We’re not just talking about “personal information assistants.” Those who complacently expect that AI in their offices will essentially be a search engine on steroids are missing the point, as are those whose AI cognizance comes mostly from the latest “Mission Impossible” movie plot.
Although artificial general intelligence (AGI), agentic AI (AAI) and whatever other advanced forms of the technology are hurtling toward us will be used eventually by public agencies and pension staff teams — first for customer services and call centers, and then for internal and online data collection, assembly and report-writing — the public funds will typically be followers who rent and use prepackaged technology, not leaders at the cutting edge of this industrial revolution. The question here is who will be left behind as AI roadkill in what some foresee as “a white-collar bloodbath.”
As with other institutional investors, only the very largest pension systems will have the opportunity and capacity to build out sophisticated cutting-edge AI capabilities to pursue market-beating returns. The other 95 percent will hire out those capabilities through consultants and money managers, especially in the alternative-investments realms of private equity, private credit and real estate. In the early years of this new technology, the impact on internal staffing will be modest, with only minor reductions in headcount as system managers realize that fewer analysts and paraprofessionals are needed to spit out the reports senior managers and trustees rely on for decision-making and fiduciary reviews.
This won’t happen overnight. Over time, as raw data becomes available through extensible business reporting applications and then gets digested by AI, a few internal positions here and there can be left vacant or eliminated. Such gradualism won’t change the systems’ actuarial contribution rates in any meaningful way. The real question investment-wise then becomes whether these AI systems can eventually provide or at least recommend important portfolio decisions directly to pension fiduciaries — and if so, whether stakeholders can trust their reported outcomes. Some skeptics therefore argue that it’s largely a waste of time, at least until the bugs get worked out.
Even so, it already makes sense to start planning soon to train qualified staff members in the use of this powerful, fast-moving technology to perform more than just the grunt work — and to head off the production of “AI slop.” Pension leaders might take a cue from angel investors who have already begun using specialized AI tools for screening and due diligence — and most of them are just part-time home gamers. At the very least, AI can already provide probing questions and make counter-arguments to challenge marketing pitches; it’s an accelerating force to be reckoned with.
Where to Start
Here the various pension associations need to focus more energy beyond hosting conferences where sponsoring marketers pitch their wares. The National Conference on Public Employee Retirement Systems has laudably set up a mostly middle-market network of chief executive and investment officers that would be especially well-suited for this forward-looking task of training the first generation of AI-savvy public pension investment professionals. To get the ball rolling, that nucleus must now be actively supported and reinforced collaboratively by the largest public pension funds’ intelligentsia. Next, the industry needs a central public pension AI academy, offering classes within a year or two once reliable AI-powered investment applications prove out.
Such an academy could offer a progressive curriculum of using, manipulating and then training large language models for internal operational efficiencies and staff augmentation, then advancing to higher-level courses on using and training AGI and ultimately AAI systems for pension-specific investment applications. Because they do not compete with each other directly, attendees can learn faster through classroom and post-graduation collaboration with peers, and ultimately build and operate a cooperative AGI-AAI platform as described in a previous column.
Outside the pension realm, a recent survey of the 25,000 members of the Government Finance Officers Association found remarkable, overwhelming interest in AI training in their field. One must suspect that there is a similar interest among public pension professionals. Nobody wants AI to turn into a career-killer. Strategically, “AI for Dummies” courses will soon become available through commercial professional-education companies that will compete to teach basic AI skills to office workers everywhere, so a public pension AI academy needs to aim higher, focusing on functions unique to the funds’ investment milieu.
As one example of the potential applications of such a cooperative network once their key staffers are trained and skilled, public pension systems and other institutional investors could use AGI to establish a collective marketplace for their private equity investments, which would provide realistic fair values for those assets to replace the Pollyanna-ish valuations often used for accounting and performance metrics and avoid being scalped by intermediaries. They could buy and sell holdings among themselves without middlemen. Eventually this platform could also serve experienced and proven entrepreneurs seeking pension capital directly without the high costs of private fund managers, a benefit for both sides.
Aside from faster rules-based, cookie-cutter report preparation, AGI will become a visible new force throughout the public pension world via the investment consulting industry. The proficiency of their professional teams in deploying its capabilities will become a competitive differentiator for firms seeking to maintain and expand their market share. A second stage of this evolution will arise when smart systems can design better portfolios that eliminate, or at least reduce, reliance on high-fee money managers. One of the nation’s premier investment consultants has already wised up to this game by going into the business of assembling private-equity portfolios itself.
Staffing and Governance
Stepping back from AI tech for a moment: There is nothing etched in stone about public pension funds needing to play the game of “chasing alpha” — trying to beat the market — when their essential mission is simply to consistently deliver reliable returns on invested capital. For public pension trustees and chief investment officers, the pursuit of investment exceptionalism is much more about egos, bragging rights and resumé-building than the resulting employer contribution rates and their reliability. For many pension plans, AGI is quite likely to eventually expose the alpha chase as a prideful waste of time, effort and money. This insight has profound staffing and governance implications.
Pension staffers, investment officers and CEOs, and their associations need to wake up now to the implications of all this in the coming years. The selection of pension consultants will increasingly focus on which firms have the technology and the capacity to supplant those high-fee money managers, which raises the question of whether internal staff has that skill set capacity. Will it be necessary for trustees to hire a selection consultant before hiring the portfolio consultant?
Here’s where trustees and senior pension staffers will face some real anguish in the coming years. There will come a day when AAI systems and related technologies may very well enable a top-tier handful of outsourced chief investment officers to produce better, lower-risk and more cost-efficient portfolios — and replace most of their entire department. For governing bodies who can eventually accept the notion that it’s impossible to outsmart markets that will increasingly become more efficient as proprietary-information advantages dissipate, an outsourced chief investment officer model may deserve consideration.
For those who find that idea to be a step too far, there will be a hybrid to consider. Savvy pension chief investment officers will soon realize that AI-powered outsourcing is likely to catch hold, and to preserve their own place in the organization chart they might best focus their trustees on the alternative-asset portion of their investment program. That costly category now accounts for a quarter to a third of pension assets, and often two-thirds or even three-quarters of all investment fees. An outsourced deputy chief investment officer, responsible for just that portfolio category, could replace the growing cadre of internal staffers and at least some of the high-fee managing partners of private equity, private credit and real estate funds who now enjoy impenetrable anti-competitive moats through their clubby social networks.
The outsourced deputy chief investment officer need not necessarily be an individual; it could also be a private advisory firm or even a large state system’s portfolio team contractually managing alternative assets of city and county systems. The California, Florida, Michigan, Pennsylvania and Texas systems might become ripe for such a cooperative sub-portfolio outsourcing strategy.
Trustees will still have a role in these processes by participating in the asset allocation and risk parameters process, and governance consultants will be there to help them think that all through. What trustees won’t be doing for too much longer is holding beauty contests for dozens of storytelling investment management marketers every year. Those sideshows will increasingly be the dumber money in this new age. As they say in poker, “If you don’t know who’s the mark once you sit down at the table, it’s you.”
A Tectonic Shift
In a few years, pension leaders should expect to see Stanford, Caltech, MIT, Harvard, the University of Chicago and other biz-tech intellectual citadels offering professional credentialing programs in “investment information management technology” along with licensing deals for their own portfolio-construction AGI sandboxes. A handful of private companies may throw their hats in the educational ring. Most likely, though, only one or two of those will ever show an interest in public pensions, which is why a niche public-pension AI academy could provide the catalyst by reaching out proactively to these knowledge and skill centers.
Investment professionals, including the 200,000-member chartered financial analysts’ institute and the various professional associations that serve staff members, are awakening to the tectonic shift now beginning. It’s foreseeable that the CFA Institute would also set up its own AI certification, but its prerequisites will likely be too arduous or academic for many public pension staffers. Employees who enter more customized and focused programs to retrain themselves to harness and proficiently operate these smarter systems will have a fighting chance to retain key jobs in the next decade as AI technology relentlessly thins the herd. For most, that will only happen if they take the initiative collectively.
As with the demise of horse-drawn buggies, this will all take several years. Public pension systems are monopolies with guaranteed funding for operational expenses and an embedded cultural bias against cost-cutting layoffs. Their CEOs will happily accept higher productivity and lower costs from their outside vendors, but significant internal replacement of labor costs by technology will come incrementally — unless or until outsourcing suddenly overtakes internal staff as the obvious and superior alternative. Remember what they say about boiling a frog: Complacency will be the enemy of those who expect the technology to come to them and not for them.
What is for certain is that the public pension associations and their professional subdivisions will have their work cut out for them as analytical technology and outsourcing replaces old-school bureaucratic decision-making processes. To state the positive: Other than AI inertia, there is no reason that they cannot place themselves at the center of their industry’s techno-revolution by hosting a focused, specialized training center of excellence and upgrading it continuously, while building out the professional community’s sandbox and database for trainees to master and deploy when they return home. As the business guru Alan Deutschman first said about embracing innovation 18 years ago, it’s now time to change or die.
This is the second of two columns on the impact of artificial intelligence on public pensions. Read the first column here.
Governing's opinion columns reflect the views of their authors and not necessarily those of Governing's editors or management.
