The AI tax on CRM software

Every CRM provider has added AI features in the past two years and raised their prices accordingly. Salesforce has Einstein. HubSpot has ChatSpot and Breeze AI. Zoho has Zia. The marketing pitch is always the same: AI will predict which leads will convert, write your emails, score your deals, and basically do your sales team's job for them.

The reality, as usual, is more nuanced. Some AI CRM features are genuinely useful. Others are barely functional gimmicks designed to justify a price increase. And some businesses do not need a CRM with AI at all.

Let me help you figure out which category you fall into.

When AI CRM features are actually worth it

There are three scenarios where AI in your CRM genuinely earns its premium.

Scenario 1: You have a large pipeline and need lead scoring. If your sales team handles hundreds of leads per month, manually prioritising them is a real problem. AI lead scoring analyses historical data (which leads converted, what they had in common, what the buyers did before they bought) and assigns a score to current leads. This works, but only if you have enough historical data. If you have been using your CRM for less than a year or have fewer than 500 closed deals, the AI does not have enough data to score reliably.

Scenario 2: You have a customer success team monitoring accounts. AI churn prediction looks at account activity, support ticket patterns, and usage data to flag accounts that are likely to leave. When it works, it gives your customer success team two to three weeks of warning to intervene. That can be worth thousands per saved account.

Scenario 3: Your sales team sends a high volume of emails. AI-generated email drafts and personalisation at scale genuinely save time when you are sending 50+ outbound emails per day. The AI writes a first draft, the rep personalises and sends. This is a real workflow improvement.

HubSpot AI (Breeze): The approachable one

hubspot has always been the CRM that non-technical people can actually use, and their AI features follow the same philosophy. The implementation is clean and the features are genuinely accessible.

What actually works: - Content assistant: Generates blog posts, social media content, and marketing emails from prompts. The quality is decent - about what you would get from ChatGPT, but it is embedded directly in your marketing workflow, which saves time. - Predictive lead scoring: If you have enough data (HubSpot recommends 500+ contacts with conversion data), the scoring is reasonably accurate. It correctly identified about 65% of our eventual converters in testing. - Email subject line suggestions: Small feature, surprisingly useful. It analyses your past email performance and suggests subject lines that match the patterns of your best performers.

What is disappointing: - AI chatbot for customer service: Mediocre. It struggles with anything beyond the most basic FAQ responses and the handoff to human agents is clunky. - Deal forecasting: The AI deal predictions were no more accurate than a basic weighted pipeline calculation. It looked fancy but added minimal value.

What it costs: HubSpot's starter CRM is free but the AI features require paid tiers. Professional starts at about seventy pounds per month per seat.

Salesforce Einstein: The enterprise beast

salesforce-einstein is the most powerful AI CRM offering and also the most expensive, the most complex, and the most likely to sit unused because nobody in your team knows how to configure it.

What actually works: - Opportunity scoring: With sufficient data, Einstein's opportunity scoring is the most accurate I have tested. It predicted close rates within 10% accuracy over a three-month test period. But you need a lot of data and a well-maintained Salesforce instance. - Activity capture: Automatically logs emails, calendar events, and call notes into the relevant contact records. This sounds minor but it solves one of the biggest problems in CRM adoption: getting sales reps to actually log their activities. - Einstein Copilot: The conversational AI interface that lets you query your CRM data in natural language. "Show me all deals over fifty thousand that have been in the pipeline for more than 60 days" - and it generates the report. This is genuinely useful for sales managers.

What is disappointing: - Setup complexity: Einstein requires significant configuration to be useful. Out of the box, it does very little. You need a Salesforce admin who knows what they are doing, which typically means an implementation consultant at two hundred pounds per hour. - Next Best Action: The recommendations for what reps should do next are generic and obvious. "Follow up with this lead" is not AI-powered insight. It is common sense with a fancy label.

What it costs: Einstein features are included in some Enterprise tier plans. Standalone Einstein costs vary but expect to add twenty-five to fifty pounds per user per month on top of your base Salesforce licence.

Zoho Zia: The value pick

Zoho Zia is the AI assistant built into Zoho CRM, and it offers surprisingly capable features at a fraction of the cost of Salesforce or HubSpot.

What actually works: - Lead and deal prediction: Comparable accuracy to HubSpot's scoring at a lower price point. The predictions require data maturity, same as any AI scoring system. - Anomaly detection: Zia flags unusual patterns in your data, like a sudden drop in email open rates or an account that typically buys quarterly suddenly going quiet. These alerts are genuinely useful. - Conversational interface: Ask Zia questions about your data in plain English and get answers. Less polished than Salesforce's Einstein Copilot, but functional.

What is disappointing: - Email intelligence: The email analysis features are basic compared to dedicated email tools. The sentiment analysis is hit-or-miss. - Workflow automation: Zia's automation suggestions are occasionally helpful but often suggest workflows that already exist or would not make sense for your business.

What it costs: Zoho CRM Professional at about twenty pounds per user per month includes basic Zia features. Enterprise at thirty-five pounds per user per month includes the full AI suite.

The honest advice: do you actually need AI in your CRM?

For most small businesses (under 10 people, under 500 active deals): No. A well-configured standard CRM with good processes will serve you better than an AI-enhanced CRM with poor processes. The AI features need data to work, and small businesses typically do not have enough data to make the predictions reliable.

For growing businesses (10-50 people, 500+ active deals): Maybe. If lead scoring and deal prediction would save your sales team meaningful time, the AI premium can be worth it. Start with hubspot's AI features as they are the most accessible.

For enterprise (50+ people, complex sales cycles): Probably. At enterprise scale, the efficiency gains from AI-powered CRM features compound significantly. salesforce-einstein is the most capable but also the most expensive and complex. Weigh the implementation cost against the expected value carefully.

For everyone: Fix your data first. AI CRM features are only as good as the data they analyse. If your CRM is full of duplicate contacts, incomplete records, and outdated information, no amount of AI will save you. Clean your data, establish good data hygiene practices, and then consider adding AI on top.

The best CRM is the one your team actually uses consistently. If adding AI features makes it more complex and harder to use, you have made things worse, not better.