There's a pattern emerging in enterprise software right now that I'm calling "acquisition as confession." The AI shopping sprees at Salesforce, Workday, and other application cloud vendors are effectively admissions that their data foundations weren't ready for agentic AI. And the implications for how we position data platforms are enormous.
The market just handed us one of the clearest competitive positioning opportunities we've had in years. This piece is your PMM playbook — competitive framing, messaging guidance, evaluation questions for the field, and ready-to-use customer scenarios.
The Market Context (Your Competitive Briefing)
Here's what happened in the last six months that you need to internalize:
Salesforce made ten acquisitions since mid-2025 — headlined by Informatica at $8B — to patch Agentforce's data layer. Early adopters hit data quality problems, inconsistent agent behavior, and a pricing model that confused buyers. The acquisition list reads like a bug report: Informatica for data governance, Momentum for call transcription, Cimulate for shopper intent, Spindle AI for forecasting, Qualified for top-of-funnel engagement. They're assembling a data fabric from acquired parts — and hoping the metadata stitches hold.
Workday made five acquisitions in 19 months — Sana ($1.1B), Paradox, Flowise, Pipedream, HiredScore — from a company that averaged roughly one acquisition per year for two decades. Their CEO was fired in February 2026, stock down 40% in 12 months, and the co-founder returned. AI products contributed just 1.5 points to ARR growth last quarter.
The NBER study published in February 2026 found that 90% of firms reported no impact from AI on workplace productivity — even as Gartner projects $2.5 trillion in global AI spend this year.
The takeaway for positioning: The enterprise AI market is entering the Trough of Disillusionment, and the cause isn't bad AI models. It's bad data architecture. Competitors are spending billions to acquire what native data platforms already have.
The Narrative: Three Pillars for Every Conversation
When you're in a competitive deal, an analyst briefing, or a customer workshop, here's the framework. It comes down to three words: native, deterministic, composable.
Pillar 1: Native Data Fabric (vs. Assembled)
This is the core differentiator. A native unified data layer wasn't bolted together from acquisitions. It connects the financial ledger, supply chain, HR, CX, and planning data through a common semantic layer with built-in business context.
The competitive contrast: Salesforce is trying to build this by stitching together Informatica's data catalog with MuleSoft's integration layer, Data Cloud's harmonization, and a growing list of point acquisitions. Workday is connecting its HR and finance core to 3,000+ external apps through Pipedream while relying on Sana to become its "front door." Both are assembling their data fabrics mid-flight — while running agents on top of them.
"Our competitors launched their agent platforms and then went shopping for a data layer. We built the data layer first — because we knew from decades of running business-critical processes that the agent is only as good as the data it reasons over."
Pillar 2: Deterministic Grounding (vs. Probabilistic Inference)
This is the concept that makes technical buyers' ears perk up. When agents reason over a native data cloud, they're accessing system-of-record truth — actual inventory positions, actual margin data, actual financial close status, actual workforce capacity. This is deterministic data. It's not inferred. It's not probabilistically stitched from metadata across six acquired systems.
"There's a fundamental difference between an agent that says 'this account looks like it might churn based on engagement signals' and one that says 'consumption dropped 40%, renewal is in 90 days, the champion left — here's the play.' One guesses. The other knows."
Use this in demos. Walk the customer through a cross-functional agent workflow — say, a collections dispute that touches finance, customer service, and operations — and show how agents collaborate using shared business context. Then ask: "How would your current platform handle this if the data lived in six different acquired systems?"
Pillar 3: Composable Intelligence (the xP&A Connection)
This is where you connect the data story to the planning and analytics story. With seamless planning capabilities, your analytics platform can use the data layer as the persistence layer. That means the same trusted data that powers AI agents also powers planning models, simulations, and Monte Carlo analyses. The agent and the human decision-maker are looking at the same truth.
"Competitors sell you an agent OR a planning tool OR an analytics dashboard. Native platforms give you one data layer that powers all three — so the agent's recommendation, the planner's simulation, and the operator's execution all draw from the same source of truth."
Competitive Evaluation Guide
When a customer is evaluating AI platforms, steer the conversation toward these questions. They're designed to expose the assembled-vs.-native gap without needing to trash a competitor by name:
Question 1: "How many acquisitions were required to build your data layer?"
This surfaces the Frankenstein problem. If the answer is more than zero, ask the follow-up: "How long until those are fully integrated into a single semantic layer?" (Spoiler: years.)
Question 2: "When your agent accesses cross-functional data — say, finance and supply chain in the same query — is it reading from a single data model or federating across systems?"
This is the deterministic vs. probabilistic distinction in buyer language.
Question 3: "Can your planning and analytics tools access the same data layer as your AI agents — without an additional integration?"
This is the xP&A differentiator. If the answer involves "we partner with..." or "you'd need to connect..." — that's a gap.
Question 4: "What's your agent governance story? Who controls what the agent can see, do, and decide?"
Integration with existing role-based access controls — plus centralized visibility for agent management — is a mature answer here. Competitors are still building this.
Messaging Do's and Don'ts
✅ Do
- Lead with the data story, not the AI story. Everyone has agents. Not everyone has a native data fabric.
- Use the "acquisition as confession" frame to contextualize competitor moves — it's factual, not inflammatory.
- Anchor on the data platform as the differentiator. The agent is the experience layer. The data layer is the moat.
- Connect everything to business outcomes: faster time-to-insight, fewer hallucinations, agents that can actually act.
- Emphasize interoperability — protocol support, ecosystem integrations, and partnerships show openness, not walled-garden.
❌ Don't
- Don't lead with "our AI is better." Model layer is commoditized. This is not a winnable argument.
- Don't trash competitors directly. Let the data do the talking — acquisition counts, analyst quotes, and the NBER stat are your ammunition.
- Don't ignore the legitimate things competitors do well. Acknowledge and reframe: "They're great at X — but the agentic AI game is won in the data layer."
- Don't forget to mention the Knowledge Graph. It's the secret weapon for semantic context and prevents hallucinations.
Scenarios for Field Enablement
Here are three customer-facing scenarios to build into your next pitch, workshop, or demo:
Scenario 1: The Collections Dispute (Finance + Service + Operations)
A collections agent detects an overdue payment. Instead of just flagging it, the agent accesses the data cloud to check the customer's full order history, verifies delivery status from supply chain data, pulls the contract terms, and cross-references the dispute with service tickets. It synthesizes a resolution recommendation — grounded in deterministic data from across the business — and routes it to the right team.
Try doing that with six acquired systems and a metadata stitching layer.
Scenario 2: The Demand Signal (Supply Chain + Planning + CX)
A spike in online engagement triggers an agent to query for current inventory positions, check production capacity, pull the demand plan from analytics, and recommend a reallocation — all within one semantic layer. The planner reviews the recommendation using the same data the agent used, runs a Monte Carlo simulation, and approves the change.
This is the xP&A story made tangible.
Scenario 3: The Workforce Rebalance (HR + Finance + Operations)
A People Intelligence Agent detects elevated attrition in a critical skill group. It pulls retention data from HR, checks budget headroom from finance, models the cost of backfill vs. retention bonuses in analytics, and surfaces a recommendation to the CHRO.
Competitors can tell you who's leaving. Native platforms can tell you what it costs and what to do about it.
The Bottom Line
The market is telling a very clear story right now: the AI model layer is commoditized, the agent UX layer is table stakes, and the data layer is the only durable moat. Competitors just validated this thesis by spending billions to acquire what native data platforms already have.
Your job as a PMM is to make sure every customer, analyst, and field rep understands three things:
- The data platform is the moat. Not the agent. The agent is the experience. The data layer is the competitive advantage.
- Native beats assembled. Every time. In every evaluation. Frame it that way.
- Deterministic beats probabilistic. When agents reason over real business data — not inferred metadata from acquired parts — the outcomes are fundamentally different. Better, faster, and trustworthy enough to actually act on.
The hype cycle is compressing. The trough is here. And the vendors who started with the data layer intact are the ones who'll emerge first.
Let's make sure the market knows it.