Software 3.0:
From Spreadsheets to the Cloud, to Agentic AI
By Stan Altshuller & Charles Poliacof
- Introduction
- Software 1.0 → 3.0 From Excel to Salesforce to Autonomous Agents
- The Big Merge Convergence of Systems of Record and Systems of Action
- A La Carte Unbundling of Bloated CRM Functions
- Impact and Examples Goodbye Dashboards, Hello “Do It For Me”
- AI-first CRM Experimentation The CRM Metamorphosis
- Parting Thoughts
Enterprise software is going through a profound structural transformation. The rise of agentic AI, or AI systems that can reason and act autonomously on disparate information, promises to fundamentally change how we manage customer relationships and data. One can argue that it could lead to a profound transformation, and in some ways a complete re-imagining, of traditional CRM software. To understand why, let’s look at how we got here and what’s coming next.
Written by:
From Excel to Salesforce to Autonomous Agents
Software 1.0 – The Spreadsheet Era
From the early 1980s to the late 1990s, tools like Microsoft Excel dominated the business world. During this period, businesses meticulously tracked contacts, sales, and finances using manual spreadsheets. While this method offered flexibility, it was also incredibly labor-intensive and error prone. This era was characterized by individuals directly manipulating data, often spending countless hours in Excel cells debugging broken formulas and chasing bad links in massive spreadsheet reports. This was a form of “software” in a very manual sense.
Software 2.0 – The Cloud CRM Era
The late 90s and 2000s brought cloud-based CRM systems like Salesforce (founded 1999) and later HubSpot. These were enabled by the relational database and the internet. Instead of ad-hoc spreadsheets, companies now had a system of record – structured data on leads, opportunities, accounts, all accessible online. This made basic automation more feasible and made reporting more collaborative and scalable beyond spreadsheets. CRM became synonymous with a central database for customer data (rows and columns, picklists, forms). Salesforce and other similar systems grew into massive suites of features, dashboards, and apps in the cloud. However, these systems still largely relied on humans to input data, interpret reports, and take action often outside of the system of record. They gave us abundant eye-candy like dynamic dashboards and charts in, but the software didn’t do much on its own beyond what we told it us. Charts and tables, tables and charts.
Software 3.0 – The Agentic AI Era
This paradigm of a divide between the systems of record and agents of action is changing fast. We are now entering a new stage of software powered and driven by AI. Software that not only writes itself but reasons and acts by itself. Instead of being just a database, today’s generation of platforms are AI-native.
AI agents have no need for table structures or visualization of data; they can tap into it directly and understand it instantly. They are built for ingesting unstructured data (text, emails, calls, video) and taking actions autonomously and with a high degree of accuracy. As Andreessen Horowitz put it, “We believe AI will… fundamentally reimagine the core system of record… Instead of a text-based database, the core of the next sales platform will be multi-modal... containing every customer insight from across the company. An AI-native platform will be able to extract more insight from a customer and their mindset than we could ever piece together with the tools we have today.” In other words, the CRM of the future isn’t a static database but a smart AI agent that is constantly learning from all customer interactions in real time and completing tasks on your behalf. This is a radical shift: the information layer and the action layer in enterprise software effectively merge into one. We can stop thinking of software as just a tool to store data and start treating it as a peer or colleague that can analyze, reason, and act. This is Software 3.0, it is here now, and it looks a lot less like forms, charts, and tables and more like an autonomous, virtual team member.
Convergence of Systems of Record and Systems of Action
The most impactful shift happening with agentic AI is the merging of “system of record” with “system of action.” In the past, we had systems of record (like a CRM database where data is stored) and separate systems of engagement or action (the emails we send in an email client, the phone calls we make, the marketing tools that help us make reports and then communicate with clients). These concepts are now all converging. Why? Because generative AI has bridged the gap between raw data and action in real-time.
AI agents are using raw communications directly as data skipping the manual step of transcribing, normalizing and recording that data. There is no need for that intermediary work as AI agents plug directly into emails, take real time call transcripts, ingest chat logs, client data bases, and even scrape online data without needing a human to re-enter any of it into a CRM form.
For example, today’s AI-powered CRMs automatically pull add data from emails, calls, meetings and even social media posts directly into customer records. Instead of waiting for a salesperson to log a call outcome or copy-paste an email, an AI agent can read the content of the interaction itself, score a lead based on established criteria and reach out just at the right time with the right content.
AI treats unstructured text, web scraped data, or voice as just another input. This means the AI is aware of customer communications as they happen, not weeks later when someone might manually update a field and decide to check in on a prospect. The traditional CRM “pipeline” (with staged data entry and handoffs) starts to disappear as the AI is listening and logging everything at once. And when it has just enough data, it acts. Here is what CRM looks like in an AI agentic world.
Agents act on insights before data hits the CRM
Because AI monitors raw feeds of customer interactions, it can glean insights and trigger actions immediately – compressing the value chain. For instance, an AI sales assistant could detect from an email thread that a customer has a product concern or a competitor’s quote before that ever gets formally noted in a CRM ticket. It might alert the account manager or even draft a proactive response right away. In the old model, that insight might be lost or delayed (“dumb data” sitting in an inbox until someone manually analyzes it). In the new model, the AI has already analyzed it. Traditional CRM implementations often end up as expensive data graveyards filled with incomplete info that no one acts on. Agentic AI turns those graves back into gardens – continuously nurturing insights from the raw data on the fly. The CRM becomes less a living interface for users and more a behind-the-scenes data store that the AI draws from. In fact, we might not even call it “CRM” anymore – it’s just the company’s knowledge, which the AI can query at will.
“Dumb data store” or smart collaborator?
If today’s CRM is a passive repository, AI aims to make the whole system proactive and intelligent. By effectively merging your system of record and system of action, using AI as the orchestrator, you end up with less reliance on humans navigating dashboards or cross-checking multiple tools and an AI agent with a full view of the customer across email, Slack, support tickets, etc., that can act – send an update, schedule a follow-up, flag an opportunity – all in one flow. The CRM database is still there underneath (you need a place for all the data to reside), but it’s the AI “brain” on top that’s doing the heavy lifting. This can even eliminate a lot of the tedious swivel-chair work employees used to do, moving data from system A to system B.
The result?
A dramatic increase in productivity and reduction in busywork. Sales teams have long complained about the tedious grunt work of CRM. The onerous job of entering notes, updating fields, creating manual reports, and logging calls have added to the drudgery of the sales operation and kept usage CRM muted. (Salespeople are famously reluctant CRM users: they want to sell, not fill out forms.) Surveys back this up: sales reps spend only ~28% of their time actually selling; the rest is eaten up by admin tasks, data entry, and meeting prep. The pie chart below highlights this imbalance clearly – over two-thirds of a rep’s week is spent on “non-selling” tasks!
Sales reps waste a huge portion of their week on administrative overhead (updating CRM, writing follow-ups, logging activities) instead of actual selling. Agentic AI aims to shrink that yellow slice dramatically by automating data entry and next-step planning.
Agentic AI directly attacks this inefficiency. By auto-capturing call notes, updating pipeline stages, and even drafting emails, AI gives sellers back their time. Early evidence is compelling – for example, we have heard current AI can already automate about 40% of sales “busywork”, and that number will only rise as natural language processing gets better. Imagine a near future where after every customer call, your AI assistant instantly updates the opportunity status, generates a follow-up email, schedules the next meeting, and flags any risks – all before you’ve even finished saying “goodbye” to the client. In essence, the AI agent is the new system of action, sometimes even suggesting the next best action, all while sitting on top of (and continually updating) the system of record. The salesperson and the AI work in tandem – the human focuses on strategy and relationship-building, while the AI handles the tedious tracking and initial responses. It’s a much more “hands-free” CRM experience. No more spending your Friday afternoons slogging through Salesforce updates – your AI sidekick has already done it.
Unbundling of Bloated CRM Functions
Traditional CRM suites became bloated over time – packed with features, add-ons, and analytics modules aimed to do everything for everyone. But the reality is most companies never use half of these features. In fact, studies show roughly 80% of features in the typical cloud software product are rarely or never used! All that extra code, sitting idle, yet you’re paying for it. (Think about it: when was the last time you used every bell-and-whistle in your CRM or ERP system? Probably never.) This overkill not only complicates the user experience – it also drives up costs. Why pay for a Swiss Army knife when all you needed was a scalpel?
AI is now unbundling these core CRM functions into smarter, focused tools. Instead of one monolithic platform attempting to do lead scoring, email marketing, forecasting, customer support, website analytics (and so on), we see a trend toward specialized AI-driven solutions for each major task. And they often perform better than the generic feature inside a big CRM suite. Let’s break down a few key functions being disrupted:
Lead Scoring & Prioritization
CRMs have long offered rule-based lead scoring (e.g., +5 points if they clicked our email, +10 if title contains “Director”, etc.), but it was often crude. AI has supercharged this. Machine learning models can analyze myriad data points like firmographics, web behavior, email engagement, even tone of communications to predict which leads are most likely to convert. The result is far more accurate rankings of your prospects. In one survey, 98% of sales teams using AI said it improved lead prioritization and focusing on the right leads. These models can even learn and adjust over time as deals either close or fall through. Salesforce’s own AI uses both your internal data and anonymous aggregated data to score leads, and refreshes those scores every 10 days as new information comes in. This is a task tailor-made for AI. AI is great at pattern recognition across many signals, something humans struggle with. So now, instead of sales reps guessing or following hunches, they get a dynamic, AI-driven “hot lead” list constantly updated in the background. Several startups (and new tools from incumbents) focus on this single area of excellence: making sure no high-potential lead slips through the cracks. In short, AI lead scoring takes the guesswork out, replacing gut feel with data-driven predictions.
Personalized Email Outreach at Scale
Another core CRM function is email marketing and sales outreach. Traditionally, you’d create templates and maybe use mail merge fields for personalization (Hi , hope is doing well…). That’s surface-level personalization. Today’s generative AI can do so much more. AI-powered sales engagement tools can research each prospect (scraping news, LinkedIn, social media) and generate a highly personalized intro or email that reads as if you spent 30 minutes researching them – but it’s done in seconds. For example, there are AI tools now that given a list of targets, will automatically find relevant tidbits (like “saw you just opened a new office in Dallas” or “noticed your CEO mentioned a focus on AI in a recent interview”) and insert those into outreach emails to dramatically lift response rates. These personalized touches at scale were nearly impossible for a human team to do manually – you’d need an army of researchers. Now an AI agent can do it 1,000 times over in minutes. The quality of AI-generated writing has also improved massively – often indistinguishable from a human rep’s style (and frankly, sometimes better, since the AI can be coached to use best-practice persuasion techniques). Buyers are seeing the difference: they’re more likely to respond when the outreach is truly relevant to them. In fact, 73% of customers expect companies to understand their unique needs, and AI helps deliver on that expectation. Startups in this space (email assistants, AI copywriters for sales like Lyne.ai, Regie.ai, Lavender, etc.) are nibbling away at the edges of traditional CRMs’ email modules. They don’t need you to log into a big CRM to send an email blast – instead, an AI layer analyzes and sends on your behalf, often connecting to email servers directly or via lightweight tools.
Pipeline Forecasting & Analytics
Forecasting has historically been a headache for sales managers – rolling up spreadsheets or CRM reports full of subjective data (rose-colored commits, sandbagged deals, etc.). AI is radically improving forecast accuracy by analyzing deal progress signals and historical patterns that managers might miss. By crunching past win/loss data, engagement metrics, and even external factors, ML models produce a more realistic picture of what revenue is likely. AI looks at actual behavior and data – how many calls happened, response times, product usage, comparable deals – to predict outcomes more objectively. It might flag a deal that’s “at 80% stage” in CRM as actually unlikely to close due to certain risk factors (no recent contact, budget not discussed, etc.), or conversely identify a “dark horse” opportunity that’s more promising than it looks. The AI essentially does scenario analysis dynamically. Clari is a well-known player in AI forecasting; so are People.ai and others embedding predictive analytics into pipeline management. The result: fewer surprises and a tighter handle on the business. And because the AI constantly refreshes with new data, forecasts become a living model rather than a once-a-quarter fire drill.
Automating Sales Drudgery (the AI Sales Copilots)
Beyond specific tasks like scoring or forecasting, there’s a general unbundling of workflow happening. Many tasks that a salesperson or support person would do manually are now being handled by targeted AI assistants. Some examples:
CRM data entry and hygiene
We all know keeping CRM data clean is painful. AI can automatically log calls, parse out key points, update contact records, and even enrich data from external sources. Tools like People.ai’s SalesAI connect to email and calendars to capture every customer touchpoint (so your CRM has complete activity data without reps typing notes). No more “if it’s not in Salesforce, it didn’t happen” – the AI puts it in Salesforce for you.
Scheduling and SDR tasks
A new breed of AI “SDR” (sales development rep) bots can handle prospecting steps end-to-end. For instance, one startup, 11x.ai, automates the SDR role – from initial outreach email to qualifying questions to actually booking a meeting on a rep’s calendar. It’s not just a chatbot handing off a lead; it tries to do the entire appointment-setting job. This unbundles the lead qualification function from the CRM and places it in an AI agent.
Live sales call support
AI copilots (like Microsoft’s Sales Copilot or others) join your sales calls (often via Zoom or Teams) and act like an AI assistant in the meeting. They can live-transcribe and even suggest answers or talk tracks to the rep. They can certainly summarize the call, extract action items, and draft follow-up emails instantly after. This replaces the need for separate notetaking and ensures nothing said in the call is lost. For example, Tools like Gong and Chorus pioneered AI “conversation intelligence” – recording sales calls and analyzing them for insights (which topics correlate with wins, or where a rep might improve their talk/listen ratio, etc.). Now that concept is being extended further with generative AI: not only analyzing but actively assisting during the call. It’s easy to see how this encroaches on CRM’s territory – a rep might rely on the copilot for deal guidance instead of checking a CRM playbook or report.
Knowledge management and proposals
Some AI solutions automatically pull in relevant content when you need it. For example, Naro is an AI that monitors your emails and will surface relevant company documentation to help answer a client’s question on the fly. Instead of you searching through the CRM’s content library or SharePoint, the AI just serves it up. Similarly, when generating proposals or custom decks, generative AI can use templates plus client-specific data to draft them quickly. These were tasks that often required fiddling within CRM or connected systems to gather info – now it’s often a single-command outcome (“Draft a proposal for ACME Corp including their recent usage stats and our ROI case” – and boom, document ready).
In essence, every chunk of the traditional CRM value chain is being peeled off and optimized by AI. As one industry observer noted, “The future of customer relationship management isn’t monolithic systems trying to be everything to everyone… What’s truly dying is the empty promise of the all-in-one CRM as the single source of truth – a promise rarely fulfilled despite massive investments.” Instead, we have a constellation of AI-powered micro-services or assistants, each very good at a slice of the process, and all talking to the core data store. This unbundling is both an opportunity and a challenge: it means you can pick the best tool for each job (better functionality than a one-size-fits-all), but it also means your tech stack can get fragmented (10+ tools where once you had one CRM). In fact, sales teams today use on average 10 different tools to close deals (from forecasting software to call analyzers to enablement platforms), and most reps feel overwhelmed by the number of apps. The irony here is that we unbundled in search of quality and ended up with complexity instead.
The likely endgame is an AI-driven platform that re-bundles the best of these specialized insights, but in a unified, user-friendly way. Many believe that agentic AI is the mechanism to achieve that: the AI can serve as the integration layer, pulling info from many narrow tools and presenting what you need when you need it. So, you interact with one AI assistant, rather than 10 dashboards. In a sense, AI could give us the best of both worlds – specialized intelligence under the hood, but a simple conversational interface as the front end.
Goodbye Dashboards, Hello “Do It For Me”
What does all this mean for businesses and software vendors? For one, the classic mode of using a CRM – with employees manually updating records and managers poring over visual dashboards – may become less important. AI agents don’t need pretty dashboards to understand the data. They consume raw data and provide answers or either suggest next best actions or can take actions themselves. The typical executive today might look at a CRM dashboard of KPIs and then decide on a course of action. Tomorrow, an AI could just cut to the chase: “Out of 1000 leads, these 5 are hot, I’ve contacted them with tailored pitches, and scheduled 3 calls for next week. Also, I predict we’re 80% likely to hit this quarter’s number, and I’ve identified two deals at risk – and here’s how I’ve already addressed one of them.” No dashboard needed there – just outcomes and insights delivered directly.
Visualizations will still exist, but primarily for human comfort and strategy, not for the AI’s sake. Instead of static reports, you get real-time narrative insights. For example, rather than a static bar chart of quarterly sales, an AI assistant might verbally brief the sales VP each morning: “We’re pacing 5% behind target in the Midwest, mainly due to two stalled deals in manufacturing. I’ve reached out to those accounts for feedback and alerted their account owners.” It’s a different way of consuming information – more conversational and actionable. Business leaders might find they ask an AI agent questions (“How’s our pipeline health today? Where should we focus?”) and get a direct answer or even a proposed action plan, instead of clicking through multiple charts.
This also has pricing and value implications for software. Traditionally, enterprise software is often sold per seat or per license – for example, you pay for 100 CRM user licenses even if those users only leverage a fraction of the functionality. If AI agents are doing more of the work, the value metric could shift. Andreessen Horowitz suggests that the rise of AI-native software could be “the kiss of death for seat-based pricing”. Why? Because companies will pay for outcomes, not just for headcount using a tool. Imagine an AI that handles customer support tickets – if it can resolve 1,000 tickets, perhaps the vendor charges per ticket resolved (an outcome-based price) rather than a flat fee for the software or per user. They gave an example with Zendesk: a company might currently pay $115 per agent per month for support software, but if AI can handle the same work, the pricing might need to align to per ticket solved or some measure of value delivered. In other words, software pricing models will likely evolve to reflect AI’s contribution. We may see more usage-based or success-based pricing. Some AI sales tools could charge a percentage of pipeline generated or deals closed (akin to how some recruiting tools take a cut of hires). It’s a shift from “we sell you a tool” to “we help you achieve X outcome, and you pay us when that happens.” This will put pressure on traditional SaaS vendors who are accustomed to per-user subscriptions. They might have to justify their cost in terms of tangible results, which AI startups will happily do.
For the big CRM players like Salesforce, Microsoft, or HubSpot, agentic AI is both a threat and an opportunity. They have massive customer bases and rich data stores and are looking to infuse infuse AI deeply (Salesforce rolled out Einstein GPT and AgentForce across their clouds, HubSpot launched Breeze as an AI chatbot for CRM, etc.). But they also face nimble startups grabbing pieces of the pie. We’re already seeing a bit of an arms race: incumbents are rapidly adding AI features to keep customers from straying. Salesforce’s CEO has been evangelizing an “AI + Data + CRM” vision, essentially to ensure Salesforce remains the central hub that these new AI features plug into, rather than get displaced. Meanwhile, countless startups are popping up offering, say, “AI-driven sales co-pilot” or “autonomous email assistant” that work alongside your CRM or even in place of parts of it.
The CRM Metamorphosis
There are startups (often stealth or early-stage) trying to build an “AI-native CRM” from the ground up – essentially an always-on AI brain that a sales team interacts with, which logs data as a secondary byproduct. One such example described by A16Z is an AI that joins sales calls, auto-documents them, and updates a “customer page” that anyone in the company can read for a quick brief. The salesperson doesn’t have to write a thing; the AI does it. In this vision, the “CRM” is not a thing you painstakingly maintain – it’s more like a Wikipedia that writes itself based on AI observations. The value to the human users is huge (no more data entry, and you get a holistic view of the customer generated for you). The traditional CRM software in this case truly becomes a “dumb data store” in the back – important, but unseen. The AI and the user sit above it, interacting in natural language. When you consider that, you realize the CRM category is morphing. It might dissolve into either a back-end utility or evolve into something quite different.
Despite some hyperbole, this doesn’t mean “CRM” as a concept of managing customer relationships is going away – it means the tools and processes are changing drastically. One provocative headline put it as “CRM is dead – and AI can’t come to the rescue fast enough.” That was a frustrated take on how poor CRM experiences have become for users and customers alike (think of those maddening phone trees and support bots – that’s a company’s CRM strategy in action, failing the customer). The author’s point was that traditional CRM philosophy has lost its way, and we need AI’s help to make customer relationships feel human and responsive again. Ironically, the solution to a dehumanized customer experience may be better AI, not more humans – because at least AI can be made available 24/7 and eventually empathetic-sounding. We’ve all had experiences where companies make it nearly impossible to reach a human agent (all in the name of “efficiency” via their CRM system). A smart AI that actually helps the customer might be better than a clunky CRM portal that helps no one. As Bob Lewis, the author of that piece, half-jokingly declared: “I hereby declare CRM to be officially dead.” – implying it’s time for a new paradigm. That new paradigm is starting to take shape now with agentic AI.
So, will agentic AI mean the end of the CRM we know? Probably – yes. At least “CRM” in the sense of a big database that salespeople begrudgingly update, and managers use to extract pipeline metrics, could become obsolete. In its place, we’ll have AI-driven relationship management. It will feel fundamentally different: more proactive, more automated, possibly even prescient. The focus shifts from managing data to orchestrating meaningful actions. Dashboards and charts give way to real-time recommendations and autonomous task completion. Software 3.0 is less about monitoring and more about delegating.
Business leaders should prepare for this shift. It might start subtly – e.g., your teams rely on a few AI assistants plugged into your existing CRM. But over time, you may find you’re investing less in traditional CRM licenses and more in AI capabilities that sit on top of (or replace) those systems. The ROI calculations will hinge on efficiency and outcomes. If an AI can do the work of 5 junior sales reps or 10 support agents, that will change hiring plans and software budgets (and yes, raise new questions about training, oversight, and ethics). There will also be a human element to manage: change management will be key, because not every employee will be comfortable trusting an AI agent to do their work. And customers will have their comfort levels, too – one survey found 93% of U.S. consumers still prefer dealing with a human over an AI for complex interactions. So in the near term, the winning approach might be AI-human collaboration, rather than fully removing humans from the loop. Agentic AI can handle the drudgery and provide intelligence, while humans handle nuance, creativity, and relationship-building. Over time, as trust and AI capabilities grow, the “autonomy” of these agents can increase.
Parting Thoughts
We are at the dawn of a new software era. The information layer and action layer are converging, especially in domains like CRM which are rich in data and routine tasks. The promise is huge: software that doesn’t just sit there waiting for you to query it, but actually works for you, continuously and intelligently. Agentic AI is poised to become “Software 3.0”, the next platform shift after cloud. Just as cloud CRM (software 2.0) disrupted on-premise software (1.0), we can expect AI agents to disrupt cloud CRM in turn. The end of traditional CRM isn’t a doom-and-gloom scenario; it’s more like a chrysalis stage. The old CRM is transforming into something new – call it CXM (Customer Experience Management), AI-driven Relationship Management, or simply a smart co-pilot for your revenue teams. Organizations that embrace this shift early stand to gain a competitive edge, freeing their teams from low-value tasks and leveraging AI insights to deepen customer engagement. Those that cling to the old dashboards-and-data-entry model might find themselves outpaced by competitors who have AI working 24/7 to delight customers and optimize sales.
The Bottom Line
Agentic AI will likely render many traditional CRM practices and interfaces obsolete. But in their place, we’ll get systems that are more powerful and far more aligned with how humans actually want to work. It’s software that acts, not just reports. As we move into this new era, it’s wise to start reimagining your processes now. The writing is on the wall (or perhaps in the training data): adapt, and you won’t be saying “CRM is dead” with anxiety – you’ll be saying it with excitement, because something much better has taken its place. The best customer relationship tool might no longer be a piece of software at all, but an intelligent agent working by your side. And that future is coming faster than a lot of CRM vendors would like to admit.
The end of CRM as we know it is also the beginning of something far more powerful. Are you ready for it?
Sources
- Andreessen Horowitz – “Death of a Salesforce: Why AI will transform the next generation of sales tech” (July 2024)
- Paul Carroll, Insurance Thought Leadership – “The End of CRM As We Know It?” (Aug 2024)
- Salesforce (News) – “Sales Reps Need a Productivity Overhaul – Less than 30% of Time Spent Selling” (2023 research)
- Pendo/WRAL TechWire – Study: 80% of software features not used (Jan 2020)
- Insightly 2025 CRM Report (via PlanAdviser) – CRM underutilization findings (June 2025)
- Salesforce Blog – “Predictive Lead Scoring + AI is a Game Changer” (Dec 2023)
- Sales-mind.ai Blog – AI in pipeline management (40% of sales work can be automated) (2023)
- SuperAGI – “AI in Sales Forecasting: 20% better accuracy” (June 2025)
- Salesforce Ben – “Is CRM dying or evolving? AI transforming industry” (quotes: W. Flaiz “data graveyards”, B. Lewis “CRM dead”, autonomous CRM) (2024)
- ServiceNow Blog – “3 agentic AI takeaways” (importance of integrating AI with record/action systems) (2025)
- CIO.com – Bob Lewis column “CRM is dead – and AI can’t come to the rescue fast enough” (2025)
- Andreessen Horowitz – examples of AI sales startups (Clay, 11x, Naro, Day, People.ai) (2024)
- Reddit quote via SalesforceBen – AI agents and convergence of CRM/ITSM (2024)
