Intapp today launched Celeste, a set of artificial intelligence (AI) agents that automate a wider range of tasks for professional and financial services firms using models from Microsoft and now Anthropic via an alliance revealed earlier this week.
Unfurled at an Intapp Amplify event in New York, Celeste, currently available in limited access, is trained to execute a series of customizable playbooks that firms can reuse or build themselves.
For example, end users will be able to compare new opportunities to past opportunities a firm may have pursued or passed on that can be directly accessed via the AI Copilot tool for Microsoft 365.
Designed specifically for firms that have adopted the Intapp platform to manage workflows, Celeste is grounded in the proprietary data that a firm, via a context engine built into Celeste, exposes to a generative AI model, says Thad Jampol, chief product officer for Intapp.
Intapp has also built a Model Context Protocol (MCP) server to make data available to third-party AI agents and applications, noted Jampol.
Additionally, Intapp has built in compliance and policy enforcement capabilities at the architecture level, he added. At the core of that capability is what’s been dubbed an AI Wall, developed to ensure confidential information is only seen by those that have the right access. “Nothing bleeds,” says Jampol.
There is little doubt that professional services firms are already using AI to automate a wide range of tasks, but Intapp is making a case to now embed AI deeper into operations.
Ultimately, each services firm will need to decide to what degree to invoke AI via a platform versus building their own agentic AI workflows. In effect, a platform-based approach may enable service providers to operationalize AI faster in a way that is still customizable. Less clear, of course, is the total cost of invoking AI agents via a platform versus building and maintaining a set of custom agents.
Regardless of approach, the output generated by an AI model needs to be reviewed by a human, especially if it is being used to deliver a professional service that might be audited. In fact, services firms should assume that auditors will also soon be using AI to surface potential violations of multiple compliance mandates.
In the meantime, professional and financial service firms should be identifying which existing manual workflows can be automated using AI agents. It’s not likely AI agents are likely to eliminate the need for existing staff but it does create an opportunity to augment personnel in a way that enables professional services firms to take on more projects without necessarily having to increase their payroll.
Hopefully, AI will lead to greater customer satisfaction, especially if it can reduce the number of billable hours that service providers need to complete a task on their behalf. However, it’s also probable those savings will not be immediately passed along until those clients start demanding to know exactly how much human expertise was required to complete a task on their behalf.
