3/23/2026 | 6 Minute Read
Topics:
There is a lot of noise right now around AI in IT, and most of it skips an uncomfortable truth: AI does not fix broken systems.
That was one of the clearest themes to come out of Spark 2026, our annual sales kickoff event at ConnectWise, and it kept coming up throughout the session. During our final session, Manny Rivelo, CEO, Lee Silverstone, SVP of Product AI Platform, Joe Mercurio, CRO, and I spent less time talking about what AI might do someday and more time talking about what actually makes it work.
This blog reflects that conversation and the thinking behind it.
Spark 2026 was an internal event, but what we talked about matters to managed service providers (MSPs) because it shapes how we build and prioritize.
The message internally was simple: IT solution providers are hitting limits. Ticket volume keeps climbing, while teams do not grow at the same rate. Automation exists, but it often adds complexity instead of removing it.
When that happens, the problem is rarely effort. It usually comes back to structure.
Most IT solution providers have plenty of data, from tickets and time entries to billing records, monitoring alerts, and security signals.
Volume is not the issue; rather, it’s how that data is structured, centralized, and accessible.
When those systems were built as standalone, they never really "cooperated” with each other in practice. That makes it hard for people to make decisions, and it makes it even harder for automation or AI to do anything useful.
You see this when teams spend hours reconciling reports, correlating data, or second-guessing recommendations. The system ends up asking humans to do the hard work because it cannot reason on its own.
We did not acquire zofiQ because we wanted another AI story. We acquired it because agentic systems only work in environments that were designed for them, environments that put common data at their core. Environments such as the ConnectWise Platform.
Lee Silverstone, co-founder of zofiQ, put it plainly during the session, based on his success of building agents inside real PSA environments from many different providers.
“ConnectWise has the properly structured data to train agents on the high levels that nobody else has. ConnectWise is the best business for that and has the best data for that. And that's how we win.”
That statement matters because Lee has seen firsthand where agents break down when data is inconsistent or shallow.
Agents do not learn from screenshots or dashboards. They learn from patterns over time. That requires a single source of truth, common extensibility, context, history, and systems that were designed to share data and services from the start.
Many platforms were assembled over time from different tools, schemas, different data repositories and different business logic. AI added later has to work around that.
ConnectWise products share data models across every service and application, such as finance, CRM, and IT operations. That was not originally done for AI. It was done to make the business run better. That design choice now becomes a real advantage and, frankly, a prerequisite for getting an ROI from AI.
zofiQ helps turn that foundation into action, without adding more charts or alerts.
It shortens the distance between what the system knows and what someone actually needs to do next.
That shows up as:
The goal is not to replace people. It is to stop wasting their time.
AI is not a checkbox or something you bolt on after the fact.
At Spark 2026, we were clear about this. AI should reduce cognitive load. It should live inside workflows people already trust. It should handle routine decisions so humans can focus on exceptions.
If AI creates more cleanup work, something in the system is failing.
A lot of vendors are layering AI onto disconnected systems. That approach often looks good in demos, but then stalls once it hits real operations because the underlying systems are not integrated, the data is fragmented, and the AI lacks the context needed to reliably perform at scale.
We focused on structure first, because intelligence only follows if the foundation holds.
Spark 2026 was about reigniting that spark in our teams around that reality. The zofiQ acquisition strengthens a foundation that already exists rather than trying to replace it.
Spark 2026 was about being honest about that and aligning execution around it. We believe ConnectWise is in a strong position because the foundation was built years ago, and zofiQ helps us use it better.
If automation depends on manual cleanup, the limitation is not AI. It is the structure underneath it.
Share:
Spark 2026 is the ConnectWise sales kickoff, renamed to reflect urgency and focus.
ConnectWise leadership: Manny Rivelo, CEO, Joe Mercurio, CRO, Lee Silverstone, SVP of Product AI Platform, and David Raissipour, CPTO.
Because agentic systems only deliver value when the data foundation already exists.
ConnectWise products share unified data models across service, finance, and operations. This depth and consistency allow AI agents to learn from real workflows, not isolated events.
AI features depend on patterns over time. Without consistent data and context, features generate noise, require manual cleanup, and fail to deliver real ROI.
Structured data means core operational data is consistently modeled, historically complete, and shared across systems so humans and AI can interpret it the same way. Service, time, billing, and operations data were designed to work together.
IT solution providers feel it when automation reduces noise, prioritization becomes clearer, and teams spend less time reconciling systems and more time solving exceptions.
