I’m looking for suggestions for existing software services that are specifically created for reviewing project specifications for quality, completeness, accuracy, etc… Ideally, it would find all the missing reference sections, TBD’s, “By Architect” and so forth. Extra points for being able to import keynotes or drawing content to verify the materials are covered in the specs.
Hi Nathan, GREAT question and many using VisiSpecs love our QA Report. Contact us for more info and we’d be happy to run your last project through it to show you the errors it picks up. Seamus
We are currently working on an AI driven QA/QC tool for specifications within Specpoint. We are going to be looking at a lot of what you mentioned, plus many more items.
Additionally in later phases we will be looking at doing keynotes and the revit model.
We’ve had really good success internally with this tool and excited to get this out to customers!
I know Document Crunch provides analysis and reporting based on specific queries. My impression is that it is still up to the user to know what prompts are needed to get useful results.
We are experimenting with AI for use with Conspectus Cloud. We have not found a use case that will produce consistent, reliable results, yet. Conspectus Cloud has global search that can include results for a single project, all projects, and all masters - at the user’s option. It also displays referenced spec sections in red text if the documents are not included in the project, the visual cue that the document is missing.
These two features have sufficed for our specification consulting services to ensure we deliver quality documents.
LLM is a “large language model” which is the basis for most artificial intelligence systems in use today. It refers to the way AI “scrapes” any text it can find on the internet to build its “intelligence”.
Specpoint just came out with Spec Intelligence which is their AI driven QA/QC tool. I have only played with it for 2 days now and it seems to be very promising. I am only running Master Office Specifications that I have written through it presently. I want to see how to manage it before I run a whole project through it.
Specpoint actually has two AI moduls. Dela which works only within the Masterspec data base. This one can be used for asking questions or comparing. The second one is Spec Intelligence which checks for completed specifications or entire projects. I am new to this forum and I find it very interesting, good job. Have a good day.
Rick, thank you for sharing more about the Specpoint AI tools. And welcome to the forum. I hope you come back often and participate in other discussions too.
Thanks for the welcome. I have spent some time going through some of the topics and find them interesting and informational. Unfortunately, I just stumbled onto this forum just recently, unfortunately it is at the tail end of my career. Would of liked something like this when I first started out, especially in these later years where I am writing specifications for some of larger projects in our office. Everyone have a great day.
Ahhh… It is never too late to participate and share your knowledge for all that come after. @Davemetzger and @George_Everding can attest to that. The forum has been in place for over 20 years.
Hi Nathan, we use InspectMind quite successfully. It is easily shared among the team, is thorough in reviewing drawings against the specs, can perform code review and report on outcomes, and responses are easily tracked within the tool.
That’s good to hear. We reviewed InspectMind not too long ago, but I think we favored the results in SpecSet in a comparison test. So cool to see how fast these tools are developing Specset: AI document control from design to operations
@jmarsili and @nwoos Conspectus just went through an review process where the owner used a product called LightTable to conduct the review. The result was hundreds of comments on the complete drawings and specs. This is my concern. The AI tools check completed docs. Then the teams must respond and correct reactively - rework at the end of the project, potentially delaying bid issue or causing addenda or modifications.
It would be so much better if this was a continuous checking process where coordination items could be flagged as they occurred. Then the team disruption and rework would be minimal. What if checking started when design started?
The scenario is more complex than you have suggested. LightTable is decent, we are using it on a project. But like many, it flags a lot of issues, some, perhaps many, are false positives, or others are simply so minor they don’t require correction. It’s separating the wheat from the chaff that is the new art form, and potentially, engages an all new level of risk. Did we mistakenly disregard or downplay a flagged issue? Can it be argued that we were aware of the issue? Clearly, it is unreasonable to address every single comment or issue, so where has the new standard of care moved to?
If we don’t run the QAQC software, but a client does, how does that change our liability? If its just the contractor running the software, how does that change our CA fee to address what could be many hundreds of RFI’s resulting from that QAQC scan?
It is clear to me that we MUST engage with some sort of QAQC program to maintain the new standard of care, if for no other reason than to keep up with the contractor and client who are doing the same. But what we do with all those comments is the real question. It will probably be, for a short time, an all new discipline of Expert, opining on the validity or seriousness of a comment, after the fact.
We have found SpecSet to be better than most when it comes to sorting and ranking the issues it finds, and more impressively, the subtle complexity of the issues it finds. Its still hundreds and hundreds of comments, but they can be filtered, sorted, and ranked with good fidelity.
