August 16, 2025 • 12 mins
Generative AI is disrupting the SaaS playbook, and pricing strategies are in the spotlight. This episode breaks down the real costs, reveals what’s working (and what’s not), and shares actionable advice for leaders navigating the evolving landscape.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Hey everyone, Andy here.
(00:01):
Welcome back to the AI breakdown, the podcast where we separate signal from noise on all things AI at work.
Today I'm tackling a question that is quietly causing panic in boardrooms and product teams.
How on earth do you actually price gen AI in SaaS? Is it as simple as bolting on another add-on? Or have we blown up the old playbook for good? Whether you are a product manager wrestling with pricing models, a SaaS, CEO, balancing your bottom line.
(00:29):
Or you're simply wondering why that tool you love just jumped up 20 quid a month.
This episode is tailor made for you.
I'm going to unpack why Gen AI has completely turned SaaS pricing on its head.
Reveal what vendors are doing to adapt their models and show you exactly where buyers are drawing their red lines in the sand. 9 00:00:49,926.94328344 --> 00:00:50,916.94328344 Let's start at the root. 10 00:00:51,396.94328344 --> 00:00:56,746.94328344 Why does gen AI break traditional SaaS pricing? Old school SAS was fairly straightforward. 11 00:00:56,956.94328344 --> 00:01:00,376.94328344 You paid your license fee usually per user, per month, and that covered everything. 12 00:01:01,324.290222215 --> 00:01:03,844.290222215 Predictable for the customer, brilliant for the vendor. 13 00:01:04,324.290222215 --> 00:01:06,394.290222215 One more user costs them almost nothing. 14 00:01:06,574.290222215 --> 00:01:08,674.290222215 And profit margins soared as you scaled. 15 00:01:09,64.290222215 --> 00:01:11,584.290222215 But with gen ai, that logic unravels. 16 00:01:12,64.290222215 --> 00:01:17,764.290222215 Every time a user fires off a prompt or gets an AI generated summary, there's a real cost to the vendor. 17 00:01:18,139.290222215 --> 00:01:21,229.290222215 Measured not in users, but in model tokens. 18 00:01:21,439.290222215 --> 00:01:25,969.290222215 API calls and compute minutes As B, c, G and Gartner have flagged. 19 00:01:25,999.290222215 --> 00:01:27,379.290222215 This is a tectonic shift. 20 00:01:27,769.290222215 --> 00:01:34,9.290222215 The new metrics, tokens, actions, sometimes even words or images, mean cost scale with actual usage. 21 00:01:34,9.290222215 --> 00:01:37,129.290222215 At the upshot, pricing is suddenly harder to predict. 22 00:01:37,519.290222215 --> 00:01:43,699.29022222 For vendors, a handful of power users can blow up their margins with heavy AI usage and for buyers. 23 00:01:44,44.29022222 --> 00:01:50,14.29022222 The bill might be double what you expected if your team gets a bit too enthusiastic with that new magic AI button. 24 00:01:50,674.29022222 --> 00:02:02,314.29022222 There's real world examples of product managers who watched cloud costs rocket within weeks of rolling out new Gen AI features, leaving them scrambling for usage controls after launch instead of before. 25 00:02:02,809.29022222 --> 00:02:04,189.29022222 Let's dig a bit deeper here. 26 00:02:04,639.29022222 --> 00:02:15,349.29022222 Unlike traditional software where your ongoing costs are basically server bills and support, genai comes with real recurring and frankly, non-trivial infrastructure costs. 27 00:02:15,859.29022222 --> 00:02:17,29.29022222 Here's the quick rundown. 28 00:02:17,419.29022222 --> 00:02:28,999.29022222 Every time someone uses your Genai feature behind the scenes, it calls a large language model like GPT-4, Claude or Gemini, and the costs add up based on what we call tokens. 29 00:02:29,749.29022222 --> 00:02:31,969.29022222 Think of tokens as bite-sized word chunks. 30 00:02:32,374.29022222 --> 00:02:40,234.29022222 For example, with open AI's GPT-4 zero, you're looking at $2 and 50 cents per million input tokens. 31 00:02:40,714.29022222 --> 00:02:42,754.29022222 That's everything your users sent to the ai. 32 00:02:42,814.29022222 --> 00:02:47,764.29022222 Like summarize this document and a hefty at $10 per million output tokens. 33 00:02:48,94.29022222 --> 00:02:55,459.29022222 All those clever responses coming back, think of it as paying a small fee every time your users have a conversation with the ai. 34 00:02:56,284.29022222 --> 00:02:58,804.29022222 Now, let me put this into perspective picture. 35 00:02:58,804.29022222 --> 00:03:03,844.29022222 A SAS customer support tool handling 10 million requests annually using Gen ai. 36 00:03:04,204.29022222 --> 00:03:10,474.29022222 And let's say each request involves approximately 2000 tokens split between input and output. 37 00:03:10,864.29022222 --> 00:03:13,174.29022222 So you're dealing with 20 billion tokens annually. 38 00:03:13,594.29022222 --> 00:03:20,74.29022222 With GPT-4 oh, you're looking at approximately $125,000 in raw API costs. 39 00:03:20,629.29022222 --> 00:03:30,769.29022222 That's before you even consider all the extra infrastructure safety mechanisms, data retrieval, performance monitoring, compliance checks, and redundancy systems. 40 00:03:31,249.29022222 --> 00:03:37,69.29022222 While this cost might seem manageable for a successful SaaS business, remember it scales directly with usage. 41 00:03:37,459.29022222 --> 00:03:45,889.29022222 Double your customer base or add more AI features, and these costs multiply accordingly, factor in the additional operational overhead. 42 00:03:46,234.29022222 --> 00:03:57,184.29022222 Which can easily add 50 to 100% to your base API costs, and you're suddenly looking at 200,000 to $250,000 annually just for AI capabilities. 43 00:03:57,574.29022222 --> 00:04:03,754.29022222 So it can add up significantly in a scaled business making cost optimization and token efficiency. 44 00:04:03,784.29022222 --> 00:04:07,234.29022222 Crucial considerations for any gen AI powered product. 45 00:04:07,774.29022222 --> 00:04:12,899.29022222 Therefore, as a vendor, if you naively price gen AI as a flat add-on and your users love it. 46 00:04:13,549.29022222 --> 00:04:20,449.29022222 You could be in serious financial pain like with GitHub copilot, who reportedly lost money due to these runaway costs. 47 00:04:20,869.29022221 --> 00:04:38,29.29022222 Every pricing conversation should revolve around delivering clear business value, ensuring you demonstrate that Gen AI capabilities genuinely improve outcomes, command a premium, and ultimately become profitable, rather than just a costly feature that bleeds your margins dry. 48 00:04:39,283.94719392 --> 00:04:43,213.94719392 Now, let's explore how SaaS companies are actually pricing gen AI today. 49 00:04:43,573.94719392 --> 00:04:56,173.94719392 Starting with the familiar seat based pricing, this AI add-on model is easy from the buyer's point of view, and now is standard for some business gen AI offerings sitting comfortably alongside base SaaS subscriptions. 50 00:04:56,653.94719392 --> 00:05:04,358.94719392 For example, Microsoft Co-Pilot and Google Duet AI have adopted this approach and a price at about $30 per user per month. 51 00:05:04,888.94719392 --> 00:05:09,313.94719392 Salesforce have also adopted seat based pricing for some of their AI offering. 52 00:05:09,703.94719392 --> 00:05:12,493.94719392 Cost varies depending on the specific product and addition. 53 00:05:13,438.94719392 --> 00:05:18,658.94719392 For Sales Cloud and Service Cloud Einstein, the list price is $50 per user per month. 54 00:05:19,78.94719392 --> 00:05:22,78.94719392 Then there's the usage based approach where you pay for what you consume. 55 00:05:22,528.94719392 --> 00:05:27,898.94719392 Think Open AI's API, Adobe Firefly Credits or Salesforce Flex Credits. 56 00:05:28,138.94719392 --> 00:05:33,358.94719392 It's essentially pay as you go measured in tokens, credits, or individual AI actions. 57 00:05:33,778.94719392 --> 00:05:37,473.94719392 But enterprise IT departments absolutely hate unpredictable bills. 58 00:05:38,308.94719392 --> 00:05:46,888.94719392 Recent surveys paint a stark picture nearly half of all potential buyers site pricing volatility as their number one reason for holding back on adoption. 59 00:05:47,338.94719392 --> 00:05:50,998.94719392 Now, where we are really seeing traction is with hybrid models. 60 00:05:51,358.94719392 --> 00:05:55,78.94719392 Companies like ServiceNow and Zendesk have struck a clever balance. 61 00:05:55,378.94719392 --> 00:05:57,598.94719392 They give you a predictable baseline license. 62 00:05:57,598.94719392 --> 00:06:07,588.94719392 Coupled with flexible usage allowances, analysts estimate that nearly 40% of AI and enterprise SaaS vendors have adopted or trialed some form of hybrid pricing. 63 00:06:08,23.94719392 --> 00:06:15,763.94719392 A strict seat based or flat source pricing no longer reflects the highly variable costs and value delivery of advanced AI workflows. 64 00:06:16,303.94719392 --> 00:06:18,13.94719392 Need more AI firepower. 65 00:06:18,523.94719392 --> 00:06:20,203.94719392 Simply top it with more credits. 66 00:06:20,533.94719392 --> 00:06:22,453.94719392 It's like having the best of both worlds. 67 00:06:22,753.94719392 --> 00:06:28,118.94719392 The financial certainty that keeps your CFO happy paired with the flexibility to scale when you need it. 68 00:06:28,668.94719392 --> 00:06:30,493.94719392 Then there's outcome based pricing. 69 00:06:30,793.94719392 --> 00:06:36,913.94719392 Think about services like Intercom Fin, who only charge you when the AI actually delivers results. 70 00:06:37,423.94719392 --> 00:06:45,943.94719392 Is the chatbot successfully resolving customer tickets? That'll be 99 pence per resolve conversation, but if it failed, it costs you nothing. 71 00:06:46,633.94719392 --> 00:06:49,363.94719392 This approach builds incredible trust with customers. 72 00:06:49,783.94719392 --> 00:06:55,813.94719392 It's the vendor essentially saying, we're so confident in our AI that we'll only charge when it works. 73 00:06:56,443.94719392 --> 00:07:00,823.94719392 But this model only functions in scenarios where success is crystal clear. 74 00:07:01,363.94719392 --> 00:07:05,233.94719392 For many Gen AI features, measuring true value isn't black and white. 75 00:07:05,593.94719392 --> 00:07:14,383.94719392 It gets tricky when you try to measure whether an AI suggestion actually saved someone time or if that auto-generated summary really made a difference. 76 00:07:14,773.94719392 --> 00:07:18,433.94719392 That stems way harder to track in any meaningful, consistent way. 77 00:07:18,793.94719392 --> 00:07:28,453.94719392 And then there's the bundled approach where AI features are included in base plans at no explicit extra charge, often to increase overall value or aid adoption. 78 00:07:28,993.94719392 --> 00:07:32,683.94719392 Some vendors like Gong include AI features right out of the box. 79 00:07:32,983.94719392 --> 00:07:34,813.94719392 Their models actually a bit of a hybrid. 80 00:07:35,203.94719392 --> 00:07:44,833.94719392 Things like conversation intelligence, deal insights, real time transcription forecasting, AI coaching and analytics all come bundled into the core platform. 81 00:07:45,193.94719392 --> 00:07:52,248.94719392 Now, they may charge more for certain modules, advanced forecasting, custom reporting, that sort of thing, but the key here. 82 00:07:52,993.94719392 --> 00:07:55,33.94719392 For your day-to-day work, it's all included. 83 00:07:55,393.94719392 --> 00:07:59,173.94719392 No fing about with the add-ons or worrying how many credits you've burned through. 84 00:07:59,503.94719392 --> 00:08:00,133.94719392 It's simple. 85 00:08:00,133.94719392 --> 00:08:01,63.94719392 It just works. 86 00:08:01,918.94719392 --> 00:08:03,118.94719392 Here's the reality though. 87 00:08:03,508.94719392 --> 00:08:05,308.94719392 There's no perfect model out there. 88 00:08:05,728.94719392 --> 00:08:11,788.94719392 Most SaaS companies are mixing and matching approaches, constantly keeping their finger on the pulse of buyer sentiment. 89 00:08:12,328.94719392 --> 00:08:15,478.94719392 They're ready to pivot at the first sign of customer pushback. 90 00:08:15,838.94719392 --> 00:08:19,108.94719392 It's essentially a massive pricing laboratory right now. 91 00:08:19,438.94719392 --> 00:08:25,18.94719392 Everyone's experimenting with different formulas, tweaking variables, and watching the results unfold. 92 00:08:26,328.19426048 --> 00:08:29,208.19426048 Now let's examine the buyer experience in practical terms. 93 00:08:29,673.19426048 --> 00:08:35,373.19426048 If you think product managers and finance leads are calmly welcoming all this pricing innovation, think again. 94 00:08:35,943.19426048 --> 00:08:49,803.19426048 Data across the board shows the number one barrier to Gen AI adoption is yes price unpredictability with 46% of IT professionals seeing it's their top concern and Salesforce's own research reports. 95 00:08:49,803.19426048 --> 00:08:54,423.19426048 90% of CIOs say AI cost management is throttling their rollouts. 96 00:08:55,8.19426048 --> 00:08:59,538.19426048 What do buyers actually want? Transparency, predictability, and control. 97 00:08:59,898.19426048 --> 00:09:06,948.19426048 They need clear upfront costs, real-time dashboards that track usage and the ability to set hard caps on AI spending. 98 00:09:07,548.19426048 --> 00:09:09,108.19426048 No one wants to be that person. 99 00:09:09,108.19426048 --> 00:09:14,238.19426048 Explaining to the board why last month's AI bill suddenly tripled on the ground. 100 00:09:14,238.19426048 --> 00:09:20,88.19426048 We're seeing procurement teams getting savvy, demanding ironclad contract clauses that limit AI costs. 101 00:09:20,418.19426048 --> 00:09:26,598.19426048 Or insisting on generous free trial quotas before committing to any tool that might spiral out of control. 102 00:09:27,228.19426048 --> 00:09:30,48.19426048 There's also real pushback on expensive add-on fees. 103 00:09:30,408.19426048 --> 00:09:38,328.19426048 When Atlassian launched rovos $20 per user, AI add-on Takeup was so low, they promptly bundled it back into their standard plans. 104 00:09:38,808.19426048 --> 00:09:44,58.19426048 Likewise, buyers trust the outcome based pricing, but only if they can clearly see the business value. 105 00:09:44,418.19426048 --> 00:09:47,418.19426048 If the bot actually resolves a ticket, go ahead and bill us. 106 00:09:48,18.19426048 --> 00:09:50,208.19426048 Random metering just pushes people away. 107 00:09:50,628.19426048 --> 00:09:53,388.19426048 Budgets are tight and no one wants a blank check. 108 00:09:53,688.19426048 --> 00:10:00,18.19426048 So the SaaS players who make a cost predictable, fair and aligned to proven value win the most loyalty. 109 00:10:00,618.19426048 --> 00:10:07,788.19426048 So what's the winning playbook for pricing gen AI in SaaS today? First, anchor your pricing to genuine customer volume. 110 00:10:08,148.19426048 --> 00:10:20,973.19426048 Can you demonstrate that your AI slashes ticket volumes by 30% or doubles team productivity? Then build your marketing and pricing narrative around these tangible outcomes, even if your actual billing mechanism works differently. 111 00:10:20,973.19426048 --> 00:10:29,493.19426048 Behind the scenes, the most successful pricing models we see and blend predictable base fees with intelligent usage or outcome linked components. 112 00:10:29,943.19426048 --> 00:10:40,803.19426048 If you can't link pricing to outcomes like when it's two subjective, then consider a hybrid usage based model where you provide a predictable baseline allowance with the option to purchase additional credits. 113 00:10:41,208.19426048 --> 00:10:46,428.19426048 This gives customers both the security of fixed costs and the flexibility to scale when needed. 114 00:10:46,878.19426048 --> 00:10:56,118.19426048 Include clear usage metrics, transparent price and tiers, and automatic alerts when approaching limits to help customers maintain control over their spending. 115 00:10:56,748.19426048 --> 00:11:07,398.19426048 The market leaders think Notion AI or Intercom all offer generous trial or free tiers that remove the risk of adoption and steadily build trust with cautious buyers. 116 00:11:08,58.19426048 --> 00:11:09,138.19426048 Here's a crucial warning. 117 00:11:09,498.19426048 --> 00:11:13,128.19426048 Don't rush to monetize features you haven't proven are valuable yet. 118 00:11:13,728.19426048 --> 00:11:15,708.19426048 Get your AI into your user's hands first. 119 00:11:16,188.19426048 --> 00:11:21,618.19426048 Gather concrete data, refine the experience, and only then layer on monetization. 120 00:11:21,678.19426048 --> 00:11:31,458.19426048 My final piece of advice, treat your pricing strategy exactly like your product, continuously test, iterate, listen to feedback, and be ready to pivot when needed. 121 00:11:32,230.4560698 --> 00:11:33,430.4560698 All right, let's wrap this up. 122 00:11:34,285.4560698 --> 00:11:36,325.4560698 I hasn't just changed what SaaS does. 123 00:11:36,595.4560698 --> 00:11:44,725.4560698 It's blown up how it's sold, shaken up margins, adoption, and trust across the industry, buyers are smarter and more skeptical than ever. 124 00:11:45,55.4560698 --> 00:12:02,5.4560698 They want transparent pricing, genuine value delivery, and absolutely zero surprises on the bill SAS companies that communicate that value keep prices simple but flexible and aren't afraid to adjust on the go, they'll be the ones customers stick with and recommend. 125 00:12:02,785.4560698 --> 00:12:10,135.4560698 If this episode struck a nerve or you've seen a Gen AI pricing model crash and burn, or work like a charm, I want to hear from you. 126 00:12:10,855.4560698 --> 00:12:13,765.4560698 Drop me a note at hello@theaibreakdown.com, 127 00:12:14,185.4560698 --> 00:12:20,275.4560698 and if you found this useful, subscribe and send the show to someone who needs a crash course on gen AI pricing. 128 00:12:20,875.4560698 --> 00:12:22,645.4560698 Thanks for listening to the AI breakdown.
Hey everyone, Andy here.
(00:01):
Welcome back to the AI breakdown, the podcast where we separate signal from noise on all things AI at work.
Today I'm tackling a question that is quietly causing panic in boardrooms and product teams.
How on earth do you actually price gen AI in SaaS? Is it as simple as bolting on another add-on? Or have we blown up the old playbook for good? Whether you are a product manager wrestling with pricing models, a SaaS, CEO, balancing your bottom line.
(00:29):
Or you're simply wondering why that tool you love just jumped up 20 quid a month.
This episode is tailor made for you.
I'm going to unpack why Gen AI has completely turned SaaS pricing on its head.
Reveal what vendors are doing to adapt their models and show you exactly where buyers are drawing their red lines in the sand. 9 00:00:49,926.94328344 --> 00:00:50,916.94328344 Let's start at the root. 10 00:00:51,396.94328344 --> 00:00:56,746.94328344 Why does gen AI break traditional SaaS pricing? Old school SAS was fairly straightforward. 11 00:00:56,956.94328344 --> 00:01:00,376.94328344 You paid your license fee usually per user, per month, and that covered everything. 12 00:01:01,324.290222215 --> 00:01:03,844.290222215 Predictable for the customer, brilliant for the vendor. 13 00:01:04,324.290222215 --> 00:01:06,394.290222215 One more user costs them almost nothing. 14 00:01:06,574.290222215 --> 00:01:08,674.290222215 And profit margins soared as you scaled. 15 00:01:09,64.290222215 --> 00:01:11,584.290222215 But with gen ai, that logic unravels. 16 00:01:12,64.290222215 --> 00:01:17,764.290222215 Every time a user fires off a prompt or gets an AI generated summary, there's a real cost to the vendor. 17 00:01:18,139.290222215 --> 00:01:21,229.290222215 Measured not in users, but in model tokens. 18 00:01:21,439.290222215 --> 00:01:25,969.290222215 API calls and compute minutes As B, c, G and Gartner have flagged. 19 00:01:25,999.290222215 --> 00:01:27,379.290222215 This is a tectonic shift. 20 00:01:27,769.290222215 --> 00:01:34,9.290222215 The new metrics, tokens, actions, sometimes even words or images, mean cost scale with actual usage. 21 00:01:34,9.290222215 --> 00:01:37,129.290222215 At the upshot, pricing is suddenly harder to predict. 22 00:01:37,519.290222215 --> 00:01:43,699.29022222 For vendors, a handful of power users can blow up their margins with heavy AI usage and for buyers. 23 00:01:44,44.29022222 --> 00:01:50,14.29022222 The bill might be double what you expected if your team gets a bit too enthusiastic with that new magic AI button. 24 00:01:50,674.29022222 --> 00:02:02,314.29022222 There's real world examples of product managers who watched cloud costs rocket within weeks of rolling out new Gen AI features, leaving them scrambling for usage controls after launch instead of before. 25 00:02:02,809.29022222 --> 00:02:04,189.29022222 Let's dig a bit deeper here. 26 00:02:04,639.29022222 --> 00:02:15,349.29022222 Unlike traditional software where your ongoing costs are basically server bills and support, genai comes with real recurring and frankly, non-trivial infrastructure costs. 27 00:02:15,859.29022222 --> 00:02:17,29.29022222 Here's the quick rundown. 28 00:02:17,419.29022222 --> 00:02:28,999.29022222 Every time someone uses your Genai feature behind the scenes, it calls a large language model like GPT-4, Claude or Gemini, and the costs add up based on what we call tokens. 29 00:02:29,749.29022222 --> 00:02:31,969.29022222 Think of tokens as bite-sized word chunks. 30 00:02:32,374.29022222 --> 00:02:40,234.29022222 For example, with open AI's GPT-4 zero, you're looking at $2 and 50 cents per million input tokens. 31 00:02:40,714.29022222 --> 00:02:42,754.29022222 That's everything your users sent to the ai. 32 00:02:42,814.29022222 --> 00:02:47,764.29022222 Like summarize this document and a hefty at $10 per million output tokens. 33 00:02:48,94.29022222 --> 00:02:55,459.29022222 All those clever responses coming back, think of it as paying a small fee every time your users have a conversation with the ai. 34 00:02:56,284.29022222 --> 00:02:58,804.29022222 Now, let me put this into perspective picture. 35 00:02:58,804.29022222 --> 00:03:03,844.29022222 A SAS customer support tool handling 10 million requests annually using Gen ai. 36 00:03:04,204.29022222 --> 00:03:10,474.29022222 And let's say each request involves approximately 2000 tokens split between input and output. 37 00:03:10,864.29022222 --> 00:03:13,174.29022222 So you're dealing with 20 billion tokens annually. 38 00:03:13,594.29022222 --> 00:03:20,74.29022222 With GPT-4 oh, you're looking at approximately $125,000 in raw API costs. 39 00:03:20,629.29022222 --> 00:03:30,769.29022222 That's before you even consider all the extra infrastructure safety mechanisms, data retrieval, performance monitoring, compliance checks, and redundancy systems. 40 00:03:31,249.29022222 --> 00:03:37,69.29022222 While this cost might seem manageable for a successful SaaS business, remember it scales directly with usage. 41 00:03:37,459.29022222 --> 00:03:45,889.29022222 Double your customer base or add more AI features, and these costs multiply accordingly, factor in the additional operational overhead. 42 00:03:46,234.29022222 --> 00:03:57,184.29022222 Which can easily add 50 to 100% to your base API costs, and you're suddenly looking at 200,000 to $250,000 annually just for AI capabilities. 43 00:03:57,574.29022222 --> 00:04:03,754.29022222 So it can add up significantly in a scaled business making cost optimization and token efficiency. 44 00:04:03,784.29022222 --> 00:04:07,234.29022222 Crucial considerations for any gen AI powered product. 45 00:04:07,774.29022222 --> 00:04:12,899.29022222 Therefore, as a vendor, if you naively price gen AI as a flat add-on and your users love it. 46 00:04:13,549.29022222 --> 00:04:20,449.29022222 You could be in serious financial pain like with GitHub copilot, who reportedly lost money due to these runaway costs. 47 00:04:20,869.29022221 --> 00:04:38,29.29022222 Every pricing conversation should revolve around delivering clear business value, ensuring you demonstrate that Gen AI capabilities genuinely improve outcomes, command a premium, and ultimately become profitable, rather than just a costly feature that bleeds your margins dry. 48 00:04:39,283.94719392 --> 00:04:43,213.94719392 Now, let's explore how SaaS companies are actually pricing gen AI today. 49 00:04:43,573.94719392 --> 00:04:56,173.94719392 Starting with the familiar seat based pricing, this AI add-on model is easy from the buyer's point of view, and now is standard for some business gen AI offerings sitting comfortably alongside base SaaS subscriptions. 50 00:04:56,653.94719392 --> 00:05:04,358.94719392 For example, Microsoft Co-Pilot and Google Duet AI have adopted this approach and a price at about $30 per user per month. 51 00:05:04,888.94719392 --> 00:05:09,313.94719392 Salesforce have also adopted seat based pricing for some of their AI offering. 52 00:05:09,703.94719392 --> 00:05:12,493.94719392 Cost varies depending on the specific product and addition. 53 00:05:13,438.94719392 --> 00:05:18,658.94719392 For Sales Cloud and Service Cloud Einstein, the list price is $50 per user per month. 54 00:05:19,78.94719392 --> 00:05:22,78.94719392 Then there's the usage based approach where you pay for what you consume. 55 00:05:22,528.94719392 --> 00:05:27,898.94719392 Think Open AI's API, Adobe Firefly Credits or Salesforce Flex Credits. 56 00:05:28,138.94719392 --> 00:05:33,358.94719392 It's essentially pay as you go measured in tokens, credits, or individual AI actions. 57 00:05:33,778.94719392 --> 00:05:37,473.94719392 But enterprise IT departments absolutely hate unpredictable bills. 58 00:05:38,308.94719392 --> 00:05:46,888.94719392 Recent surveys paint a stark picture nearly half of all potential buyers site pricing volatility as their number one reason for holding back on adoption. 59 00:05:47,338.94719392 --> 00:05:50,998.94719392 Now, where we are really seeing traction is with hybrid models. 60 00:05:51,358.94719392 --> 00:05:55,78.94719392 Companies like ServiceNow and Zendesk have struck a clever balance. 61 00:05:55,378.94719392 --> 00:05:57,598.94719392 They give you a predictable baseline license. 62 00:05:57,598.94719392 --> 00:06:07,588.94719392 Coupled with flexible usage allowances, analysts estimate that nearly 40% of AI and enterprise SaaS vendors have adopted or trialed some form of hybrid pricing. 63 00:06:08,23.94719392 --> 00:06:15,763.94719392 A strict seat based or flat source pricing no longer reflects the highly variable costs and value delivery of advanced AI workflows. 64 00:06:16,303.94719392 --> 00:06:18,13.94719392 Need more AI firepower. 65 00:06:18,523.94719392 --> 00:06:20,203.94719392 Simply top it with more credits. 66 00:06:20,533.94719392 --> 00:06:22,453.94719392 It's like having the best of both worlds. 67 00:06:22,753.94719392 --> 00:06:28,118.94719392 The financial certainty that keeps your CFO happy paired with the flexibility to scale when you need it. 68 00:06:28,668.94719392 --> 00:06:30,493.94719392 Then there's outcome based pricing. 69 00:06:30,793.94719392 --> 00:06:36,913.94719392 Think about services like Intercom Fin, who only charge you when the AI actually delivers results. 70 00:06:37,423.94719392 --> 00:06:45,943.94719392 Is the chatbot successfully resolving customer tickets? That'll be 99 pence per resolve conversation, but if it failed, it costs you nothing. 71 00:06:46,633.94719392 --> 00:06:49,363.94719392 This approach builds incredible trust with customers. 72 00:06:49,783.94719392 --> 00:06:55,813.94719392 It's the vendor essentially saying, we're so confident in our AI that we'll only charge when it works. 73 00:06:56,443.94719392 --> 00:07:00,823.94719392 But this model only functions in scenarios where success is crystal clear. 74 00:07:01,363.94719392 --> 00:07:05,233.94719392 For many Gen AI features, measuring true value isn't black and white. 75 00:07:05,593.94719392 --> 00:07:14,383.94719392 It gets tricky when you try to measure whether an AI suggestion actually saved someone time or if that auto-generated summary really made a difference. 76 00:07:14,773.94719392 --> 00:07:18,433.94719392 That stems way harder to track in any meaningful, consistent way. 77 00:07:18,793.94719392 --> 00:07:28,453.94719392 And then there's the bundled approach where AI features are included in base plans at no explicit extra charge, often to increase overall value or aid adoption. 78 00:07:28,993.94719392 --> 00:07:32,683.94719392 Some vendors like Gong include AI features right out of the box. 79 00:07:32,983.94719392 --> 00:07:34,813.94719392 Their models actually a bit of a hybrid. 80 00:07:35,203.94719392 --> 00:07:44,833.94719392 Things like conversation intelligence, deal insights, real time transcription forecasting, AI coaching and analytics all come bundled into the core platform. 81 00:07:45,193.94719392 --> 00:07:52,248.94719392 Now, they may charge more for certain modules, advanced forecasting, custom reporting, that sort of thing, but the key here. 82 00:07:52,993.94719392 --> 00:07:55,33.94719392 For your day-to-day work, it's all included. 83 00:07:55,393.94719392 --> 00:07:59,173.94719392 No fing about with the add-ons or worrying how many credits you've burned through. 84 00:07:59,503.94719392 --> 00:08:00,133.94719392 It's simple. 85 00:08:00,133.94719392 --> 00:08:01,63.94719392 It just works. 86 00:08:01,918.94719392 --> 00:08:03,118.94719392 Here's the reality though. 87 00:08:03,508.94719392 --> 00:08:05,308.94719392 There's no perfect model out there. 88 00:08:05,728.94719392 --> 00:08:11,788.94719392 Most SaaS companies are mixing and matching approaches, constantly keeping their finger on the pulse of buyer sentiment. 89 00:08:12,328.94719392 --> 00:08:15,478.94719392 They're ready to pivot at the first sign of customer pushback. 90 00:08:15,838.94719392 --> 00:08:19,108.94719392 It's essentially a massive pricing laboratory right now. 91 00:08:19,438.94719392 --> 00:08:25,18.94719392 Everyone's experimenting with different formulas, tweaking variables, and watching the results unfold. 92 00:08:26,328.19426048 --> 00:08:29,208.19426048 Now let's examine the buyer experience in practical terms. 93 00:08:29,673.19426048 --> 00:08:35,373.19426048 If you think product managers and finance leads are calmly welcoming all this pricing innovation, think again. 94 00:08:35,943.19426048 --> 00:08:49,803.19426048 Data across the board shows the number one barrier to Gen AI adoption is yes price unpredictability with 46% of IT professionals seeing it's their top concern and Salesforce's own research reports. 95 00:08:49,803.19426048 --> 00:08:54,423.19426048 90% of CIOs say AI cost management is throttling their rollouts. 96 00:08:55,8.19426048 --> 00:08:59,538.19426048 What do buyers actually want? Transparency, predictability, and control. 97 00:08:59,898.19426048 --> 00:09:06,948.19426048 They need clear upfront costs, real-time dashboards that track usage and the ability to set hard caps on AI spending. 98 00:09:07,548.19426048 --> 00:09:09,108.19426048 No one wants to be that person. 99 00:09:09,108.19426048 --> 00:09:14,238.19426048 Explaining to the board why last month's AI bill suddenly tripled on the ground. 100 00:09:14,238.19426048 --> 00:09:20,88.19426048 We're seeing procurement teams getting savvy, demanding ironclad contract clauses that limit AI costs. 101 00:09:20,418.19426048 --> 00:09:26,598.19426048 Or insisting on generous free trial quotas before committing to any tool that might spiral out of control. 102 00:09:27,228.19426048 --> 00:09:30,48.19426048 There's also real pushback on expensive add-on fees. 103 00:09:30,408.19426048 --> 00:09:38,328.19426048 When Atlassian launched rovos $20 per user, AI add-on Takeup was so low, they promptly bundled it back into their standard plans. 104 00:09:38,808.19426048 --> 00:09:44,58.19426048 Likewise, buyers trust the outcome based pricing, but only if they can clearly see the business value. 105 00:09:44,418.19426048 --> 00:09:47,418.19426048 If the bot actually resolves a ticket, go ahead and bill us. 106 00:09:48,18.19426048 --> 00:09:50,208.19426048 Random metering just pushes people away. 107 00:09:50,628.19426048 --> 00:09:53,388.19426048 Budgets are tight and no one wants a blank check. 108 00:09:53,688.19426048 --> 00:10:00,18.19426048 So the SaaS players who make a cost predictable, fair and aligned to proven value win the most loyalty. 109 00:10:00,618.19426048 --> 00:10:07,788.19426048 So what's the winning playbook for pricing gen AI in SaaS today? First, anchor your pricing to genuine customer volume. 110 00:10:08,148.19426048 --> 00:10:20,973.19426048 Can you demonstrate that your AI slashes ticket volumes by 30% or doubles team productivity? Then build your marketing and pricing narrative around these tangible outcomes, even if your actual billing mechanism works differently. 111 00:10:20,973.19426048 --> 00:10:29,493.19426048 Behind the scenes, the most successful pricing models we see and blend predictable base fees with intelligent usage or outcome linked components. 112 00:10:29,943.19426048 --> 00:10:40,803.19426048 If you can't link pricing to outcomes like when it's two subjective, then consider a hybrid usage based model where you provide a predictable baseline allowance with the option to purchase additional credits. 113 00:10:41,208.19426048 --> 00:10:46,428.19426048 This gives customers both the security of fixed costs and the flexibility to scale when needed. 114 00:10:46,878.19426048 --> 00:10:56,118.19426048 Include clear usage metrics, transparent price and tiers, and automatic alerts when approaching limits to help customers maintain control over their spending. 115 00:10:56,748.19426048 --> 00:11:07,398.19426048 The market leaders think Notion AI or Intercom all offer generous trial or free tiers that remove the risk of adoption and steadily build trust with cautious buyers. 116 00:11:08,58.19426048 --> 00:11:09,138.19426048 Here's a crucial warning. 117 00:11:09,498.19426048 --> 00:11:13,128.19426048 Don't rush to monetize features you haven't proven are valuable yet. 118 00:11:13,728.19426048 --> 00:11:15,708.19426048 Get your AI into your user's hands first. 119 00:11:16,188.19426048 --> 00:11:21,618.19426048 Gather concrete data, refine the experience, and only then layer on monetization. 120 00:11:21,678.19426048 --> 00:11:31,458.19426048 My final piece of advice, treat your pricing strategy exactly like your product, continuously test, iterate, listen to feedback, and be ready to pivot when needed. 121 00:11:32,230.4560698 --> 00:11:33,430.4560698 All right, let's wrap this up. 122 00:11:34,285.4560698 --> 00:11:36,325.4560698 I hasn't just changed what SaaS does. 123 00:11:36,595.4560698 --> 00:11:44,725.4560698 It's blown up how it's sold, shaken up margins, adoption, and trust across the industry, buyers are smarter and more skeptical than ever. 124 00:11:45,55.4560698 --> 00:12:02,5.4560698 They want transparent pricing, genuine value delivery, and absolutely zero surprises on the bill SAS companies that communicate that value keep prices simple but flexible and aren't afraid to adjust on the go, they'll be the ones customers stick with and recommend. 125 00:12:02,785.4560698 --> 00:12:10,135.4560698 If this episode struck a nerve or you've seen a Gen AI pricing model crash and burn, or work like a charm, I want to hear from you. 126 00:12:10,855.4560698 --> 00:12:13,765.4560698 Drop me a note at hello@theaibreakdown.com, 127 00:12:14,185.4560698 --> 00:12:20,275.4560698 and if you found this useful, subscribe and send the show to someone who needs a crash course on gen AI pricing. 128 00:12:20,875.4560698 --> 00:12:22,645.4560698 Thanks for listening to the AI breakdown.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
NFL Daily with Gregg Rosenthal
Gregg Rosenthal and a rotating crew of elite NFL Media co-hosts, including Patrick Claybon, Colleen Wolfe, Steve Wyche, Nick Shook and Jourdan Rodrigue of The Athletic get you caught up daily on all the NFL news and analysis you need to be smarter and funnier than your friends.