The inflection point for intelligent capture has arrived. The question for you now is not whether to proceed, but rather how best to proceed.
It is clear in 2020 that advanced technologies for identifying and extracting information from unstructured content are maturing to the point of readiness for document-intensive enterprises. While vendors have tantalized us with “parlor tricks” over the past 2-3 years – with narrowly-defined demos that made implicit promises their technologies couldn’t actually deliver on – this will be the year of legitimate, substantial progress.
The feeding frenzy in this space is well evidenced by the list of participants. From the traditional major capture players – IBM, Kofax, ABBY, et al – to the major cloud service providers – Amazon, Google, Microsoft – to an array of boutique startups with their bullpens of newly minted, applied math PhDs, the rush is on to fulfill the long-anticipated promise of intelligent capture.
It’s About the Customer Experience
Right now, it’s about human key-stroke elimination, improved business insight better customer experience, and enhanced operational accuracy and compliance.
And the underlying technologies involve lots of heady descriptors – artificial intelligence, machine learning, cognitive analytics, and so on. But beneath it all, both the value propositions and the technologies are compelling.
So how do you proceed in taking advantage of this progress? After all, the winners are not clear at this point and evolution promises to remain highly dynamic for a while. But now is not the time to simply play “wait and see” – the business value is too great. How do you prepare yourself to take advantage without potentially locking into dead-end paths?
Here are Some Thoughts:
Differentiate between capture “platform” and capture “services.”
Think of your capture platform as the enabling framework of your enterprise capture world, managing content ingestion channels (both paper and electronic), authentication and security, business rules, and workflow supervision. These platforms are foundational and not easily replaced with each new trend.
On the other hand, most of the intelligent capture capabilities mentioned above are consumable as standalone web services that can plug and play. Make them continually earn their place in your environment, and be wary of emerging capture services that require their own proprietary capture platform.
Ensure your capture platform supports the unique demands of this journey.
The optimal capture platform is highly configurable, based on open architecture, and plays well with others — all while solving primary workflow issues.
Support for Sample Content Curation.
No AI or ML-based tool is any better than the samples you provide it. You will need training sets and blind testing sets, and we find that organizations consistently underestimate the effort and importance involved in getting this right.
Support for Efficient A/B Testing.
Many of the emerging intelligent capture services have particular sweet spots. Some are better on form-based content, some provide contextual awareness, some are “black box,” and some are “white box,” some have unique “entity extractors” specific to target industries and use cases. To cut through the promises and noise, you need the tools to quickly and accurately compare intelligent capture services to each other.
Ability to Orchestrate Multiple Intelligent Capture Services.
Once you understand which capture services best address particular content types, you need a tool to orchestrate the 2dynamic application of the right tool to the right job.
Integration with Process Automation.
It is increasingly accepted that the reach of capture platforms should expand to support the more robust acquisition and disposition of content. How do you automate requests for needed content, or distribute extracted data to business systems? Whether through integration with 3rd party RPA tools or through expanded built-in automation, this feature expands the value of intelligent capture.
Audit trail.
As automation increases, the need to have a complete, transparent record of your content’s journey only grows.
Focus on value.
Take the time to understand where intelligent capture can be most impactful in your organization, and start there. Don’t become paralyzed by the missing 10% in a 90% great solution – 90% great is still great.
Build your Own Intelligent Capture Competency.
Few technologies you’ll adopt over the next few years have the potential to be more transformative to your operations. You owe it to your organization to get this right, and it’s worth dedicating some of your best internal team members to become experts. The right capture vendors will work best with informed customers.
Measure Progress.
Establish success metrics that align with business impact. It’s easy to become enthralled with isolated metrics like auto-classification and extraction rates and false-positive rates. Unfortunately, definitions of these metrics can be tricky and sometimes become altogether separated from the more important question of whether your business is running better.
So good luck as you go forward! Be aggressive in pursuing value, but avoid locking yourself into experimental paths. This year promises extraordinary opportunities for those who do it right. We can help you there.