Do Leaders Really Want Five Different AI Assistants?
Every software company seems to be making the same announcement. "We've added AI."
Accounting platforms are launching AI assistants, CRM providers are introducing conversational interfaces, Project management systems are rolling out AI-powered insights, and Professional services automation platforms are embedding AI throughout their workflows.
In many cases, these announcements are accompanied by a familiar message: new capabilities, the promise of increased productivity, and a corresponding increase in subscription fees. On the surface, this makes perfect sense since AI has become the most important technology shift in a generation, and software vendors are understandably racing to incorporate it into their products.
But as I watch these announcements continue to roll out, I find myself wondering whether we're solving the right problem. The question isn't whether these business critical tools incorporate AI, but whether executives, senior managers, and operators actually want to use a different AI assistant for every application they own.
In a typical professional services firm, their accounting data lives in QuickBooks or NetSuite, their pipeline lives in Salesforce or HubSpot, resource planning happens in a PSA platform, and employee information sits inside an HR system - not to mention that forecasts and budgets, operating plans, and goals often still live in spreadsheets and Google docs.
If each one of those platforms introduces its own AI assistant, the accounting assistant can answer questions about invoices and financial statements, the CRM assistant can answer questions about deal velocity and win-rates, the PSA assistant can answer questions about utilizations, and the HR assistant can answer questions about team experience and skills. Individually, the answers to these questions are certainly useful, but using AI is not much of an advancement beyond the existing reporting tools that were already part of the platforms.
The real issue is that a CEO, for example, is rarely asking questions such as: "What was revenue last month?", instead, they're asking questions like:
Can we afford to hire ahead of demand?
Are we staffed appropriately for the opportunities expected to close next quarter?
What happens if our largest opportunity slips by sixty days?
Can we take on this new project without impacting existing client delivery?
Which actions would have the greatest impact on EBITDA this year?
and the answers to these questions don’t live in one of these systems, they exist across all of systems and require the unification of all of the data within them. This is where the current generation of application-centric AI begins to show its limitations, and most AI assistants are only as intelligent as the application they sit within. The limitations become quickly apparent:
A CRM assistant understands pipeline, but often has no visibility into staffing constraints.
An accounting assistant understands financial performance, but typically has no understanding of future sales activity.
A PSA assistant may understand utilization and project delivery, but not necessarily how future bookings impact hiring decisions.
Each assistant has access to a piece of the puzzle, but none of them can see the entire picture. As large language models rapidly become commodities, the competitive advantage will increasingly come from how well organizations connect their data and provide AI with the business context needed to understand how they operate.
As we have discussed in other blog posts, every services business we work with operates differently. The general concepts of utilization, pipeline management, and resourcing are similar, but their own definition is nuanced and specific to their organization and way of working. The relationships between these metrics vary from company to company as well. Two firms may have identical data and arrive at completely different conclusions because their operating models vary.
This is precisely why simply pointing an AI model at a database or embedding a chatbot into an application rarely produces meaningful business intelligence. The model may understand the data contained within that system, but it does not understand the business itself.
What AI ultimately needs is context. It needs to understand how pipeline converts into bookings, how bookings translate into project work, how project work drives capacity requirements, and how capacity influences revenue, profitability, and ultimately enterprise value. It needs to understand the operational model behind the numbers, not just the numbers themselves.
This is where we believe the industry is headed. The current wave of application-specific AI assistants represents an important step forward. They make information more accessible and allow users to interact with software in more natural ways. But they are still fundamentally tied to the applications in which they live.
Rather than creating intelligence within every individual application, organizations will increasingly need a unified intelligence layer that can understand information across all of them.
That belief has shaped how we've built Executive Companion.
Instead of treating AI as a feature within a single system, we've focused on creating an operating intelligence layer that sits across the systems already running the business. Executive Companion connects data from CRM, PSA, finance, HR, and other operational systems, combines that information with the operating logic that defines how the business works, and makes that context available wherever teams choose to work.
That's one of the reasons we've embraced Model Context Protocol (MCP). We don't believe organizations should have to choose a single AI interface. Whether leaders prefer ChatGPT, Claude, Microsoft Copilot, or the next generation of AI tools, they should be able to access the same trusted business context. The interface may change over time, but the context shouldn't.
It’s with this approach that we can help leaders answer the questions that matter most. Not what happened last month, but what is likely to happen next. Even more exciting, it creates the opportunity to move beyond reporting and into recommendations. If gross margins are eroding, most reporting tools can tell you where the problem exists. An intelligence layer that understands the relationships across the operating model can help identify which operational intervention is most likely to have the biggest impact going forward.
This is where we believe the greatest long-term value of AI will emerge. Not from making reports easier to access, but from helping leaders better understand the consequences of their decisions and the actions most likely to improve business performance.
The companies that create the most value from AI over the next decade will not necessarily be the ones that adopt the most AI features. They will be the ones that successfully connect their data, define their operating model, and provide AI with the context required to understand how their business actually functions. Because executives don't need five different AI assistants, they need one intelligence layer that understands the entire business.