Dear Reader,
Welcome to Ask Achan. Who's Achan, you ask? He's one of our in-house HubSpot experts and the Head of HubSpot Delivery here at Concentrate.
Each month, we dive into your HubSpot questions with simple, actionable tips to help you get the most out of the platform. This month's question comes up a lot in sales leadership circles right now — so let's get into it.
Got a question? Send it my way.
Quick answer: AI can write a CRM prototype in a weekend. It can't take on 24/7 uptime, data security, integration upkeep or the next ten years of maintenance and that's where the real cost of "building your own" shows up.
It's worth addressing the premise head-on. Yes, AI can write a working CRM interface fast but there's a real difference between an AI writing code for you and a platform built on AI trained across millions of customer interactions, like HubSpot's predictive lead scoring and forecasting. A homegrown tool calling an AI API is doing something shallower, and HubSpot's own research shows even mature platforms still face real barriers around AI accuracy and data quality. Those problems don't disappear when you build it yourself they just lose their support team.
Here are the three risks that get missed most often.
AI can spin up forms, pipelines and dashboards quickly. What it can't do is make the years of decisions that determine whether your data actually holds up under real business use. For example, how contacts, companies and deals relate to each other, how duplicates get prevented, how your data model scales as you add products, regions or business units.
Then there's integrations. Your CRM needs to talk to email, calendar, your website, your accounting platform, maybe a quoting tool and every one of those connections needs to keep working every time the other platform updates its API. That's not a one-off build. It's an ongoing job.
To put the scale of that in perspective: HubSpot's App Marketplace alone lists nearly 2,000 pre-built, vendor-maintained integrations, and the average customer connects more than five of them. That's not five things to build once it's five things somebody has to keep working, forever.
This all means that while an AI-built custom CRM might give you functions quickly and cheaply, it can’t easily leverage what’s most important - context. That mine of information you have about customers, prospects and partners.
Achan top tip: "AI can write the code for an integration. It can't tell you what happens to your reporting when that integration silently breaks at 2am on a Friday, and nobody notices until month-end."
A CRM holds your most sensitive commercial data for example, customer, deal values, contract terms, sometimes payment details. That makes it a genuine security target.
The stakes are real: IBM's 2025 Cost of a Data Breach Report puts the global average cost of a breach at USD $4.44 million. Platforms like HubSpot invest heavily in access controls, encryption, audit logging and compliance certifications backed by dedicated security teams and uptime commitments. When you build in-house, all of that responsibility sits on your business including keeping pace. If your CRM goes down the morning of a big sales push, there's no support line to call. It's whoever on your team can be reached first.
The build is the cheap part. The next ten years are where the real cost lives.
Someone has to patch it, host it, back it up, and fix it when it breaks. Someone has to update it every time your sales process changes. And because most in-house CRM builds are the work of one developer or a small team, that person becomes a single point of failure. If they leave, they take the only real understanding of how the system works with them and every future fix starts with someone else reverse-engineering their own company's software.
Left unfunded, the build also stops evolving so what may seem cheaper up front is usually more expensive in the long term. With a CRM product, you get access to the combined wisdom of hundreds of thousands of businesses and their input into features. Wisdom that is updated into your software continuously.
Achan top tip: "The real question isn't 'can we build a CRM?' It's 'do we want to be a software company as well as running our actual business?' For most teams, the answer is no, they just haven't costed it out yet."
None of this means AI has no place in your CRM strategy. Quite the opposite.
Platforms like HubSpot are already investing that engineering effort at scale - purpose-built data architecture, maintained native integrations, enterprise-grade security, and AI features built and tested by dedicated product teams rather than one AI-assisted developer. You get the benefit of that investment without carrying the risk of it.
There's also an irony in the "ownership" argument: a custom build is usually harder to walk away from than a SaaS platform, not easier and you can export your HubSpot data and migrate any time, but unwinding a bespoke system nobody else understands is a much bigger job.
What HubSpot doesn't do on its own is set itself up around your sales process, your data, and your team's way of working. That's where a local implementation partner earns its keep, configuring the platform properly the first time, keeping it clean as you grow, and being a phone call away when something needs fixing, rather than a support ticket into the void.
If you can't answer all three with confidence, that's usually your answer.
If you're weighing up build-vs-buy for your CRM, the team at Concentrate can help you think it through. Contact the Concentrate team today.
See you in the next edition of Ask Achan.
Got a question for Achan?
What's your biggest HubSpot challenge? Ask below.