A couple of weeks ago Fin365 wrapped up our 2^nd annual Symphony Conference. By all measures it was a resounding success. Two days jam packed with insightful discussion about how technology is the key to helping financial services firms deliver their valuable services to a greater number of Australians.

Of course, AI was the hot topic. So hot that most of the room skipped morning tea to pepper us with questions after our live demo session in which we unveiled the new AI agents we’re developing.

However, reflecting on the conference during my 24-hour flight back to Texas, the thing that stood out most was not the hunger to learn more about AI. Rather, it was the pervasive understanding that AI would only deliver results if the underlying foundation was sound. While 2025’s winner #whynotmicrosoft still made an appearance, #datadatadata became the top hashtag of the conference. Perhaps I shouldn’t have been surprised. Data and process have been two of the key themes in Fin365’s hymnbook since our inception. But, in an age where our industry is once again being inundated with sparkly new toys, it was refreshing to hear such a large choir choosing classics over trendy chart-toppers.

Unfortunately, in the marketplace of ideas, especially the crowded, noisy arena of social media, sizzle almost always outsells substance. And right now, there’s a seductive new melody making the rounds that threatens to drown out the #datadatadata message entirely and replace it with #aiwillfixeverything.

The narrative goes something like this: AI is coming. Traditional CRM (or Saas) is dead. Stop investing in your systems, wait for the agentic revolution, and the shiny new tools that will solve all your problems.

You’ll find versions of this argument in newsletters, on LinkedIn, and on conference stages. Bill McDermott, CEO of ServiceNow, who, perhaps not coincidentally, recently launched a new CRM product, declared agentic AI “an extinction-level event for traditional CRM systems” on Bloomberg last year. The CEO of Reevo and others have echoed this sentiment. The SaaSpocalypse crowd is loud, well-credentialled, and very confident.

There’s just one problem. They’ve described the solution perfectly and then told you not to invest in it.

What is CRM?

The term “CRM” has been so thoroughly colonised by software vendors that most people now treat it as a product category; a Salesforce licence, a HubSpot subscription, a Dynamics dashboard. This conflation is understandable, but not helpful … unless you’re trying to declare its extinction.

CRM (Client Relationship Management) is not a software product. It is a business discipline, i.e. the structured practice of acquiring, understanding, serving, and retaining clients over time.

For the purposes of this argument, and with a focus on the financial services industry, CRM, properly understood, comprises seven interdependent functional pillars:

1. Client Data & Identity: who your clients are: their profiles, goals, risk tolerances, family structures, life events, and history. The atomic unit of everything else.
2. Files, Documents & Activity: the longitudinal record of everything that has happened in the relationship: meeting notes, advice documents, compliance records, signed authorities, review packs, and audit trails. The answer to “what has actually happened with this client, and when?”
3. Workflow & Process Management: the defined, repeatable sequences of actions that move a client relationship forward: onboarding flows, review cycles, compliance checklists, paraplanning handoffs, implementation steps.
4. Client Communications & Engagement: the managed cadence of outreach, education, and engagement across the client lifecycle: review reminders, targeted campaigns, life event triggers, referral programmes.
5. Reporting, Intelligence & Performance: pipeline visibility, retention metrics, advice capacity utilisation, compliance adherence, and business health dashboards. The feedback loop that turns outputs into decisions.
6. Integration & Ecosystem Connectivity: the APIs, middleware, and data feeds that connect client data across portfolio platforms, advice tools, document management systems, and communication platforms. The nervous system of the operating stack.
7. Compliance, Consent & Governance: the accurate, auditable, tamper-evident record of client consent, advice given, disclosures made, and obligations met. Non-negotiable in a regulated industry.

None of these pillars are going away. Not one of them becomes irrelevant because AI is getting smarter. In fact, and this is the central argument of this article, every single one of them becomes more important as AI becomes more capable.

Why the sensational claims?

Let’s take a moment to examine where this narrative is coming from, and who benefits from it.

Bill McDermott’s “extinction-level event” declaration arrived on Bloomberg at the same time ServiceNow was launching its own CRM product, built from the ground up on agentic architecture. McDermott is a brilliant CEO and a masterful communicator. He is also, in this instance, doing what brilliant CEOs do: framing a market shift in terms that position his company’s new product as the inevitable winner.

Similarly, the SaaSpocalypse narrative, amplified by a sharp correction in SaaS stock prices through late 2025 and early 2026, has been enthusiastically adopted by consultants, analysts, and newsletter writers whose incentive is to appear ahead of the curve. The more alarming the prediction, the more newsletter opens, the more conference keynote slots.

This is not to say the underlying technology shift isn’t real. It absolutely is. AI is transforming how software is built, sold, and used. Agentic systems will change the way client-facing workflows are executed. These are genuine and significant shifts.

But “the tools are changing” is a very different claim from “your data, your files, your processes, and your client relationships no longer matter.” The first is insightful. The second is reckless.

Without a hint of irony

Humorously (not really), the very same voices telling you AI will make your CRM obsolete are, in the same breath, telling you that AI requires clean data, integrated systems, accurate activity records, and well-defined workflows to function.

They’re right about the second part. They just haven’t connected the dots.

Let’s be precise about what AI needs to deliver value in a client-serving business:

• Structured, accurate client profile data – so it knows who it’s working with
• A complete activity and document history – so it understands the context of the relationship
• Defined workflow rules and process logic – so it knows what actions to take and when
• Integration across the technology stack – so it can pull data from advice platforms, communication tools, and portfolio systems

Translation? AI NEEDS your CRM. Or, as the old saying goes Garbage In, Garbage Out^[1]

A firm that neglects its CRM because “AI is coming” is like a restaurant that stops stocking the pantry because a robot chef is arriving. The robot still needs ingredients.

What “Systems of Action” actually need to act