Six months ago, on an engagement with a global, multi-product business, I made a bet.
The client is the kind of company built through acquisition: global, multi-product, multi-business-unit, serving customers across a complex value chain of designers, retailers, manufacturers, and contractors. It is also a business where a single bundled deal can touch more than one CRM.
The conventional play in that situation is to scope a multi-quarter integration project. Hire engineering. Bring in a consulting firm. Wait three quarters for something to ship.
I recommended a different play. Bring Zapier’s Enterprise plan into the stack, put it in the hands of operators, and see what they would build.
Six months later, the model has settled. The integration work that used to wait in someone else’s queue is moving at the operator layer. AI sits inside the workflows, not bolted on the side. The build team is RevOps, Sales Ops, and Management Information Systems. Very few of them carry an engineering title.
The two arguments most operators have heard
In my experience, a decade building GTM systems across fast-scaling, acquisition-built companies, including leading GTM Systems at Multiverse, post-close systems integration tends to come down to two arguments.
The first is standardization. Adopt the buyer’s tech stack. Migrate the acquired company’s processes, definitions, and data model onto it. The case for this approach is a clean unified architecture, predictable operating cost, and a single source of truth.
The second is plug-and-run. Keep the acquired company’s stack and team. Integrate at the holding-company layer. The case for this is speed of close, retained acquisition economics, and preserved autonomy.
The third path I’ve been building is bottom-up. The operators closest to the work design the integration strategy, anchored to their constraints.
Narrow standards, broad freedom
The third path isn’t no-standards. It’s narrow standards.
Two things the client standardizes at the architecture layer. Definitions of revenue and customer status. A live master customer record that syncs across every system.
Beyond those two, the operating units plug and play. CRM consolidation is staged by operational impact. For the cross-system handoffs that span CRMs, we use Zapier. Lead routing across CRMs. Opportunity passing on bundled deals. Cross-sell visibility between business units.
The build the team is proudest of
One specific build illustrates the model.
A cross-CRM workflow that lets reps work bundled deals spanning business units without losing visibility, attribution, or compensation accuracy. The work that previously required manual reconciliation across systems now stays in sync.
The build pulled in RevOps, Finance, and Product. It was built by operators. A bundled deal used to move through four hands before it cleared. The rep stitching it across systems, a second pair of eyes verifying the pass, finance checking deal structure, and commissions checking payout. The workflow collapsed that chain. Roughly 230 hours a year that used to live in reconciliation now don’t.
AI sits inside the program
The forward-thinking edge is what AI does inside these workflows.
AI enriches the lead before a rep sees it. AI summarizes opportunity context so the rep walks into the call already briefed. AI surfaces cross-BU patterns the rep would otherwise have to find on their own. The rep still makes the call. The operator-built automation is the connective tissue. AI is what makes it intelligent.
Two shifts made the model defensible at enterprise scale. Agentic AI moved enrichment, summarization, and pattern recognition into the operator layer. Automation platforms designed for non-engineers gave that layer enterprise-grade governance. Together they made operator-led integration not just possible but the better choice.
That’s the bet we made when bringing Zapier’s Enterprise plan into the stack. Not as a replacement for the consolidation roadmap. As the connective layer that lets the consolidation roadmap proceed without disruption.
The constraints that bound the work
Two rules drove the design.
Do not disrupt the order-to-cash motion.
Do not let the consolidation roadmap deteriorate the customer experience.
Customers should not feel the CRM transition. Reps should not lose deals because the data ran ahead of the systems. Finance should not have to reconcile two parallel realities at quarter close. Anything that ships has to clear that bar.
Governance is what made it move fast
The instinct against operator-built automation is institutional flinch. The first time someone without an engineering title ships a workflow that touches the CRM, the finance system, and the support stack, engineering wants to know who reviewed it. Security wants to know what permissions it has. Compliance wants to know what data it touches.
The answer to all of those questions has to be ready before the build starts.
In Zapier Enterprise, anyone on the team can build a workflow in a sandbox. Only designated approvers can promote it to production. Build velocity stays at the operator layer. Accountability stays where it belongs.
Today, the client has over 45 builders across the org who have shipped more than 170 production workflows under that governance model.
The wider signal
The “AI strategy” conversation in fast-scaling companies tends to argue about which model. Which platform. Which fine-tune.
The practical question that compounds is how to remove the noise and connect systems efficiently in a way that compresses time to value. The answer points one layer upstream from the model debate, to who, in the org chart, is allowed to ship the integration work that touches the customer.
