AI, Its Forecasted Development and Communication Future
AI's strengths, needs for communication improvement, narrowing gaps and what is to come
Artificial intelligence (AI) in business is constantly developing and increasing the capacity of what it can do and offers.
There are questions surrounding it that Communication Intelligence wanted to ask and learn about, such as what are 1) AI’s most notable strengths 2) what are its current shortcomings to be further developed 3) how can the gaps between what customers/clients/users want and expect and what is being offered to them now be narrowed or bridged and 4) where is the technology likely to be in 1, 3 and 5 years.
The answers to those questions are in this Communication Intelligence Special Report.
“The greatest strength of AI today,” says Erik Severinghaus, founder and CEO at Bloomfilter, which makes software development measurable and efficient, “is its ability to process large amounts of information and respond almost instantaneously.
“Whether it's answering customer service questions, helping a sales team prioritize leads or even managing routine emails, AI can handle simple, repetitive tasks, which often take a lot of time… when it's installed properly, it looks seamless.
“In fact, when a customer receives a personalized response in real time, the speed alone can create a sense of satisfaction that they don't even realize is powered by AI.”
The technology can make challenges less arduous.
“AI helps businesses get access to information in real time which makes problem solving much easier,” says Oindrila Mandal, a senior game product manager at Electronic Arts, a video game company.
“It can simplify and accelerate routine tasks such as note taking, summarization, search, data analysis, making presentations, content generation and more. Another great advantage of AI is that it can perform consistently and dependably with high levels of accuracy. This leads to consistent and predictable stakeholder interactions.”
There are shortcomings, areas of development where AI is not yet as strong as it can be and will be.
“One of its weaknesses, in my opinion, is actually the fault of humans and that is the fact that it is often used ineffectively,” says Edward Tian, the CEO at GPTZero, an artificial intelligence detection software to identify artificially generated text.
“It doesn’t need to be used for everything. You can send a quick chat to a coworker without putting a prompt into a generative AI platform in order to write out your message.
“It can be more valuable to personally write messages to your investors so that they don’t feel less prioritized. Sometimes using generative AI actually isn’t more efficient — or it can be negatively received.”
“AI is not without its limitations,” says Ghazenfer Mansoor, founder and CEO at Technology Rivers, a software development agency.
“One of the significant challenges is its inability to always grasp the emotional nuances of human communication.
“If you're feeling frustrated or down, AI might not pick up on that and could come across as a bit too robotic. AI can sometimes struggle with complex or multi-step problems. It's like using a calculator to solve a complex puzzle, it needs a more human touch to get it right.”
The talk about the emotions of communication is regularly brought up by sources.
“The primary shortcoming of today’s models for communication revolves around sentiment analysis,” says Jeff Oskin, the founder at NewLogiq, which helps small-to-midsized companies improve operations.
“The spoken or written human word has sentiment — emotion — embedded in the words or phrases that are chosen. Today’s AI models struggle to pick up on these nuances, which in turn degrades the experience of the recipient.”
“Despite its strengths, AI as a communication tool has several shortcomings or areas in need of development,” says Matt Rosen, founder and CEO at Allata, a company that helps clients to craft unique customer experiences, identify revenue generating opportunities, and improve operational efficiencies.
“One major limitation is its difficulty in understanding nuance. AI often struggles with grasping subtleties and emotional tones in communication, which can result in misunderstandings.
“Additionally, AI lacks deep contextual awareness, leading to responses that may be irrelevant or inappropriate. Its effectiveness is heavily dependent on data, requiring large volumes of high-quality data and if biased or poor-quality data is used, it can significantly affect performance.
“Furthermore, AI exhibits limited creativity, as it may struggle with creative problem-solving and generating novel ideas.”
The technology, while helpful and valuable, still requires work to become more widely respected, utilized and approved.
“AI still has a long way to go to truly communicate with humans on a deeper level,” Severinghaus says. “I mean sensitivity, nuance, context — what makes us human. Think about it: When someone is frustrated by a delayed order, they don't just want to know when it arrived, they want someone to acknowledge that the delay has ruined their plans. Currently, AI often fails to achieve this goal.”
Communicating factually is not enough when you are interacting with humans and their complexity of emotions and psychology.
“I once saw an AI that gave a technically correct answer, but looked like a robot,” Severinghaus says. “It does not have the ability to adapt to the emotional context of the situation. Artificial intelligence can't even detect subtle signals like tone or frustration in a voice.
“Of course, some systems are starting to look into this area with sentiment analysis, but it is still in its early stages.”
He commiserates with a too-common experience involving frustrating user interactions, which is a call for improvement.
“Another area where the AI fails to maintain a stable and ongoing relationship,” Severinghaus begins, adding that, “If you've ever had an AI-assisted interaction where you had to explain the same problem multiple times, you know what I'm talking about.
“AI often treats each interaction as an isolated event.”
Not everyone is at the point where they feel sufficiently comfortable with the technology to where they believe it will help them.
“It is still hard to completely trust AI, especially dealing with complex tasks or research,” says John Rich, founder at A Rich Opinion, a public relations and marketing firm. “To an extent, high-value outputs should be validated by a human,” he contends.
Impatience and annoyance can emerge with people when AI is considered to be missing the mark.
“Users can get frustrated when the AI can’t understand them in the same way a human would, especially in scenarios that require a more thoughtful or creative approach,” says Aaron White, CEO at Outbound.com, a marketing platform that utilizes AI to automate advertising for small businesses. “That’s an area where I think we’ll see a lot of improvement.”
“The major shortcomings of current implementations of AI is that it works in limited context. AI use cases typically are small and very specific,” Mandal says.
“While AI is great at completing the specific task it was trained for under the ideal conditions of operation, it fails when these conditions are not met or when the request deviates even slightly from what the AI knows.”
She proposes a response.
“One clear area of improvement for AI is creating algorithms and solutions that are widely applicable or that can be easily plugged into a new business context,” Mandal says.
There are additional challenges with the technology.
“AI cannot think, it merely does what it was trained to do,” Mandal points out. “This is why in Generative AI we see the issue of AI hallucination where the AI does not know the exact answer so it starts generating random, incorrect output.”
It is important to understand what can happen — and what remains necessary for better or ideal outcomes.
“Hence for communication use cases, AI might get stuck if complex decision-making is involved,” Mandal reminds leaders. “AI still needs a human in the loop to address the edge cases in communication solutions for stakeholders.”
Price remains an obstacle.
“Another shortcoming is the cost of AI. AI compute and infrastructure costs create a high barrier to adoption,” Mandal says. “Resources must become more affordable to be able to generate a meaningful ROI so that businesses can justify AI adoption.”
The question that reasonably emerges is what needs to be done with to where the gaps are narrowed between what stakeholders expect of AI as far as fullness and thoroughness of communication.
“Laws and regulations need to be created regarding AI usage. We are very behind in this regard,” Tian says. “Until we know what guidelines we can safely operate within when it comes to AI usage, many businesses and stakeholders are going to be apprehensive about using AI.”
It’s important to know to go to the root to work on the tasks to drive improvements.
“First, we must work on language processing (NLP). This makes AI (better) at information, not just the phrases humans use, but also the emotions at the back of them,” Mansoor writes. “It could be notable if AI could not only deliver accurate answers but also understand that you're having a tough day.
“Another aspect is making AI more remarkable adaptable. Right now, it can get thrown off via sudden conditions, so we need to make it more flexible and capable of coping with a broader range of interactions.”
Not everyone is a believer that full AI is the answer in business communications.
“I think the crux of person-to-person communications will remain inherently human,” Rich asserts. “Unless explicitly stated, as soon as a customer finds out he or she has been communicating with a robot, they will lose trust in that brand. I think creating more efficiency in conversations is where AI communications tools will go.”
There could be a deeper issue to tackle, at the foundation of the technology.
“While the AI models will continue to improve and become more sophisticated, the major limiter for most organizations is the cleanliness and centralization of their data,” Oskin states.
“Nearly all models deployed in a corporate setting to support interactive communication require access to data from which the models learn how to behave and interact.
“Voids or inaccurate data will cause the models, no matter how sophisticated, to miss expectations as they will produce inaccurate or misleading responses to user queries.”
Returning to how humans process and function is a critical area for development.
“To narrow the gaps in AI communication, it is essential to enhance its understanding of context and emotion by developing systems with improved natural language processing capabilities,” Rosen says. “This involves creating more conversational AI that supports dynamic, two-way interactions with humans.”
He goes deeper to discuss a core responsibility
“Ethical data use should be prioritized to ensure AI models are trained on diverse and unbiased datasets, thereby improving their reliability and fairness,” Rosen advises.
“Encouraging cross-disciplinary collaboration among AI developers, linguists, psychologists, and ethicists can contribute to crafting a more human-like AI communication experience.”
Severinghaus advocates for it too.
“We need more advanced emotional AI,” Severinghaus advocates. “The technology can begin to recognize not only what a person says, but also how they feel when they say it. Think of it as AI gaining emotional intelligence.
“The more AI understands feelings —whether a person is happy, angry or confused — the more effectively it can respond in a way that feels human.”
He believes the macro is as important in its own right as the micro.
“Beyond emotional intelligence, AI must better see the big picture,” Severinghaus states. “It's not just about answering questions once but about creating a conversation that extends over time. AI systems must use a broader context, ranging from a person's previous interactions to their long-term relationship with a brand or company.”
The reasoning is clear.
“In this way, the AI not only deals with this problem, but also takes into account the relationship and the general history,” he says.
Fluidity should come with time and skill. It needs to do so.
“AI is still pretty rigid,” White contends. “It works best when it follows predefined scripts or algorithms. To meet stakeholder expectations, it needs to become more flexible and capable of handling unexpected or nuanced situations.
“This will likely involve advancements in natural language processing and emotional intelligence within AI systems, so they can better gauge a user’s intent and provide more meaningful, human-like responses.”
People should not excluded even when the technology becomes advanced.
“One way to narrow the gap between stakeholder expectations is to build AI solutions with a human-in-the-loop where a human expert can step in when the AI performance starts deteriorating,” Mandal advises.
“Another way is to ensure the AI is trained on large volumes of high-quality contextual data that is clean, accurate and labeled effectively so that the AI solution can answer all edge cases.”
Companies should remain engaged with users, at least presently.
“An interim solution is to also provide training to customers, clients and business stakeholders on how to use the AI and get the most effective answers from it.”
The near future could bring rapid improvement and capacity of function.
“I strongly believe AI will become the de-facto mode of customer communication in the next 5 years,” Mandal says. “With AI models becoming more advanced and compute infrastructure becoming inexpensive, AI will get much better at addressing customer and client concerns.
“There will be less and less need for human-in-the-loop solutions as organizations improve data quality and labeling. I think AI communication solutions will be able to satisfactorily address most edge cases and complex customer problems within the next 3-5 years.”
Other sources who spoke to C.I. see similar progress coming.
“In the next year, AI is expected to better understand context and provide more relevant responses. It will be more like conversing with someone who truly knows you,” Mansoor forecasts.
“In three years, AI might even start exhibiting more empathy, thanks to advancements in emotional intelligence. Imagine an AI that not only answers your questions but also empathizes with how you feel.
“And in five years, AI could handle complex conversations almost as adeptly as a human. It will be able to tackle specific issues while still maintaining a personal and engaging tone,” he adds.
As this happens, there will a changing out of technology.
“Over the course of the next year or two, leading organizations will look to replace existing chatbot or IVR technology with AI chatbots or AI voicebots,” Oskin says, going on to explain the reasoning and impetus behind it.
“These tools are inherently more flexible than traditional chatbot-IVR technology that required each question and resulting response to be scripted. In this paradigm, if a user asked a question that was unknown to the technology, it would not be able to respond, which in turn frustrated users.
“AI chatbots or AI voicebots differ in that they are completely unstructured and therefore do not require scripted questions for each and every scenario a customer may pose. Rather these tools simply require access to historical data such as ticketing systems and from the data are able to dynamically respond to user queries.”
Within three years, Oskin predict there will be further evolution within workplaces.
“AI technologies will mature to the degree that they sit ‘side-by-side’ with employees, also known as AI teaming,” Oskin says.
“This collaboration will aim at enhancing productivity, where AI handles routine inquiries, data processing and provides real-time insights, allowing human employees to focus on more complex problem-solving and personal customer interactions.”
Further down the road, he sees even more assistance for businesses and their people.
“AI will be able to take on full business processes such as sales and sales demonstrations,” Oskin says.
“Imagine the sale of software where the entire experience, including demonstrating user-specific scenarios, proposal development, contract negotiations, etc. is handled by an AI agent. Similar approaches can be done for other industries,” he details.
Communication that now may seem inadequate or poor will develop to an acceptable or excellent level.
“Within a year, we can expect improved chatbots that offer more natural and contextual interactions, along with better integration into customer service workflows, as well as enhanced analytics tools providing more predictive insights and automated reporting,” Rosen says.
“Looking ahead to three years, AI is anticipated to achieve advanced personalization through refined communications tailored by real-time data analysis and predictive behavior models, combined with seamless integration across various business platforms to enhance operational efficiency and customer experience.
“In five years, AI is projected to reach augmented communication capabilities, assisting in real-time translation, summarization and complex report generation with minimal human input, while also attaining near-human levels of understanding and empathy, achieving human-like interactions with improved creativity in communications,” he adds.
Potential is significant.
“I think we are on the verge of something really exciting,” Severinghaus says.
“In the coming year, AI will become more conversational. It will be able to answer more complex questions and anticipate needs better than today.
“In three years, we will probably start to see AI that can detect signals. Imagine a customer interaction service where AI can not only detect frustration, but also change its tone accordingly, giving a more empathetic response when things go wrong, so if you are faced with a more complex problem like scheduling an appointment or solving an invoice error, the AI will be able to handle it more easily.”
“I think that in five years, AI will begin to blur the line between human and machine interaction. We will see hybrid systems in which AI takes care of routine tasks and humans intervene when more complex or emotional situations arise.
“And in some cases, AI can even outperform humans in data analysis and analytics. It wont just react to problems; it proactively offers you solutions before you know there is a problem.”
As much as A.I. offers currently, progress could prove exponential.
“We are only scratching the surface of what is possible,” Severinghaus says. “The technology is efficient and scalable but it lacks some essential elements that make human communication so powerful: empathy, context and continuity.
“In the coming years, I am convinced that we will close these gaps and that AI will not only meet our expectations, but exceed them, becoming an essential part of how companies interact with customers and their stakeholders.
“As exciting as this journey has been so far, I am confident that the best is yet to come.”
To sponsor this publication, place an ad or have links published, check this page.
Follow on the LinkedIn company page for Communication Intelligence