables, charts, summaries—all drawn from our CRM, pipeline data, and performance records.
The fear-based way of seeing this is that he is creating an agent to do his own job. That is not how we see it. His value is not in manually building reports. It is in understanding the business, the data, and what good output looks like. That’s the hard part. The tools to build AI agents are now accessible to people with strong domain knowledge and data literacy—you don’t need to be a developer, though you may need support from one initially. That is precisely why transitioning an existing ops person makes more sense than hiring an engineer from outside.
Humans and AI in our department
Once you have someone dedicated to building AI capability inside a commercial team, the applications go well beyond reporting. We can create deep, real-time insight into our market, our customer base, and how we are executing—understanding prospective customers, existing clients, and our own performance in ways that were previously too slow or too manual to sustain.
The pessimistic view of this is that it replaces people. The optimistic view is that it gives them all superpowers.
The reality, as we see it, is somewhere in between. In the short term, we may not need as many people in certain functions while we get to grips with the power of AI, but we are not reducing headcount. We view AI not as a replacement for humans—but those that embrace it and use it effectively are undoubtedly more valuable.
Humans provide the creative spark, the compliance oversight, and the directive guardrails. Longer term, the economy and the world of work will transform and adapt.
There will be disruption and discomfort along the way. But in the best case, AI improves speed, quality, depth, and outcomes. It is not without its risks.
Building and buying
Our situation is specific. We’re an AI-enabled technology business with a CTO building advanced AI into our products. We have the resource to build in-house and a commercial reason to do so.
Not every business is in that position. There are strong off-the-shelf AI products that deliver real improvements in quality, productivity, and speed. For some, the right approach is to select the best tools available and learn to use them well. For others, it is a combination. The important thing is to start rather than defer.
How we approach AI-led services
We are custodians of important data across payroll, HR, and expense management. We have deep compliance expertise and significant technical resource. That puts us in a position to build real capability on top of our product set—tools that do not just retrieve information but surface patterns, identify risk, and generate insight from the data our clients already trust us to manage.
Our expense management platform already demonstrates this. Users input data in natural language and request reports on demand—no templates, no manual configuration. The same principle applies across our product range: trusted data, accessible on demand, in the format you need.
That is what a CFO wants when reviewing spend across cost centres. It is what a payroll manager needs before a compliance deadline. It is what an HR director is looking for when pulling headcount by department, tenure, and location.
What comes next
This is a journey we are starting, not a finished case study. The technology exists, the approach is proven by trailblazer sales and marketing departments, and the logic is straightforward: commercial teams need someone dedicated to building AI capability and someone accountable for managing it.
We need to deliver our clients real, tangible value while remaining compliant, secure, and steady. The key is to get started and move quickly—but carefully.
If you are working through similar decisions in your business, I would be happy to compare notes.
