Program report · filed November 15, 2026
AI in SalesMeasured report
Enterprises Flip AI Spend to 80 Percent Standardized Core Workflows
Cisco and ConnectWise leaders at Gong Celebrate '26 detail a build-or-buy framework as enterprise AI spend flips to roughly 80 percent standardized core workflows, not individual experiments.
By Rebecca Stone4 min read717 words
Program notes
- Roughly 80 percent of enterprise AI effort has shifted from individual experimentation to standardized core workflows.
- ConnectWise built an internal agent, Account Brain, pulling from five third-party platforms including Gong across a 35-plus product portfolio.
- Cisco reserves in-house builds for proprietary data and business-unique use cases, Oscar Canon said.
- The framework was presented on a Gong Celebrate '26 panel with Cisco and ConnectWise leaders.

Enterprise AI effort has flipped from roughly 80 percent individual experimentation to about 80 percent standardized core workflows and best practices, with the remainder left for creativity at the edges. That shift, described at Gong Celebrate '26 by panelists from Cisco and ConnectWise, frames a decision framework for the build-or-buy question every revenue and technology leader now faces. The numbers are directional, not a formula, but the change is real.
The panel, titled "Decide: Build, buy, or both," featured Oscar Canon, AI COE Leader at Cisco, and Austin Reichart, Sales Enablement Manager at ConnectWise.
How did enterprise AI adoption unfold?
The past year ran through three phases. In phase one, capabilities exploded and anyone could create an agent. Boards called for bold moves before anyone defined a plan or an objective. Within about six months, phase two arrived: fragmentation. Small groups spun up hundreds of their own agents, the same work got done a dozen ways, and managers could not tell whether people followed coached behaviors or agreed best practices.
Phase three, where most enterprises now sit, is intentional architecture: one operating rhythm built around business outcomes, designed from the start with leading indicators tied to financial return. Efficiency still matters, but effectiveness matters more.
Why define the problem before the tech?
Pressure to keep up pushes teams to buy or build before defining the problem. "You truly need to understand the work," Oscar Canon said — the people, the workflows, and the specific tasks inside them. He also advised an honest inventory: "Before you start building a duplicate solution, truly understand what you have." The cheapest capability you will ever add is the one you already pay for.
ConnectWise found exactly that. Its assessment showed reps pasting Gong transcripts into general-purpose AI assistants to craft next best actions — something Gong's AI does automatically. "It was an enablement challenge," Austin Reichart said. "We had the capability, so we started there."
What does the knowledge layer add?
Every AI architecture rests on data, but companies underestimate the layer above it: the knowledge graph. It maps how information relates, what matters to the business, and the rules around governance, privacy, and security. Raw data gives a seller a summary and action items. A knowledge layer can tell the seller what to have said, which behaviors were missed, how the conversation connects to the buyer's decision process, and whether key influencers were looped in. Data tells you what happened; the knowledge layer tells you what it means and what to do next.
How should leaders sort the work?
Every job fits one of three buckets:
- General-purpose AI platforms that reason across many sources
- Purpose-built platforms with specialized domain depth, embedded in workflows
- In-house builds for what no vendor can provide — Cisco reserves building for proprietary data and business-unique use cases
These are not competing choices, and open standards like MCP ease the connections. Without integration, sellers face five or ten disconnected tools and a harder job.
ConnectWise shows a well-justified build. With more than 35 products and reps working across five interfaces, the team built an internal agent called Account Brain. It pulls from five third-party platforms, including Gong, so reps see internal data, recent conversations, product usage, and customer sentiment from Reddit and G2 in one place before a meeting.
What does building truly cost?
The build is the cheap part. Costs keep coming after launch: engineers, maintenance, new releases, the full lifespan of the solution. "If the system goes down, the buck stops with you," Reichart said. Canon also weighs speed to value, because every month of delay is a missed opportunity. ConnectWise ran a low-code pilot in a controlled environment, learned where costs ran high, then optimized.
Finally, partner rather than just purchase. Reichart tells vendors what his team plans to build; the answer is often to wait two or three months because it is on the roadmap. Canon advises against competing with vendors on their core strengths.
Canon summed up the framework in one line: buy the capability, build the differentiation, and integrate it into the flow of work. As capabilities keep changing, leaders will make these calls again — take stock of what you own, standardize the core, and build only what makes you different.
via Gong Labs Blog (Source)
More from Rebecca Stone
Show full bio
Market editor covering people, launches and day-to-day practice in Sales technique and methodology, sales enablement at Quota Brief.
139 articles