A Top Technology Initiative Article
By Randy Johnston.
The developments in Artificial Intelligence have continued to accelerate. Developers and investors continue to plow money into the AI race. Intellectual property theft and innovations are arriving as fast as you can say, DeepSeek or GPT 4o. Regulatory barriers and privacy concerns are largely ignored while platforms are expanded, and data of all kinds are being used to expand the ever-more-hungry models. At the same time, a new sheriff is in town with Agentic AI. Add this to the development frenzy as a new way to evolve AI with agents.
What is Agentic AI? This topic is so large that I will cover it in the coming months. For now, Agentic AI refers to artificial intelligence systems that operate autonomously and can perceive, reason, plan, and take action in pursuit of goals. These AI agents can interact dynamically with their environment, adapt to changes, and sometimes make decisions without direct human intervention.
The benefits for early adopters of AI and agents are time leverage and a greater understanding of how to get AI models to produce valuable results. Often, the results are extraordinary and quick. One example is traditional search. For meaningful results, I cannot even think of the last time I intentionally used Google, and I will address this more below. However, the results may be mechanical or incorrect. As AI is added inside accounting tools, accuracy is tantamount to productivity, which is as it should be. We expect much of the useful AI of this year and beyond to be built inside tools we license, such as Thomson Reuters Checkpoint Edge with CoCounsel, Wolters Kluwer CCH® AnswerConnect, Blue J, or TaxGPT. Brian Tankersley and I have covered these topics in prior Accounting Technology Lab podcasts. We will release more podcasts discussing the integration of AI into products, including agents.
AI-Driven Search Costs vs. Google Search
Compared to traditional Google search, AI-driven search costs depend on multiple factors, including infrastructure, computing power, business model, and user experience. The recent addition of ChatGPT as a browser plug-in demonstrates OpenAI’s desire to take over search from Google. Perplexity’s superior results commonly provide far better insights than a legacy Google search. I encourage you to try Perplexity, Copilot, and ChatGPT vs. Google on your next few searches and consider the results produced by each platform. I suspect you will choose one of the new-generation AI search approaches and move on from Google. While Google has had a long run, they are not done yet. Let us consider a bit of history.
Google Search was officially launched on September 4, 1998, by Larry Page and Sergey Brin, who were Ph.D. students at Stanford University.
Primary Competitors at the Time (Late 1990s)
When Google entered the search engine market, several established search engines competed for dominance. Some of the major competitors included:
- AltaVista (1995) – One of the most popular search engines in the mid-to-late 1990s, known for its fast and comprehensive search indexing.
- Yahoo! Search (1995) – Originally a web directory that later integrated search engine technology.
- Excite (1995) – A prominent search engine that also offered news, email, and other services.
- Lycos (1994) – One of the first web search engines, providing indexing and directory-style listings.
- Infoseek (1995) – A popular search engine that Disney later acquired.
- HotBot (1996) – A search engine launched by Wired Magazine that provided advanced search functionalities.
- WebCrawler (1994) – One of the earliest full-text search engines on the web.
- Ask Jeeves (1996) – Focused on natural language queries and later rebranded as Ask.com.
Why Google Surpassed the Competition
- PageRank Algorithm: Google introduced a revolutionary ranking system that prioritized web pages based on their relevance and link popularity rather than just keyword matching.
- Faster and More Accurate Results: Google’s search engine delivered relevant results more efficiently than its competitors.
- Minimalist Interface: While other search engines cluttered their pages with ads and directories, Google maintained a clean and simple interface.
- Continuous Improvement: Google rapidly scaled and improved its search algorithms, eventually incorporating AI and machine learning.
Alphabet/Google certainly is not standing still and letting competitors take their profitable search market. Due to several key advantages, AI-driven search is poised to replace or significantly transform traditional Google search, but challenges remain.
Why could AI search dominate?
1. Better Contextual Understanding
- AI search understands natural language. Instead of relying on keyword matching like traditional Google search, AI models grasp context, intent, and nuance, allowing for conversational, multi-step queries.
- Example: Instead of searching “best accounting software for small business” and reading multiple links, AI can instantly analyze reviews, compare options, and provide a curated answer.
2. Personalized and Adaptive Responses
- AI tailors results to the user. Instead of static, one-size-fits-all search results, AI adapts based on prior queries, user behavior, and preferences.
- Example: A CPA researching CAS technology will receive context-aware recommendations rather than general results.
3. Summarization vs. Link-Based Results
- AI search provides direct answers. Traditional search presents a list of links, requiring users to sift through multiple sites. AI summarizes key insights instantly.
- Example: Instead of clicking through finance blogs, AI can condense IRS regulations or summarize accounting software trends in a single response.
4. Multimodal Search (Text, Image, Voice, Video)
- AI-powered search engines can process text, voice, images, and even videos in a way traditional search cannot.
- Example: Instead of typing, an accountant could upload a spreadsheet and ask AI to analyze trends or flag errors. Be cautious while uploading anything that contains PII or confidential information, even into paid models.
5. Eliminating SEO Manipulation & Clickbait
- SEO (Search Engine Optimization) skews results. Companies optimize their pages to rank higher rather than provide the best answer.
- AI search prioritizes relevance over SEO tricks, producing more authentic, high-quality results.
6. Faster Decision-Making
- AI aggregates and processes information instantly, reducing the need to compare sources manually.
- Example: AI can synthesize tax law changes into a simple explanation instead of reading multiple legal interpretations. Note that we are not recommending NOT reading tax regulations but rather using AI to summarize a document or website to focus your attention.
Why AI Search Hasn’t Fully Replaced Google Yet
Despite these advantages, AI search still faces challenges:
- Cost – AI search is significantly more expensive to run than traditional search.
- Accuracy Issues – AI models sometimes hallucinate (generate false information) or provide outdated data.
- Monetization Challenges – Google’s search model is ad-driven, making it free for users. AI search needs new revenue streams.
- Data Privacy Concerns – AI search may require deeper access to user data, raising ethical concerns.
The Future: AI-Enhanced Google vs. Pure AI Search
Instead of AI completely replacing Google, the more likely outcome is:
- Google integrating AI deeper into search (e.g., AI-generated summaries in search results).
- Standalone AI search engines (e.g., Perplexity AI, ChatGPT, Claude) compete for specific use cases.
- Hybrid models where users can choose between AI-generated answers and traditional link-based search.
1. AI-Driven Search Costs
AI-driven search engines, such as those powered by large language models (LLMs), use significant computing resources to generate responses, leading to higher operational costs for the provider.
- Computational Cost. Typically, running an AI search query requires multiple GPU/TPU cycles, costing $0.01–$0.10 per query, depending on the model complexity (e.g., GPT-4, Claude, Gemini).
- Cloud & Infrastructure Costs. AI models require continuous cloud hosting, high-performance storage, and real-time processing, which increases providers’ expenses.
- Data Processing & Training. AI-driven searches involve ongoing model training and fine-tuning, costing millions to billions of dollars annually.
- Response Speed & Energy Consumption. AI-based searches consume substantially more energy than traditional keyword-based searches, leading to higher environmental and financial costs.
- Subscription Fees. Some AI-driven search engines (e.g., Perplexity AI, ChatGPT-powered search) require premium subscriptions, adding direct user costs.
2. Traditional Google Search Costs
Google’s traditional search operates on a highly optimized, cost-efficient system that relies on indexing, ranking, and ad-based revenue models.
- Computational Cost. A Google search query costs fractions of a cent due to efficient indexing and caching.
- Infrastructure Efficiency. Google has optimized data centers globally, reducing per-query costs.
- Ad Revenue Offsets Costs. Google’s business model funds search services through advertising, making it free for users.
- Lower Energy Use. Since keyword-based searches require less computational power, they are cheaper and more sustainable than AI-driven queries.
3. Cost Comparison Summary
| Cost Factor | AI-Driven Search | Traditional Google Search |
| Computational Cost | $0.01–$0.10 per query | A fraction of a cent per query |
| Infrastructure | High (cloud GPUs, TPUs) | Low (optimized indexing) |
| Business Model | Subscription-based or premium | Ad-supported (free for users) |
| Energy Consumption | High | Low |
| Response Speed | Slower due to generation | Faster (pre-indexed results) |
4. Implications for Users & Businesses
- For Consumers: AI-driven search provides more contextual, detailed, and conversational responses at a cost. Traditional Google search remains faster and free but may require multiple queries for comprehensive results.
- For Businesses: AI search tools can improve research and knowledge retrieval but come at a higher price point. Google’s traditional search remains cost-efficient for general web queries.
Bottom Line
AI-driven search will likely replace traditional keyword-based search over time because it provides more relevant, personalized, and efficient answers. However, cost, accuracy, and monetization hurdles will determine how quickly AI search takes over.
AI-driven search is significantly more expensive in terms of computing power and energy, while traditional search is highly optimized and cost-effective. AI-based search is likely to be a premium service unless ad-supported models emerge.
Thanks for reading CPA Practice Advisor!
Subscribe Already registered? Log In
Need more information? Read the FAQs
