Could ‘intelligent’ software programs revolutionise hitherto laborious document control processes?
I was recently contacted by Ritesh Tripathi of TransBit Technologies, a software development company based in Hyderabad in India. Its lead product, CollabWrite, is a web-based enterprise content management system which he says is used extensively to capture project records and automate workflows among the multiple firms involved in major engineering, procurement and construction (EPCs), particularly in power, mining and oil and gas.
In its Indian sub-continent core market, CollabWrite has won some substantial client projects, and occasionally competes with international SaaS collaboration vendors such as Aconex. However, there is still considerable untapped potential in these markets, partly due to ongoing cultural and process change challenges. Document controllers are still deployed to manage internal systems, and many document transactions rely on separate email and attachments, or FTP or similar file-sharing tools, and – sometimes – SaaS collaboration or on-premise ECM platforms (Ritesh named Enovia, SharePoint and Documentum). As a result, files still need to be downloaded, validated and entered into internal enterprise systems by document control staff, whose daily work may involve creating cover-sheets and transmittals, stamping drawings, etc. Paper remains important with some items repeatedly scanned, stamped and manually signed (Ritesh showed me examples of drawings that were extensively marked-up by hand, and which went through numerous analogue to digital iterations and file format changes – eg: DWG to TIFF to PDF).
Bring on the RoBOTs
To automate and accelerate this document management process, TransBit has developed RoBOTic Document Control Assistants, which can aid human document controllers. These are not physical robots but software tools. Combining computer vision, artificial intelligence, digital image processing and advanced decision-making algorithms, these digital assistants can create cover-sheets and transmittals for out-bound documentation, interrogate incoming digital files, capture data – from drawing title blocks, for example – automatically, manage compliance processes, and, when necessary, highlight exceptions (errors, conflicts, etc). The latter can then be resolved by human document controllers.
Ritesh said the RoBOTic Document Control Assistants can fetch emails from project in-boxes and can be authorised to work with SaaS collaboration platforms and similar systems. In each case, the ‘Assistant’:
- fetches the email or transmittal
- identifies the type of document and its purpose
- visually checks the files, including identifying decision stamps
- performs quality checks and data reconciliation with previous transactions
- updates the internal system with metadata details of the decision codes, transmittal numbers and files received
- notifies relevant people, or initiates other relevant system workflows, and
- in case of conflicts, errors, etc, gets the human document controller to process exceptional items.
He demonstrated a process involving transmittal of a complex A3-size image, and the whole document acceptance process was completed in 47 seconds – far faster than could be manually achieved by the fastest human document controller. Ritesh told me that practical simulations had demonstrated that RoBOTic Document Control Assistants could reduce the work of the conventional human document controllers by up to 85%. Such productivity benefits may also be accompanied by improvements in the accuracy of enterprise record systems, and do not involve replacing existing technologies or processes: “This technology is simply an additional server in their datacenter – that can be integrated with the existing ECM.”
In the UK and similar developed economies, the use of SaaS collaboration tools to exchange digital information and manage electronic workflows is well-understood, and since these platforms’ introduction in the late 1990s/early 2000s, the AEC industry has developed the necessary tools and processes to manage the ‘less paper’ (still not quite ‘paperless’) working world. However, in other markets, these approaches are still in their infancy, and many projects and the multiple organisations involved still require centralised internal records management systems.
It was clear from Ritesh’s demonstration that the manpower requirements to run such platforms could be significantly reduced by automation, and he sees opportunities to partner with SaaS vendors (and on-premise ECM providers) to make inroads into the Indian market by providing this layer of technology to extend and expedite routine document management and control. Aconex is clearly on his radar as one firm already active in the region, and the TransBits focus on EPC clients could also be interesting to McLaren Software, which has a track record of supporting major mining, oil & gas and other natural resource projects, to 4Projects which is targeting similar work in Australia (post) and the US, and to QA Software (whose QDMS is deployed by one of TransBit’s customers).
Update (3 June 2013) – Ritesh has forwarded a case study to me with some test results from running the TransBits RoBOTic Document Control Assistants on a Fortescue Metals Group mine project which was using a QA Software EDMS. The simulation results showed:
- 100% Accuracy in transactions performed by RoBOTic Document Control Assistants
- 75% of valid test cases processed successfully.
- 48% of total test Cases processed successfully, meaning ability to reduce the work of human document controllers by about 50%.