Why Agentic AI’s Real Score Is Actions, Not Chatbots
The enterprise AI race has shifted from chatbot counts to measurable actions. Here's what the latest funding rounds and data say about where CRM automation is heading.
Eighteen months. That’s all it took for Wonderful AI to go from founding to a $5 billion valuation on the back of a $550 million Series C led by Insight Partners, with Salesforce joining the round. That kind of trajectory doesn’t happen in a market that’s still experimenting. It happens when buyers are spending real money on a real problem – and right now, that problem is making AI actually do things rather than just answer questions.
The broader enterprise software backdrop reinforces this. Snowflake posted a significant earnings beat in its most recent quarter, following strong results from both Salesforce and Workday – three major platforms signalling that AI-driven spending hasn’t cooled. For RevOps and GTM leaders, the signal is clear: the budget is there, and the expectations attached to it are rising fast.
From Chatbot Count to Action-to-Output Ratio
The Salesforce Agentic Enterprise Index, now in its second edition, introduced a metric that’s more useful than anything the industry has been measuring for the past two years: the action-to-output ratio. The idea is straightforward. Don’t ask how many AI agents a company has deployed – ask how many real-world actions those agents completed per unit of output. That shift matters enormously for sales and CRM teams who’ve been sold on volume and left wondering why their sales pipeline looks the same as it did before the AI rollout.
The chatbot era rewarded announcement. The agentic era rewards execution. A team running five tightly scoped agents that autonomously qualify leads, update CRM records, trigger follow-up sequences, and flag at-risk deals is operating at a different level than a team that deployed fifty bots to answer FAQs. The metric isn’t about sophistication for its own sake – it’s about whether AI is compressing your sales cycle or just decorating it.
The Salesforce Agentic Enterprise Index, released in its second edition in August 2026, found that enterprise AI conversations have shifted decisively away from chatbot volume toward the action-to-output ratio as the defining measure of agentic maturity.
If your team is still measuring AI success by the number of tools adopted, that’s the wrong scorecard. Start tracking actions per opportunity – specifically, how many steps in your workflow an agent owns end-to-end without human intervention.
What Adobe’s Rilo Acquisition Says About the Marketing Automation Gap
Adobe’s acquisition of Rilo, a PeakXV-backed AI startup, tells a specific story. The company was explicitly building AI for Fortune 500 marketers, and Adobe’s decision to absorb it rather than build internally suggests the gap between what enterprise marketing teams need and what existing platforms offer is still wide enough to justify acquisition. That’s a meaningful signal for go-to-market teams evaluating their marketing automation stack right now.
The acquisition expands Adobe’s position in AI-powered campaign execution – the kind of automation that doesn’t just personalise a subject line but makes downstream decisions about audience segmentation, content sequencing, and timing based on real behavioural signals. For B2B teams in particular, this directly affects customer acquisition cost. When a well-tuned AI layer reduces the manual work between lead capture and first meaningful touchpoint, cost per acquisition drops without a proportional cut in quality.
Worth watching: whether Adobe integrates Rilo’s capabilities into Experience Cloud in a way that makes it accessible to mid-market teams, or whether it stays a feature set for enterprise contracts. That distinction will matter a lot to buyers who don’t have a dedicated MarTech team to manage the implementation.
Wonderful AI and the Venture Signal for CRM Buyers
Wonderful’s funding story is worth reading carefully. The company raised $150 million at a $2 billion valuation in March 2026, then closed $550 million at $5 billion just six months later – a 2.5x valuation step-up in half a year. Insight Partners led both rounds. Salesforce’s participation in the Series C is the part that should catch CRM professionals’ attention: it’s a strategic bet, not just a financial one.
When a platform like Salesforce puts capital into an AI automation company, it’s either building toward integration or signalling where it sees the market moving. Either way, buyers evaluating their CRM tools should factor in how well their current stack connects to the emerging layer of AI agents that sit on top of core systems of record. The CRM itself is increasingly the data foundation. Agents are becoming the execution layer.
For teams thinking about Net Revenue Retention, this has practical implications. AI agents that proactively identify expansion signals or flag churn risk based on product usage data, support ticket patterns, and engagement trends don’t just protect revenue – they give customer success teams a defensible prioritisation framework. That’s the kind of outcome that shows up in NRR, not just in demo videos.
What CRM and GTM Teams Should Do Differently Right Now
The funding rounds, the earnings beats, and the index data all point in the same direction. Agentic AI isn’t coming – it’s already being bought and deployed at scale by companies that have moved past the pilot phase. For most GTM teams, the question isn’t whether to adopt it but how to structure adoption so it produces the action-to-output ratio that justifies the spend.
A few concrete steps worth taking before your next planning cycle:
- Audit your current AI deployments against actual completed actions, not logins or usage sessions. If an agent isn’t closing a loop autonomously, it’s still a chatbot by another name.
- Identify the two or three highest-friction handoffs in your pipeline – lead routing, opportunity updates, renewal outreach – and scope those as your first agentic automation targets.
- Map your Ideal Customer Profile criteria into your AI qualification layer explicitly. Agents that operate on vague intent signals produce vague results.
- Check whether your CRM vendor has a published integration roadmap for third-party AI agents. If they don’t, ask directly. The answer tells you a lot about where your stack is heading.
The teams pulling ahead aren’t the ones who launched the most AI projects in 2025. They’re the ones who picked fewer, scoped them tightly, and measured ruthlessly. That’s still the playbook. For deeper context on building AI into your sales process, the CRM Guides section has practical frameworks worth reviewing – and if you want to stay current as this space moves fast, the CRM Daily Newsletter covers developments like these as they happen.
