Is "automation" at the core of all modern solutions or is it just the new “sizzle” being sold to make the same old steak look more attractive?
In some cases, it’s the former and in other cases it’s the latter. Here’s how to tell the difference – and it will require work on your part. There is no substitute for doing your own due diligence in these matters.
Practitioners look to their software vendors to help them automate tasks and make their work more efficient. Successful firms are making remote work an option for their teams with browser-based applications, cloud hosting, and next generation automation tools.
Many vendors, hosting companies and consultants offer a range of services to help firms offer flexible work options to their staff while maintaining efficiency and service quality. Unfortunately, firms often don’t redesign their internal processes and workflows to work with the new toolsets until they have had a mediocre implementation – which reduces the credibility of the tool and those charged with implementing it.
What to Look For
When you select and implement a tool, you should understand how the vendor envisions its use and be prepared to change some parts of your processes to make your tools work for you and your team. We think of the services provided by vendors as technology solutions, but the reality is that human work still plays a role in many tools.
Some applications use technologies like optical character recognition to perform rote tasks like extracting data from invoices or tax forms, others import data by downloading it from financial institutions, a third group simply uses inexpensive offshore human labor to accomplish the same tasks, and the most successful ones use a combination of all three approaches – OCR, integration, and human verification – to extract accurate information from paper source documents.
If these tools solve the problem of getting data into a structured data format so it can be used by another application, some of the approaches are truly technology solutions, while others are simply shifting work from your employees to an offshore service center which does data entry for thousands of customers. All of these approaches can play a role in your processes, and you should ask enough probing questions so that you understand which approaches are being used, and can adjust your expectations to match the reality of how the solutions work in a production environment.
Manage Your Expectations
It’s critical to set proper expectations for what your solution can do so that your implementation gets a fair chance. Some early salespeople of computer text recognition boasted of “99 percent+ accuracy without human verification” and when the software only delivered 85-90 percent accuracy (partially due to poor scanning procedures by the firm), the solution was deemed to be a failure.
Checking the performance claims of vendors by doing a pilot project with some of your data in your environment will help you figure out how a new automation tool will work with your processes and your client data. Not all efforts will succeed, but by investing some time in trying new approaches, you will learn more about the solution and identify potential issues in your firm before you try to sell an implementation to your entire firm.
What are You Signing?
Some vendors offer big discounts on their solutions in exchange for signing a long-term non-cancellable contract for multiple years. If the solution performs fine for them, it’s not a big deal, but if it doesn’t work – or if they change it in a way you do not like, the firm is left holding the bag.
When some of these dissatisfied users read the fine print, they learn that they have signed what I call a “Hotel California” contract – one where “you can check in anytime you like, but you can never leave.” Some of these contracts have early termination fees of 75-90 percent of the fees for the remainder of the contract term after day one, which means that you have little recourse if you and the vendor cannot make their solution work for you.
We can easily have an honest misunderstanding of a conversation with a salesperson months after the fact, but the contracts you sign (on paper or by clicking on them) will usually spell out clearly what you are (and are not) entitled to receive as part of the agreement. It’s critical that someone at your firm reads (and understands) the contracts before anyone signs on behalf of the firm.
At the same time, while you can’t eliminate risk from purchasing a new solution, you can reduce the risk of a failed implementation using the steps I mentioned above. Here’s a quick checklist:
Make sure your firm is willing to change its processes to take advantage of any new automation tools you select. Without process change, the possible benefits of any application are more limited.
Understand how the solution works, the types of automation which are used, and adjust your expectations to reality before you sell it to your partners.
Try out new technologies on a limited basis to see how well they work for you– and make sure you also understand where they fall short of expectations– before you deploy on a widespread basis.
Read and understand all of the fine print in the license agreement, service level agreement, terms of service, and all other legal documents before you sign a long term contract so that you don’t have unexpected surprises later.
Finally, I’d suggest that you review all of your applications periodically in light of your long-term strategy for the firm. If you plan to have your firm acquired by another firm in three years, it may still make sense to implement a new solution, but it may not be wise to sign a seven year non-cancellable contract.
Similarly, it may not make sense to invest in a new solution for a practice line which you plan to discontinue in the intermediate term. Good luck in your efforts!
Brian Tankersley CPA CITP is a technology consultant, educator, writer and serves as Director of Strategic Relationships for K2 Enterprises, where he works with vendors serving the industry to understand their existing and new offerings.