|
Los tipos de archivo clásicos dejarán de estar disponibles a partir de enero de 2021. Puedes migrar tus archivos clásicos o descargar un PDF. Más información

Workiva AI - XBRL

Solicitar demostración
  • Editado

In light of Workiva incorporating AI within some of its offerings' functionality, is there a plan to utilize such for XBRL?

Seems a great case for AI generative responses would be for tagging XBRL.  For instance, it would seem that the current Workiva AI capabilities should be able to take the context from narratives / tables for related numbers, as requested by a user, and source suggestions for tags based on the Workiva XBRL Taxonomy Research tool.  Of course this would all take place within the XBRL module of WDesk.

At this time, unless I'm mistaken, there isn't such a tool to provide suggested tags, but only detective validation methods when a tag may not make sense or is misplaced, or other improper use thereof.

2

Comentarios

11 comentarios

Fecha Votos
  • Lucas Aguirre

    I find that the LLM's are pretty terrible with XBRL, and will just make up tag suggestions that don't actually exist in the taxonomy. I agree that this would be very value-add from Workiva if they can refine their tool for this purpose. 

    2
  • Isabel Messore

    Hi Pat Flanakin, 

    Currently, Workiva does not have AI capabilities specifically designed to assist with XBRL tagging. While our AI tools are integrated into various workflows to enhance productivity, such as intelligent search and document assistance, we do not yet provide generative AI features for suggesting XBRL tags based on context or taxonomy research. The existing tools focus on validation methods to identify misplaced or improper tags rather than suggesting tags proactively
    0
  • Pat Flanakin
    • Editado

    Thanks Isabel.  I was more asking about a plan to address the current status of no AI functionality for XBRL yet.

    0
  • Pat Flanakin

    Lucas Aguirre
    RE: "LLM's are pretty terrible with XBRL"
    I expect they won't be for long.  Would seem to be low hanging fruit for the first SEC filing development platform to get this in operation.

    1
  • Pat Flanakin

    Mike Davis
    Mike, maybe you could help us understand why EDGARsuite has this functionality but Workiva doesn't.

    1
  • Mike Davis
    Great convo here! Regarding Workiva's AI functionality with XBRL + AI, this year we will be looking into possible tagging assistance. That said, do not have committed timelines at this stage though we'll be sure to keep our sites updated.
    A couple of questions we have for folks on this convo:
    • Which reporting solutions/taxonomies are you using Workiva for today?
    • Would you have a desire for our Product team to reach out on additional learnings?
    • if Workiva were to have an EAP or Beta program, would you be interested in possibly participating?

    Let me know what follow-ups you have and chat soon!

    0
  • Pat Flanakin

    Ok thanks Mike for the additional info.

    For us, we are using Workiva and we always are using the latest taxonomies for SEC filing reports 10-K / 10-Q / 8-K.
    We at this time don't need additional education on Workiva's current AI capabilities; Workiva AI's website providing such information appears sufficient and I think that if anyone has used one of the AI engines available, they can see what Workiva AI is doing makes sense.
    As far as a Beta program, I would be interested in seeing what such would look like and possibly provide feedback if I participated...would depend on timing (due to SEC filing deadlines and commitments for the day-job) and commitment weight.

    0
  • Remco Bredenhoff

    Mike Davis

    "if Workiva were to have an EAP or Beta program, would you be interested in possibly participating?"

    My team is interested in this beta program, hope to hear when there are updates on this

    0
  • Isabel Messore

    Hi Pat Flanakin and Remco Bredenhoff, 

    I've passed along your interest to our product managers! I'll be sure to report back if a beta program for this functionality becomes available. Thanks for sharing here :) 

    0
  • Pat Flanakin

    Lucas Aguirre Do you mind sharing how you tested LLMs on XBRL tagging?  Did you just copy and paste a footnote into an LLM engine and ask it what tags would be appropriate?
    Let's assume for the moment the above is the case, I would not expect to get good results; however, within Workiva, I expect this could easily be added functionality.  They already have Workiva AI and an engine particularly trained and coded specific to provide results conditioned on the Workiva environment.

    Given this, with the coding already in place within the Workiva WDesk XBRL tagging platform, they are already using it to determine tagging outliers internally; therefore, in my opinion, it is likely that expanded functionality for XBRL in terms of Workiva AI is closer than we think.

    0
  • Lucas Aguirre

    Nothing in-depth at all. Essentially have used it as a glorified search engine with prompts, to bad results.

    Agreed, on all counts. 

    0

Iniciar sesión para dejar un comentario.