I caught a recent briefing from PSA vendor Kantata regarding their new AI powered Expertise Engine. The Expertise Engine has four new applications that complement other new innovations from Kantata. But these new applications/capabilities aren't like traditional software applications. Each of these is designed around especially difficult professional services problems (e.g., the ability to know how one project's staffing issues affect all other projects in the service firm's portfolio). Kantata's approach to developing these new capabilities was to imagine how, in one Kantata executive's words, "Would PSA look like around this use case if AI were a partner, not just an assistant?" In other words, how can Kantata transform a process/workflow and not just add an incremental bit of new capability?
Service firms sell and deliver work quite differently than product-based companies. Consultancies (among other service sub-markets) sell the time and expertise of their professional staff to clients. These people-based projects are often tough to scale as:
My career, like that of thousands of other consultants/integrators/architects/etc., faced a number of challenges in selling and delivering client work. We needed:
But this information is not commonly available in many service firms as the key people who are most familiar with these kinds of projects are often unavailable or unwilling to contribute, synthesize, edit and/or create these materials. Those individuals are in high demand and often lack the time to work on such an 'internal' (i.e., not client facing) effort. Creating great intellectual property (IP) for others to use or benefit from is often seen as less valuable to advancing one's career. Chargeable work is more visible and drives promotions more than creating/editing/refining IP.
Those points resonate with me as I ran a couple of groups at a prior employer doing just those things. Prior to AI, the amount of time needed to collect project working papers, sifting through them (like a gold miner panning for nuggets) for unique and/or high value additions, reconciling differing process steps or outputs, placing everything in a common template, etc. was enormous. And, of course, you have to deal with different languages, currencies, paper formats (A4, Legal, etc.), project tools, and much more.
Frankly, it's a wonder these firms are able to field proposals with a consistent look and feel when their source materials are all over the map.
Kantata recognized the professional services problems above and crafted new solutions to address these. This is newsworthy as many software vendors are using AI tools in very unimaginative and incremental ways. Other vendors see AI tools as a means of dealing with smaller business issues that somewhat affect productivity or efficiency.
Kantata came up with four solutions (all part of the Expertise Engine) that seek to make a material difference in how service firms sell, staff, deliver and monitor/optimize client projects. These four tools are referenced in the following graphic:
The Sales Accelerator appears to have the potential to dramatically shrink the time needed to create a proposal, get all proposal documents into a consistent format, determine an available (and great) team to place on the project, price the work, and, recommend content for the sales presentation. Prior to AI, this kind of effort often takes days to complete with some larger projects needing many weeks of work to get all of this together.
The Resourcing Accelerator takes the assignment of personnel to different proposals and projects to a whole new level. I wrote an extensive piece on this just two weeks ago. You can read that article here. This capability goes beyond Kantata's project specific resource management tool of recent years. It can continuously monitor project staffing and availability changes across ALL of a service firm's clients, recommend staffing changes to optimize current efforts and even support what-if scenarios.
The Delivery Accelerator taps into the firm's IP to serve up appropriate deliverables, work papers, data and insights at the point of need. I also liked how it can surface potential project risks and craft mitigation strategies for them.
Lastly, there's the Forecast Accelerator. This tool brings new kinds of visibility to margins and other service metrics. It also provides powerful forecasting (and re-forecasting) capabilities that go beyond a single project.
A Kantata source indicated that all four accelerators should be available in 2026.
Kantata's AI investments go beyond these four accelerators. Kantata can interoperate with other customer agents via Model Context Protocol (MCP). Customers can also use Kantata's Platform to build their own agents. That platform has an Agent Studio, Workflow tools, Knowledge Graph and the MCP.
According to Kantata:
Where the market is doubling down on building agents on generic LLMs, Kantata is building our own domain-specific small language model (SLM) with a proprietary services ontology built into it. We believe that is going to be the crucial thing that makes the Expertise Engine premise work. With a domain-specific SLM referencing a maintained knowledge graph of your business that maps the relationships that matter, the Engine is going to empower agents (ours & yours) with a level of context about what works and what doesn't. This will make the curation of and arming with IP possible. The embedded benchmark data will also be crucial in grounding that SLM in an awareness of the industry and what average and high performance looks like for (services) businesses.
The 'embedded benchmark data' will include licensed external data from SPI Research (Service Performance Insight).
The following graphic shows the summary view of the five key pieces to Kantata's Expertise Engine:
Kantata apparently started its investigation into the use of agents and AI by first evaluating the material problems service professionals are facing and then identifying technologies that could make a big difference operationally and financially for customers. Other software firms don't do this. They look at their applications and ask where can we insert some AI magic into these products? This is a big difference as the former can trigger the development of transformative solutions while the later lets a vendor wallow in incrementalism.
Kantata told me that they are building a 'system of expertise' not just operational apps. I can see that. The trick will be for Kantata's software implementers to also possess great change management skills as these tools will likely drive significant alterations to their customers' projects, operations, metrics and more. As Kantata sells a markedly more transformative solution, its own solutions personnel (as well as those of its partners) will need to upgrade their capabilities to deliver a more transformative end state for customers.
All of these four new tools can look at the totality of a service firm's business - not just aspects of an individual project. That difference in scope is huge as it involves the access to a form factor greater number of projects, data, working papers, estimates, staff, etc. The power of the new tools is beyond the ken of a mere mortal. These usher in a new way of running service firm operations.
That full visibility into all projects, proposals, IP, etc. also enables their customers to become more cosmopolitan (i.e., to been keenly aware of one's surroundings and environment). I'm sure Kantata will hear numerous revelatory comments from customers once they start using the new enterprise-wide tools.
There are significant implications with these tools, though. For example, many firms lack quality IP documents. What materials they have may be outdated, non-standard, stored on inaccessible media, etc. Before the new Kantata tools can deliver on their full promise, buyers may need to put forth considerable effort in cleansing, standardizing and codifying their knowledge assets, project methods and more. This will not be an insignificant task.
Finally, these tools are still going through the shakedown phase with new capabilities not available until next year. Prospective buyers should double-check product availability before buying.
