Underdeveloped technology no substitute for bedrock principles of content management
PJ Bellomo, CEO of Proposal Software
In November, I grabbed breakfast in Saratoga Springs with Peter Dean, a former global sales executive at Federal Express who founded and now serves as president of RenderTribe, a boutique technology marketing firm. At one point, our conversation turned from the locally-sourced breakfast ingredients to something decidedly more geek: the future role of artificial intelligence (AI) in the proposal management space.
Peter posed a cheeky question: “When it comes to AI and proposal management, what would you like first: the bad news, the bad news or the bad news?”
With my eggs and smoked bacon behind me and a warm cup of decaf in my hand, I settled in for the first piece of bad news.
Peter then went on a small rant. “It’s unprofessional and frankly irresponsible for small- and mid-size software companies to make promises related to AI unless that’s the sole focus of their firm. Even then, it’s likely hype. The challenges remain large, and unless a software company is thoroughly focused on AI, keeping those promises will be an expensive undertaking for a number of years while customers wait for the type of practical AI results that deliver value in daily business operations.”
Peter went on to mention the still-evolving Einstein initiative at Salesforce as a case in point. And he’s not alone in his observation about Einstein.
Business Insider reflected that “Salesforce’s AI initiative only started to gain traction in 2014 after it acquired RelateIQ, and although it has spent nearly $1 billion aimed at boosting its AI capabilities, including the addition of 175 data scientists, it will likely take a little while before it catches up to some of the more experienced competitors in the field.”i
Peter predicts that, over time, Einstein will prove out as another great success for Salesforce, but he cautioned about the hype from others.
“Companies that are not building in-house AI solutions simply jump on the acronym band wagon,” Peter said. “Without acquisitions, even the big boys struggle to produce true AI functionality. Frankly, we see companies describing their products with ‘AI’ and selling dreams instead of realities.”
Not quite ready for prime time
With a little less energy and color, Sheila McGee-Smith made similar observations earlier this year. In her case, she mentioned Watson in the same breath as Einstein.
“I think in 2016 that the hype began for Artificial Intelligence and bots,” McGee-Smith said. “2017 will continue the hype, but we will likely have the inevitable disillusionment as companies realize it’s not as easy as turning on an application, but collection and integration of a variety of data sources will be required. Artificial Intelligence is going to take time. We’ve seen the beginning with IBM Watson and Salesforce Einstein, but I believe we’re still at the pure technology phase. It will take time before real applications are deployed at real companies.”ii
With any other breakfast companion, I’d have hidden the high test by then. But Peter apparently enters this world pre-caffeinated and doesn’t drink coffee. He gulped his OJ and moved on to the second piece of bad news.
This time, he qualified his remarks; bad news for some will seem like good news for others.
“In the future, AI will reduce the number of jobs in the proposal writing field, just like AI will reduce the number of jobs in countless other fields,” Peter said. “For some large companies with large proposal writing staffs, AI will drive significantly lower labor costs.”
As machine learning, natural language processing, and other AI fields become more advanced and accessible, they can be leveraged to solve many types of structured recurring tasks. Receiving a request for proposal (RFP) and creating a first draft proposal is a good example of such a task. At some point, variations in the format and language of an RFP won’t impede true AI functionality, as they currently do. Eventually, AI will replace much of the manual labor involved in generating at least the first draft of a proposal. Furthermore, the quality of AI results will only strengthen with use, continuing to reduce the need for human intervention.
An investment with inherent risks
Peter questioned whether the proposal management space made sense as the next best domain to utilize AI.
“Sure, AI will take work away from proposal writers at some point,” Peter explained, “But if I’m a big financial institution or insurance company, do I really believe that RFP response is the top problem to attack with AI in 2018? Really? Imagine a proposal manager and a CIO from a major corporation meeting and agreeing to put precious resources into proposal management AI.”
As Peter described the scenario, I imagined how those conversations might go. Outside of controlled AI beta testing, all I could see was a race to determine which bozo reached the unemployment line first.
For us at Proposal Software, Peter’s comments hit close to home. In the last year, two of our customers, a major insurance company and a large bank, asked to work with us on this exact problem.
The insurance company’s team of data scientists, notably absent of bozos, found themselves testing AI use cases, and in this instance, the IT guys soberly approached the proposal management team as a low-risk testing ground.
After several months, the insurance company’s data scientists abandoned the initiative. They concluded that, given the current state of technology, the resource investment didn’t make sense. When they shut down the project, the lead scientist lamented: “The juice isn’t worth the squeeze.”
As for the bank, we developed a custom integration, allowing them to feed data into their AI system for machine learning purposes. Months have passed. No results yet.
Just a stone’s throw from one of the most historic horse racing venues in America, Peter offered his final piece of bad news.
“We’ve got a cart-before-the-horse problem here,” he said. “Reorganizing mountains of data in order to produce a truly practical and easy-to-use AI solution may prove to be a huge investment of capital and time that might be better used elsewhere. Otherwise, AI may simply produce a fast first draft riddled with wrong answers. Who’s the bozo that thinks that’s a good idea?”
Peter continued: “They have bigger fish to fry before they put AI in a technology that no sales person is going to trust.”
Peter put down his orange juice and winked because he knew he was preaching to the choir, including his repeated reference to bozos.
Content is still king
Regardless of evolving technology, the discipline of content management remains the top priority in the proposal management space.
For some time now, we’ve worked with our customers to uncover bedrock practices in proposal management. This brings us to the “AI cart” and “content horse” issue referenced above.
In 1996, Bill Gates famously wrote his “Content is King” essay and thousands have rushed to repeat his proclamation ever since.iii Among the many voices echoing Gates, our own Tammy Dungan spent time at the Association of Proposal Management Professionals (APMP) in 2016 presenting on the key role content plays in optimizing RFP responses.iv
Since Tammy’s 2016 APMP workshop and her related webinar series, we’ve continued to work with our most advanced customers to uncover best practices. For our customers, “winning” means producing fast, accurate and high-quality proposals that put their sales team in the finalist round.
Content discipline opens doors
The VP of Proposal Management at a large insurance company summed it up this way at our recent user conference.
“Despite inflated claims by some colleagues in the industry, a proposal team never wins a deal. It can lose a deal, but really never win one. For us, success means getting our company on the short list. And we get there with streamlined processes, a talented team, and hard work. But it all starts with content.”
Like this insurance company, our most advanced proposal management customers disproportionately make the short-list, allowing their sales team to pitch a finalist presentation. And they all do it with content. More accurately, our best-in-class customers practice the discipline of content management.
We refer to content management as a discipline in a strict dictionary definition sense of the word, where discipline means “a rule or system of rules governing conduct or activity.”v
Proposal management teams, and for that matter all the top-performing customer-facing teams we know, exercise discipline around content management. And, at the core of this discipline, sit six fundamental elements which we recall with the “STOMP-D” mnemonic: strategy, tools, organization, metrics, process and data structure. In the coming months, we will publish more on the discipline of content management.
Incidentally, my breakfast with Peter at Park Side Eatery in Saratoga ended with little fanfare. The track guys had long since gone to work and we both headed off to tackle immediate and practical customer problems. You see, both of us run companies with a “no bozos” policy.vi Absent bozos and the coffers of a huge company like Apple, Google, IBM or Salesforce, neither of us intended to contemplate AI again for the rest of the day.
i Eugene Kim, “Salesforce’s big new product ‘Einstein’ receives mixed reviews despite all the hype,” Business Insider, October 10, 2016.