Why Your Next Investment Writer Will Be a Machine

The next time you need an investment writer, don’t bother calling HR. That was BEFORE. Before a firm in Chicago made software to report out baseball game stats. Before artificial intelligence began writing stories. Before AI went from sports to finance. Before algorithms jumped from high-frequency trading to monthly reports.

Generating stories at the click of a bottom based on deep data sets–that’s NOW–and Chicago-based Narrative Science has been a pioneer and the most visible face of this fin tech gadgetry since 2010. ‘Gadgetry’ might be the wrong term, since this development promises to upend asset managers’ reporting processes and force a rethink of the investment writing function, along with the need to hire an actual human being for an investment writer role.

 An umbrella for reporting season

With this new tool, Narrative Science CEO Stuart Frankel says producing something like an investment strategy [pooled fund or SMA] report “goes from the job of a small army of people over weeks to just a few seconds.” As anyone familiar with the month-end and quarter-end rush periods knows, this represents a dramatic improvement over the status quo.  The firm says that their asset manager clients have been able to ‘reduce the number of days spent producing portfolio commentaries by 50% to 75%’ delivering reports that are marketing-approved and compliance compliant.

If it all sounds to good to be true, that’s not the word I would use. In my opinion, it’s something else entirely: eery. I actually got chills watching the 3-minute video below on how Narrative Science’s Quill is being used to expand Credit Suisse’s investment research coverage. The algorithms identify trends and report in plain (analyst) language what’s going on with a given company (eBay in this case).


(For website owners, you can enjoy Quill Engage –their Google Analytics app for site analytics– for free.)

 The new wave

Turning data and analytics into stories clearly has momentum. In March 2014 the Financial Times reported that Fidelity was testing technologies offered by Narrative Science and Kensho, among others. Narrative Science counts American Century Investments and Nuveen Investments along with T. Rowe Price, Credit Suisse, and USAA among its clients. The firm’s portfolio commentary tool for equity strategies had 15 clients by November 2014. A fixed income version was recently made available. (Personally, I’m curious how the package handles non-benchmarked strategies in either asset class).

 Welcome, our new robotic overlords

Kensho’s software, Warren, may be more ambitious in that it allows investment firms to test ideas by unleashing what CEO Daniel Nadler calls ‘a quant army.’ With these technologies, it would seem to follow that the need for flesh-and-blood investment analysts, writers and editors would be reduced significantly. That’s a bit premature. Adoption will take time. Piecing together the headline story on a stock from a bunch of data is impressive, but it doesn’t provide a buy/sell/hold value judgment. It will clearly reduce the need for manual low value-added tasks such as number crunching and descriptive writing considerably. It appears that Narrative Science has for several years been combating the perception that their technology will enable firms to kick workers to the curb. In 2013 Frankel said “It’s less about replacing people, and more about leveraging those folks that are already there.” To be fair, I know some veteran writers and analysts who are overworked and looking for more challenging, creative tasks. Perhaps these new tools will help them make that shift.

Tomorrow, the world

No doubt the AI technology will improve as time goes on, and there are several areas for development. For one, I have yet to see any information about multi-asset or alternative strategy capabilities, though that’s probably just a matter of time. Second, while English is hands down the lingua franca of finance, asset managers highly active in Europe and Asia would have a hard time refusing a reporting tool that delivers material in multiple languages with speed and precision. Once the technology overcomes these hurdles, you can expect to see it implemented everywhere. The question will then become whether these tools become victims of their own success. For reporting, it’s easy to imagine the entire industry racing to reduce cost pressures and improve competitivity. But when it comes to testing investment ideas, it’s hard to believe that firms will be willing to use the same testing environment as their competitors and run the risk of all using the same playbook.