It is honestly impressive how we can spend over one hundred and twenty-six billion dollars as an industry on a single category of software and still have more than half of the people using it say it does not work for them. I was looking at the latest Wave Connect report from earlier this month, and that fifty-five percent failure rate for implementations is just staggering. It makes you wonder if we are actually buying tools or just buying digital chores.
Herman Poppleberry here, and you are hitting on the fundamental tension of the enterprise right now. Today's prompt from Daniel is about exactly that, the state of the customer relationship management market here in March of twenty-six. It is a fascinating moment because the CRM has become the largest enterprise software category in the world, yet we are seeing this massive split between the old guard and the new AI-native entrants.
It feels like the industry is finally hitting a wall with what I call manual debt. You know, that friction where a salesperson spends more time logging calls than actually making them. Daniel is asking us to look at the contrast between the monoliths like Salesforce and the newer players like Attio. And honestly, looking at the numbers, the gap is becoming a canyon. We are moving from what we used to call systems of record to what the industry is now labeling systems of intelligence.
That is the crucial distinction. For thirty years, a CRM was a digital filing cabinet. You had to open the drawer, put the paper in, and organize it yourself. A system of intelligence, however, is more like having a chief of staff who watches everything you do and organizes the files before you even think to ask. We are talking about a market projected to hit one hundred and twenty-six point one seven billion dollars this year, and yet the most successful players are the ones trying to make the software invisible.
Which brings us to the heavyweight champion. Salesforce is still the undisputed leader with nearly twenty-two percent of the market. They are looking at over forty-one billion dollars in revenue for this fiscal year. But Herman, the vibe around them has changed. Their stock is down thirty-four percent over the last twelve months. Why is the market cooling on the king?
It is a classic valuation reset. Investors are no longer looking at Salesforce as this hyper-growth disruptor. They are treating it like a mature value stock, almost like a utility company. Even though they just reported eight hundred million dollars in recurring revenue just from their Agentforce suite in the fourth quarter, the growth narrative has shifted. People are looking at the implementation times. For a mid-market firm, you are still looking at an average of four and a half months to get Salesforce fully operational. In twenty-six, four months feels like an eternity.
Especially when you compare it to the new kids on the block. We have to talk about Attio. They are London-based, led by Nicolas Sharp and Alexander Christie, and they are basically the poster child for this AI-native shift. Their average implementation time is sixteen days. Not months, days. When you look at that sixteen-day window versus the four-and-a-half-month slog with a monolith, you start to see why the legacy players are nervous.
And it is not just about speed; it is about the architecture. This is where we need to get into the weeds a bit. Attio is built on an AI-first architecture. They recently launched Ask Attio last month, which is a conversational interface that uses something called the Model Context Protocol, or MCP. We touched on this back in episode fifteen hundred when we talked about the new era of agentic AI, but MCP is really the secret sauce here.
Explain that for the non-engineers, Herman. Why does MCP matter for a salesperson trying to hit their quota?
Think of the old way as a walled garden. If you wanted Salesforce to talk to your email, your Zoom, and your LinkedIn, you needed these massive, brittle custom APIs. You often had to pay a consultant fifty thousand dollars just to make the pipes connect. MCP acts like a universal translator for AI agents. It allows the CRM to reach out and interact with data across your entire tech stack without needing those custom-built bridges. It turns the moat into a bridge.
That explains the adoption numbers Daniel pointed out. Attio has a ninety-one percent feature adoption rate. Salesforce is sitting at fifty-two percent. If you are paying for a massive suite of tools but your team only uses half of them because they are too complex to set up, you are essentially lighting money on fire.
And the cost difference is wild. If you look at the total cost of ownership over three years, Attio is coming in at around twelve hundred dollars per user. Salesforce? You are looking at over eight thousand dollars per user on average. That is an eighty-five percent reduction in cost. For a mid-sized tech company with a hundred users, that is the difference between seven hundred thousand dollars in savings or a massive line item that might not even be fully utilized.
But Salesforce is not just sitting still. Marc Benioff is betting the entire company on Agentforce. They are leaning into these autonomous agents that can handle customer service, lead qualification, and scheduling. Eight hundred million dollars in revenue for a new product line in one quarter is a massive signal that their existing customers want this.
They do want it, but they are struggling with the underlying substrate. This goes back to our discussion in episode six hundred and sixty-seven about the evolution from AI-washing to AI-first. Salesforce is trying to bolt a jet engine onto a horse-drawn carriage. The carriage is their legacy data model—objects, leads, accounts, all defined decades ago. To make Agentforce work, you still need that four-month implementation to clean up your data and map everything correctly. Attio started with a blank slate where the data is fluid and the AI is the primary way you interact with it from day one.
It is the difference between a system that waits for you to tell it something and a system that observes. I saw a stat that sixty-five percent of sales teams using mobile-first CRMs are meeting their quotas, compared to only twenty-two percent for those who are tethered to a desktop. That tells me the "system of record" model is failing because sales does not happen at a desk anymore.
It happens in the world. And if your CRM is not in your pocket, proactively telling you what to do next based on a transcript it just ingested from your last meeting, it is just a digital filing cabinet. We are seeing this massive productivity gap. If you are a salesperson and you have to spend your Sunday nights logging data from the week, you have what we call manual debt. That debt compounds until the data in the CRM is so out of date that it becomes useless.
That brings up a big debate Daniel mentioned: is the CRM actually going to survive, or will AI just replace it? If the AI knows everything from my email and calendar, why do I even need a structured database?
This is the "Brain versus Memory" debate. The industry consensus in twenty-six is that AI is the brain, but the CRM is the memory. If you let an AI operate purely on unstructured data—just raw emails and chats—it eventually loses the thread. It starts hallucinating details about contract terms or what stage a lead is in because it lacks a ground truth. You need a structured system of record to keep the AI honest. The winner in this market is whoever makes that structured record the easiest to maintain without human intervention.
So the CRM becomes the guardrail for the AI. It provides the facts so the AI does not make things up. That makes a lot of sense. But we also have to look at who is actually using these tools. Tech companies are at ninety-four percent adoption, manufacturing is at eighty-six percent, but micro-businesses—companies with under ten employees—are only at fifty percent. Why is half of that market still holding out?
It is the complexity floor. If you have five employees, you do not have a CRM administrator. You do not have time to learn a complex system. For a long time, your only options were a spreadsheet, which is free but dumb, or a CRM that was too expensive and too complicated. This is where HubSpot has been dominating. They have sixty-two percent of installations in the small and medium business space because they have managed to keep the interface relatively simple.
And they have their Breeze AI agents now, trying to bridge that gap. But even HubSpot is feeling the pressure from these hyper-simplified, mobile-first tools that just ingest data and stay out of the way. It feels like the "middle ground" is a dangerous place to be right now. You either need to be the massive enterprise vault like Salesforce or the invisible AI partner like Attio.
Speaking of the enterprise vault, we should talk about the regulatory side. There is a major piece of legislation we are watching: the American AI Leadership and Uniformity Act, or H R fifty-three eighty-eight. This is a big deal for the CRM world.
Right, because currently, it is a total mess of state-level regulations. If you are a company in Texas using an AI agent to process data from a customer in California, you are navigating two different sets of privacy and automation rules.
H R fifty-three eighty-eight is aimed at creating a federal framework that supersedes this patchwork. For the big players like Salesforce, they have the legal teams to handle the chaos. But for a startup like Attio or a mid-sized firm trying to deploy AI agents, the compliance overhead is a massive barrier to entry. A federal standard would level the playing field. It is a pro-growth move that reinforces American leadership in the AI space by making it easier to scale these autonomous systems across state lines.
It is also about trust. If there is a uniform set of rules for how these "systems of intelligence" handle our data, businesses will be more comfortable letting the AI take the wheel. Right now, only sixteen percent of companies have successfully integrated AI into their core CRM workflows, even though eighty-three percent are using some kind of generative AI feature. That gap is largely due to fear of hallucinations and regulatory uncertainty.
And that brings us back to the "manual debt" crisis. If you do not trust the AI to log the data, you are stuck with the human doing it. And we know humans are bad at it. That is why fifty-five percent of implementations fail. They do not fail because the software is broken; they fail because the data inside them is garbage.
So, if you are a business leader listening to this, what is the actual playbook for twenty-six? Because it feels like the old advice of "just buy Salesforce and you will be safe" is not necessarily true anymore.
The first takeaway is to stop buying systems of record and start evaluating systems of intelligence. If a vendor's pitch involves your team spending hours on data entry, you are buying a legacy problem. You want to prioritize tools that offer automated ingestion from the jump. If it does not automatically pull from Zoom, email, and LinkedIn, it is not a twenty-six CRM.
I would add that you need to look at the integration philosophy. We talked about the Model Context Protocol. If a vendor is still talking about proprietary APIs and trying to lock you into their "ecosystem," they are building a moat that will eventually trap you. You want support for open protocols like MCP so your CRM can talk to whatever AI agents you deploy in the future.
And do not ignore the mobile-first workflow. The data is just too compelling. If your sales team is not meeting their quotas, check their CRM usage. If they are tethered to a desktop to update their pipeline, you are losing forty percent of your potential productivity right there. The CRM has to be where the work happens, not a place where you report on the work after it is done.
It is about reducing the distance between the action and the data. The closer those two things are, the better your intelligence will be. Herman, what do you make of Marc Benioff's recent move? Salesforce just started a twenty-five billion dollar accelerated share repurchase on March sixteenth. That is a massive amount of cash to put into buying back their own stock.
It is a defensive move. When you can no longer point to thirty percent year-over-year growth, you use your massive cash reserves to prop up the stock price and provide value to shareholders. It is a signal that the "Salesforce era" of hyper-growth is transitioning into a "Salesforce era" of being a stable, cash-generating utility. They are still a powerhouse, but they are no longer the ones setting the pace of innovation. That pace is being set by the companies that are making the CRM invisible.
It is the classic innovator's dilemma. Salesforce has the revenue and the customers, but they also have the legacy code and the legacy mindset. Meanwhile, Attio is proving that a smaller, more focused architecture can deliver better results at a fraction of the cost. That sixteen-day implementation versus the four-month implementation is the stat that should keep every enterprise executive up at night.
It really should. We are seeing the death of the digital filing cabinet and the birth of the autonomous partner. Whether Salesforce can successfully pivot with Agentforce or if we are watching a long, slow transition to a new market leader is the big question for the rest of the decade.
I think we have given Daniel a lot to chew on here. The CRM market is bigger than ever, but it is also more divided. You have the monoliths trying to evolve and the AI-natives trying to disrupt, and the winner is going to be whoever can actually kill the manual debt once and for all.
Well said. It is a big world out there in the CRM space, and it is getting smarter every day. I think we covered the main threads Daniel was looking for—the contrast in players, the adoption trends, and the regulatory shifts.
Definitely. We should probably wrap it up there. Thanks as always to our producer Hilbert Flumingtop for keeping the gears turning behind the scenes and making sure we do not hallucinate our own facts.
And a big thanks to Modal for providing the GPU credits that power the generation of this show. We could not do these deep-dive explorations into the state of tech without that kind of infrastructure support.
This has been My Weird Prompts. If you are finding these deep dives useful, a quick review on your favorite podcast app really helps us reach more people who are trying to make sense of this fast-moving landscape. It helps the algorithm find us, and we appreciate the support.
You can also find us at myweirdprompts dot com for the full archive and all the ways to subscribe. We have got all our previous episodes there, including those deep dives on agentic AI and the agency evolution we mentioned earlier.
Catch you in the next one.
See you then.
