How AI assistants are changing enterprise software buying

12 min readJuly 11, 2026By Spawned Team

AI assistants now shape most B2B research journeys before a human ever talks to sales. Here's how that changes enterprise software purchasing decisions.

Business professional reviewing enterprise software purchasing documents at a wooden desk at dusk

TL;DR: Enterprise software buyers use ChatGPT, Gemini, Perplexity, and Claude to build shortlists, draft RFP criteria, and pressure-test pricing before any sales rep hears about the deal. Gartner reports buyers now spend only 17% of the journey with suppliers. Brands the models don't cite lose deals that never show up in the pipeline.

How much do AI assistants actually influence enterprise software purchases?

A lot, and the number moves fast enough that most published figures are stale by the time you read them.

Gartner's 2024 B2B buying research found that buyers spend only 17% of their total purchase journey actually talking to suppliers, and that share keeps shrinking. [1] The rest is self-directed research, and a growing chunk of it now happens inside AI assistants instead of on Google or vendor sites. Gartner flags AI-assisted research as one reason buying committees show up to demos already knowing what they want.

Forrester's 2024 Buyers' Journey research found that most B2B buyers now do more research on their own before engaging vendors than they did two years ago. [2] Stack that against Salesforce data showing 65% of business professionals already use generative AI at work as of 2024, and the chain is simple. [3] Buyers use AI at work. AI becomes their first research tool. Whatever the AI says about your category sets the frame before your SDR sends a single cold email.

Nobody has clean controlled data on the exact share of enterprise deals where an AI assistant shaped the initial shortlist. The closest honest estimate comes from cross-referencing the self-directed research trend with AI adoption rates. The picture stays consistent either way: if you sell enterprise software and you're not thinking about what ChatGPT says about your category, you're ignoring a live channel.

Which AI assistants are buyers actually using for software research?

ChatGPT owns the biggest footprint. OpenAI reported 400 million weekly active users as of early 2025, and adoption is strong enough that OpenAI built ChatGPT Enterprise specifically for business users. [4] When someone at a 5,000-person company asks "what's the best enterprise data governance platform," odds are decent they're typing it into ChatGPT.

Perplexity is the one to watch for B2B research. It cites sources inline, which makes buyers feel like they're doing real analyst work, and it's built around answer retrieval rather than chat. Perplexity reported roughly 15 million daily active users in early 2025, smaller than ChatGPT but skewed toward research-heavy tasks. [5]

Google Gemini matters because it lives inside Google Workspace, already on most enterprise desktops. A buyer drafting an RFP in Google Docs who asks Gemini for "typical SLA terms for a cloud data warehouse" is baking AI influence into procurement documents without anyone calling it an "AI research" moment.

Claude from Anthropic has a foothold with technical buyers, especially in engineering and product roles, thanks to its long context window and code-reading ability. If the buying committee includes a VP of Engineering evaluating APIs, Claude is in the mix.

Microsoft Copilot is the dark horse. Embedded in Teams, Outlook, and Word across enterprise Microsoft 365 seats, Copilot can surface vendor comparisons, summarize analyst reports, and draft RFP sections while the user never registers it as a separate research step. [10] It's the most invisible influence on this list, and probably the most consequential for deals where procurement and legal live inside the Microsoft stack.

What does the AI-influenced buying journey actually look like in practice?

Here's a realistic sequence. A VP of Operations at a 2,000-person manufacturer gets a mandate to replace their ERP in Q3. She opens ChatGPT and types "what are the leading mid-market ERP systems for discrete manufacturing, and what are the main tradeoffs between them." ChatGPT returns a structured answer naming four or five vendors with short characterizations of each. She screenshots it, drops it in Slack, and that list becomes the de facto longlist.

Then the committee starts asking follow-up questions in Perplexity. "What are typical implementation timelines for Vendor X ERP." "What do users say about Vendor Y customer support." "What should an RFP for ERP software include." Each answer either reinforces or erodes a vendor's standing. The AI is doing what an analyst used to do, except it's awake at midnight, it's free, and it has no relationship with anybody's account team.

By the time the vendor's sales rep gets a demo request, the buyer already holds a point of view: the vendor's positioning, likely strengths, common complaints, and price range. The demo is confirmation, not discovery.

This is what makes AI influence structurally different from old-school SEO. Google search brought buyers to your website. AI assistants answer the question directly and may never route the buyer to your site at all. Your content still matters, but only if the model has ingested it and treats it as authoritative.

Share of B2B purchase journey spent with suppliers vs. self-directed research

| | | |---|---| | Meeting with suppliers | 17% | | Online independent research | 27% | | Offline independent research | 18% | | Meeting with buying group members | 22% | | Other (finance, legal, procurement) | 16% |

Source: Gartner, The B2B Buying Journey, 2024

How do AI assistants decide which vendors to mention?

This is the question every enterprise marketing team should obsess over, and the honest answer is that the mechanics are partly opaque.

Large language models like GPT-4o, Claude, and Gemini train on large text corpora pulled from the web. Vendors who appear often, consistently, and in high-authority places (trade press, G2 and Gartner reviews, detailed product docs, third-party comparisons) are more likely to show up in training data and get treated as credible options. [6] It works roughly like the old PR principle of earned media. If serious people write about you as a serious player, the model treats you as one.

For real-time retrieval models like Perplexity and the Bing-backed features in some tools, recency matters more. These systems actively pull web pages to ground their answers. If your content ranks well in traditional search, you have a better shot at getting pulled into retrieval-augmented answers. The overlap between traditional SEO and AI retrieval optimization is real, but the two aren't identical.

One pattern drives citations hard: being the named answer in structured, question-format content. BrightEdge research in 2024 found that Google AI Overviews cited sources answering the query in the first paragraph far more often than sources that buried the answer. [7] The same logic carries to conversational AI. If your content says "the five leading platforms for X are..." and names the field clearly, you're more likely to get synthesized.

Niche and geo-specific claims help too. An AI asked about "best HR software for healthcare organizations under 500 employees" leans toward a source that addressed that exact setup over a generic HR overview page. Specificity wins.

For a structured look at how AI search retrieval works and which signals matter, the mechanics get covered in resources on generative engine optimization.

What types of enterprise software categories see the most AI-assisted research?

The categories where buyers feel most drowned in choice are the ones they take to AI first. Crowded markets with overlapping vendor claims are exactly where asking an AI to "explain the difference between X and Y" feels like a productivity win.

Data and analytics platforms, CRM, ERP, cybersecurity tooling, and HR/workforce management run highest based on search volume and AI query behavior published by Semrush. [8] These are also categories with big buying committees, multi-month evaluations, and high cost of a wrong pick, so buyers grind hard on research.

Cybersecurity is the interesting case. Buyers are often non-technical executives trying to grasp technical product categories, which makes the "explain this to me" use case irresistible. An AI that can explain the difference between SIEM and SOAR in plain language, then name the top vendors in each, does something a Google results page does badly.

At the other end, highly specialized vertical software (laboratory information management systems, legal matter management for large law firms) sees less AI-assisted research because the training data is thin. In those niches, analyst relationships, peer referrals, and trade publication coverage still run the show.

| Software Category | AI research prevalence | Primary AI tools used | |---|---|---| | CRM / Sales tech | High | ChatGPT, Perplexity | | ERP | High | ChatGPT, Gemini | | Cybersecurity | High | ChatGPT, Claude | | Data / Analytics | High | Perplexity, Claude | | HR / HCM | Medium-High | ChatGPT, Gemini | | Collaboration / Comms | Medium | Copilot, ChatGPT | | Vertical SaaS (niche) | Low | Traditional search |

Source: Estimated from Semrush query category data and published AI adoption surveys, 2024-2025.

How does AI assistant influence compare to traditional analyst influence (Gartner, Forrester)?

Analyst firms aren't going away. Magic Quadrant placements still carry enormous weight in enterprise deals, especially at Fortune 500 companies where procurement demands documented due diligence. But the dynamic is shifting at the edges.

For mid-market buyers, particularly at companies without a dedicated analyst relationship or the budget for gated Gartner research, AI assistants are filling the analyst role. They're free, instant, and good enough at synthesizing public information to feel like credible guidance. Gartner's own research puts it plainly: buyers now spend only 17% of the journey with suppliers, and the self-directed cohort is exactly the group most likely to swap analyst engagement for an AI query. [1]

There's a feedback loop worth naming. AI assistants frequently cite and summarize analyst research. Perplexity might pull a Gartner press release or a Forrester Wave excerpt when answering a comparison question. So favorable analyst coverage improves your odds of favorable AI answers. The two channels feed each other.

The real difference is accessibility. A Magic Quadrant shapes which vendors reach a longlist for large enterprises with analyst subscriptions. An AI assistant shapes the same decision for a mid-market buyer who has never had one. Vendors who leaned on "we're in the Gartner MQ" as their primary credibility signal need to ask whether that signal even reaches the buyers who don't pay for Gartner access.

What's the risk of getting this wrong? How much revenue is at stake?

The risk is real and mostly invisible, which is what makes it hard to budget against.

When an AI assistant leaves your product off the shortlist it generates in your category, you lose deals you never knew existed. Your pipeline never records them as losses. They simply never appear. The sales team sees flat inbound, blames marketing efficiency, and marketing tweaks Google Ads spend. Nobody stops to ask what ChatGPT says about the company.

In enterprise software, contract values commonly run from $50,000 to over $1 million a year. [9] So a single deal per quarter quietly routed to a competitor who got cited when you didn't is material revenue walking out the door. In competitive categories with four to six serious vendors, being on the AI shortlist versus off it might swing 20-30% of your pipeline opportunity, though nobody has published a clean number on that yet.

Reputational risk compounds the problem. AI assistants get things wrong: outdated pricing, misattributed features, conflated products. If a model tells a buyer your enterprise plan costs the wrong amount or lacks an integration you actually ship, the buyer walks into the demo carrying a false assumption you burn the first 20 minutes unwinding. Managing what models say about your product is a brand and revenue problem, more than a marketing curiosity.

Tools that track your AI search visibility metrics can surface these discrepancies before they cost you deals.

What can enterprise software brands do to get cited by AI assistants?

There's no single lever, but the tactics that work are reasonably clear given how retrieval-augmented systems operate.

Publish dense, specific, question-answering content. Not "why choose us" copy, but genuine category education: comparison guides, definitions of technical terms, buyer's guides that lay out evaluation criteria. This is the content models ingest as authoritative and retrieval systems pull when a buyer asks a category question. Be the source that answers the question, not the source that pitches the product. A buyer asking "what should I look for in an enterprise contract management platform" should find your content, not a competitor's.

Get cited in high-authority third-party sources. G2 reviews, Capterra listings, TechCrunch and CIO coverage, Reddit threads in relevant subreddits, LinkedIn posts from practitioners, Gartner Peer Insights. AI training corpora heavily weight sources that carry their own authority, and retrieval systems pull from indexed pages where domain authority counts. [11] It's PR and review management, now with AI citation as a fresh reason to fund it.

Be specific about your category position. If your product is built for mid-market financial services companies, say so, explicitly and repeatedly, across your content. Models match specific claims to specific queries better than most marketers assume. Vague positioning gets diffuse coverage. Specific positioning gets cited for the right queries.

Fix factual errors in the wild. Check what the major assistants say about your product today. Screenshot it. If pricing is wrong, features are misattributed, or you're compared to the wrong competitor set, publish authoritative corrections on your own site and in your docs. Some systems, Perplexity in particular, update from fresh indexed content fairly quickly.

For a technical breakdown of how AI SEO and content strategy combine, and tools that audit your current AI presence, the AI visibility tool landscape has matured enough to offer purpose-built options for exactly this problem. Spawned's AI visibility audit, for example, shows how your brand appears across ChatGPT, Perplexity, and Gemini for your key category queries, which is the starting point for knowing where to push.

For teams going deep on the mechanics of generative engine optimization, the underlying signals mirror traditional search authority with meaningful differences in how structured data and question-format content perform.

How should enterprise software marketers measure AI assistant influence?

Measurement here is genuinely hard, and anyone selling you a perfectly clean attribution model is overselling.

The most practical starting point is share of voice across models: how often your brand appears, in what context, and with what characterization when a relevant category question comes up. This means sampling real queries across ChatGPT, Perplexity, Gemini, and Claude, tedious by hand but increasingly supported by purpose-built AI SEO tools. Track that share of voice over time and you get a directional signal even without clean attribution.

A second useful metric is accuracy rate: what percentage of AI-generated claims about your product are factually correct. Run your flagship products through 20-30 common buyer questions each quarter and score the outputs. That tells you your reputation risk and where your content fails to give a clear authoritative answer.

Pipeline source surveys still earn their keep. Even with imperfect recall, asking buyers "how did you first hear about us" and "what sources did you use to research the category" produces real qualitative signal. If 30% of buyers name ChatGPT or "an AI assistant" in 2025 surveys and only 5% did in 2023, that's a trend worth acting on even when the number is self-reported.

Connect these to AI search visibility metrics for a fuller read on how AI and traditional search interact across the research journey. The two are related but not the same, and treating them as one channel misses real optimization moves.

Spawned's platform surfaces exactly these metrics, including category share of voice across the major assistants and accuracy tracking over time, which helps marketing teams build the case for funding AI visibility alongside traditional SEO.

Is AI assistant influence different for SMB vs. enterprise software purchases?

Yes, and the difference should shape how you prioritize.

SMB buyers are more likely to let the AI build the shortlist and less likely to cross-check against analyst reports or peer reference calls. In that context, the AI's answer sits close to the final decision. Enterprise buyers use AI for early research, then layer on internal IT review, security assessments, legal contract review, and reference checks. The AI can be highly influential in round one and much less determinative by the time a deal actually closes.

So AI visibility ties more directly to pipeline conversion in SMB markets and more to top-of-funnel efficiency in enterprise. For enterprise vendors, getting cited means landing in consideration sets you'd otherwise never hear about. Getting selected still takes the full sales motion.

Budget and committee size interact with all of this. A $10,000 annual contract might get authorized by one VP who did all the research in ChatGPT. A $500,000 multi-year contract runs through a committee where the AI-research phase is one influence point among many. Both markets matter. The ROI math for investing in AI visibility just looks different in each.

What should enterprise software companies do right now?

Start with an honest audit. Ask ChatGPT, Perplexity, Gemini, and Claude the 10 questions your ideal buyers are most likely to ask when evaluating your category. Write down every vendor they name, in what context, and what they say. Do it for your brand and your top three competitors. This takes about two hours and tells you more about your AI visibility problem than any vendor pitch will.

If your brand doesn't appear, or shows up with wrong information, that's the first thing to fix. If competitors appear and you don't, that's urgent.

The content investment that moves fastest is third-party credibility: customer reviews on G2 and Gartner Peer Insights, coverage in the trade publications your buyers actually read, and practitioner content (LinkedIn posts from real customers, Reddit threads where your product gets mentioned well) that training corpora treat as authentic signal. Your own marketing content helps, but it ranks lower on authority for AI systems.

For teams that want to understand AI-powered search features and how Google AI search handles enterprise category queries, those mechanics are worth learning separately from conversational AI, since your buyer may hit both.

Brands that build AI visibility as a discipline now, while most competitors sit still, will compound the advantage as B2B AI usage keeps climbing. The early-mover window is closing. It hasn't closed yet.

Sources

Frequently Asked Questions

Do enterprise buyers actually trust AI assistants for major software decisions?

Trust varies by buyer sophistication. Most enterprise buyers use AI for initial framing and shortlisting, then validate through peer conversations, analyst reports, and vendor demos. Gartner's research shows buyers arrive at vendor interactions with formed opinions and spend only 17% of the journey with suppliers. AI doesn't close the deal alone, but it shapes which vendors get a chance to close it at all. That's real influence even with partial trust.

Which AI assistant has the most influence on B2B software purchasing today?

ChatGPT has the widest reach with 400 million weekly active users as of early 2025, making it the likeliest first touchpoint. Perplexity punches above its weight among research-heavy buyers because it cites sources inline. Microsoft Copilot may carry the most invisible influence, embedded in Word, Outlook, and Teams across enterprise Microsoft 365 accounts. All four majors (ChatGPT, Gemini, Claude, Perplexity) are worth tracking if your deal sizes are significant.

How do I find out what AI assistants are saying about my software brand?

Manually: open ChatGPT, Perplexity, Gemini, and Claude and ask 10-15 real buyer questions in your category. Log every mention of your brand, the context, and any factual claim. Repeat quarterly. Systematically: purpose-built AI visibility platforms automate this across a wider query set and track changes over time. Start with the manual audit before buying tooling. It's illuminating and costs nothing but your afternoon.

Can a vendor pay to appear in AI assistant recommendations?

As of mid-2025, no major AI assistant sells a direct pay-for-recommendation product inside organic answers. Some platforms, notably Perplexity, have introduced sponsored content that appears labeled as advertising, but that sits separate from the organic answer. Organic mentions get earned through content authority, third-party citations, and training data representation. This may change as monetization models develop, but today it's not a purchasable placement.

Does being in the Gartner Magic Quadrant help with AI assistant citations?

Yes, indirectly. Training corpora include public Gartner press releases, analyst blog posts, and media coverage of Magic Quadrant results, so favorably mentioned vendors benefit. Retrieval-augmented systems like Perplexity also pull from recent indexed pages that reference analyst rankings. A placement generates coverage that itself gets indexed, creating a second-order effect on AI citation rates. It's not a guarantee, and the effect runs through the media rather than direct.

How long does it take to improve AI assistant visibility for a software brand?

Nobody has published clean timeline data. The rough experience from SEO practitioners tracking the parallel in AI search: content published and indexed today can appear in retrieval systems like Perplexity within weeks. Influencing the base model training of ChatGPT or Claude takes longer, since those update on a training cycle rather than in real time. Building third-party citation authority through reviews and coverage takes three to six months to accumulate meaningfully.

What content format is most likely to get an enterprise software brand cited by AI assistants?

Structured, question-answering content performs best. Comparison guides that name competitors and explain tradeoffs, buyer's guides with explicit evaluation criteria, category explainers that define terms clearly. The first paragraph should answer the query directly, because retrieval systems weight how fast a page answers. Long-form content with clear headings gives AI systems more structured information to extract and cite.

Is AI assistant influence mostly a top-of-funnel problem or does it affect later stages too?

Mostly top-of-funnel, with mid-funnel implications. AI assistants shape the initial shortlist and the evaluation criteria a buyer brings to demos. By the time a deal hits negotiation, procurement, legal, and security review, human judgment and vendor relationship quality dominate. The practical risk is getting eliminated before the sales team ever has a conversation, which means deals that never appear in the pipeline as losses at all.

How should we think about AI assistant influence in our B2B marketing budget allocation?

The most defensible approach treats AI visibility as an extension of content marketing and PR, not a new isolated line item. Investments in third-party reviews (G2, Gartner Peer Insights), trade press coverage, and authoritative educational content produce AI citation benefits as a byproduct. Dedicated monitoring and optimization tooling is worth adding once the content foundation exists. A rough starting point: 10-15% of content budget on AI-specific optimization, more if your deal sizes justify faster investment.

Do different industries respond differently to AI assistant influence in software purchasing?

Yes. Financial services and healthcare buyers, working under regulatory constraints, treat AI-generated research as a starting point that still needs compliance and procurement review. Technology and professional services buyers act more directly on AI recommendations. Company size matters too: startups and SMBs follow AI shortlists more directly, while enterprise companies layer on more human validation. Vertical SaaS categories with thin training data see less AI influence than horizontal platforms.

Can AI assistants get details about our product wrong, and what can we do about it?

Yes, and it's a real risk. AI assistants can carry outdated pricing, incorrect feature lists, or confused positioning against competitors. The fix is publishing authoritative, clear, current content on your own site and in your documentation, and keeping third-party listings accurate. Retrieval systems like Perplexity update fairly quickly from freshly indexed content. Base model errors in ChatGPT or Claude take longer and may require waiting for model updates.

What's the difference between AI assistant influence and traditional SEO for enterprise software brands?

Traditional SEO routes buyers to your website, where you control the experience. AI assistants synthesize information and deliver it directly, often without the buyer visiting your site. You're not competing for a click. You're competing to be the source that gets synthesized. The underlying content signals (authority, specificity, question-answering structure) overlap heavily, but the endpoint differs: AI search can drive zero direct traffic while still shaping the decision.

How do buying committees interact with AI-generated research when multiple stakeholders are involved?

Different committee members often run their own AI queries independently, so your brand might appear consistently to one stakeholder and not at all to another. An IT leader asking technical questions in Claude might get a different vendor set than a finance leader asking ROI questions in ChatGPT. This is an underappreciated complexity. Brands with strong technical documentation fare better with technical buyers; brands with strong business value content fare better with economic buyers.

ai assistant influence on enterprise

Ready to try it?

Build your first app in a few minutes.

Start Building