October 10, 2023
In recent posts, I’ve written about Aspire’s ongoing study concerning the opportunities and challenges associated with the use of generative AI in customer communications and experience management. In this blog, I’ll offer an update on our progress and go over a few pointers buyers should consider when evaluating AI offerings.
At Aspire, we have spoken at length about how changing buyer personas have reshaped the trajectory of customer communications development. Centralized IT and operations have traditionally selected and purchased Customer Communications Management (CCM) solutions, but marketing and customer experience (CX) professionals have gained a stronger voice in communications and interactions management while buying power continues to shift to the line-of-business or digital/marketing teams who increasingly prefer SaaS solutions that enable business users to be in full control.
CX-leaders more likely to use generative AI
While we’re still conducting interviews and analyzing our research, an initial review of our findings reveals that organizations that are further along the maturity curve from simple, reactive CCM to more purposeful and conversational Customer Experience Management (CXM) report significantly more interest in leveraging generative AI in their customer communications. In addition, these mature businesses are more likely to source artificial intelligence from a third-party vendor that has integrated AI capabilities into its own solutions (as opposed to relying on internal IT to develop and train AI engines on their own.
This, of course, leads us to an obvious question: What do these organizations look for in their preferred vendor? I think there are a number of considerations they must keep in mind to be successful.
1. AI readiness: Most vendors are jumping on the AI bandwagon, but some have yet to build up a strong track record of success. In the CCM space, companies like Messagepoint have been talking about AI for many years. In fact, at a conference in 2018, I met Messagepoint’s AI lead, a PhD computer scientist tasked with building a team to create an engine (now called MARCIE) that would optimize content for CCM. That same year, Adobe announced that AI and data experts were its top hiring priorities. OpenText also has a long history of AI investments, though some of it is through acquisitions. I think it was 2015 when I saw its AI-engine IDOL (from Micro Focus) at work, when it was still part of HP (through the Autonomy acquisition).
2. CCM fit: This is particularly relevant for larger vendors like OpenText, Adobe, and Oracle, among others. Many of them have invested in AI or have acquired companies with a strong AI background, however, because of their large size and the wide variety of use cases they address, these AI endeavors may not prioritize CCM. Next week, OpenText will hold its World Event in Las Vegas, so we will soon learn how it will apply AI to CCM through its new offering, Aviator. (In my previous post, I explained how there are multiple areas where AI can make a difference in CCM. Check it out here if you’re interested.)
3. Depth and breadth of CCM-specific AI use cases: Many CCM vendors have invested in some form of AI. For example, Quadient has InspireXpress and Smart Communications comes with Migration Studio (AI solutions for semi-automatic template migration). Both vendors are also beginning to build content intelligence solutions into their offerings as well, but there are much wider opportunities for orchestration, inbound processing, personalization, and approvals. When choosing a new CCM software solution, we recommend clients pay specific attention to a vendor’s AI roadmap and strategy.
4. Speed to market: A key question to ask is:
Has this new technology has been released?
If so, have customers embraced it?
Even if they’ve released the new technology, many vendors may still be in beta, limiting the project’s scope to just a few select customers. While that’s normal for new developments, there are some vendors who handle tech development in more agile ways. Crawford Technologies, O’Neil Digital Solutions, Messagepoint, Adobe, and several others currently under NDA have or will soon launch generative AI applications in the area of CCM.
5. Private AI mode: In our research we found that sharing proprietary data with cloud-based AI engines is a top concern for enterprise executives. Personally identifiable information is a no-go for obvious reasons, but content fragments from existing webpages or communications are also considered sensitive intellectual property that should not fall into the wrong hands. Open technologies like ChatGPT learn from inputs and may feed that information back to other users, potentially creating risks to security and intellectual property. Many hyperscalers are actually building confidential or incognito modes, and in our research, we will explore how enterprises, technology vendors, and service providers can mitigate those concerns.
6. Productivity gains: This speaks to the power of the AI on offer. An AI engine that can point out language that’s difficult to read or outside brand guidelines can be useful, but a more powerful one will be able to rewrite the content into a more acceptable and accessible form. For business users, the ability to use a text prompt to generate documents or even code workflows is very powerful and offers tremendous productivity gains.
7. Commercial Model: This is a major focus of our research. Generative AI consumes a lot of computing power and is therefore an expensive technology. Will vendors absorb this cost into their offerings or will they begin charging for it? What would their enterprise customers prefer? If we look at the MarTech space, vendors like Hubspot have integrated ChatGPT at no additional cost (so far). The same goes for Word 365, where Microsoft provides insights and suggested rewrites as business users write. However, for more specialized solutions such as CCM, the potential labor savings may justify an additional software fee, especially since they have fewer customers over which to spread these costs.
As you can see, there’s still a lot left to uncover. If you’re interested in our research, please leave your details here and we will contact you once its released.