From click to agent: the next shift in European commerce

E-commerce has evolved relentlessly over the past two decades, from desktop to mobile, from search to social, from browsing to buying. A new transformation is now emerging: agentic commerce.

Defining agentic commerce: AI as an active agent

In the agentic commerce model, AI no longer merely assists users; it acts autonomously on their behalf. AI agents are expanding beyond product recommendations to perform complex purchasing tasks.

Current AI capabilities in commerce include:

  • Product recommendations
  • Search and comparison of products
  • Price negotiation
  • Transaction execution
  • Post-purchase customer service management

Sopra Steria conducted the first European-led benchmark study on consumer attitudes toward agentic commerce.

The survey included:

  • 8,400 consumers interviewed across the study
  • 8 European countries represented in the research

Download the report

From click to agent: the next shift in European commerce
[00:00] Agentic Commerce. [00:05] AI that doesn't just recommend, it buys. Agentic AI is on the verge of deployment across Europe, a market worth [00:13] a potential 310 billion euros. Consumers are interested, but cautious. 55% already know about Agentic Commerce, but awareness varies by country [00:25] and is strongest in Northern Europe. 74% of European consumers face friction when buying online. Agentic Commerce changes that: better [00:36] decisions, quicker turnaround, lower cost. But trust is everything. 94% of consumers want to stay in control. They want AI [00:49] to assist, not decide. Agentic Commerce holds the potential to transform retail. Whether it fulfills that potential remains to be [00:59] seen, but the shift from click to agent has begun. Download the report now.

Key findings: Market size, consumer readiness, and trust

  • The estimated value of e-commerce transactions to be made via by agentic commerce in Europe within just 10 years is €310 billion.
  • 55% of Europeans have already heard of agentic commerce, with 13% saying they understand it well.
    • 74% of online shoppers report at least one pain point in their purchasing journey.
    • Key issues include product discovery (46%) and comparison (44%).
  • 94% of consumers still prefer to approve each transaction before an AI agent executes it on their behalf.
    • 27% of consumers identify banks as the most trusted providers of AI shopping agents.
    • 10% identify technology companies as trusted providers.
    • 6% identify retail companies as trusted providers.

    Strategic questions for industry leaders

    The emergence of agentic commerce raises the following strategic questions for businesses:

    • Customer ownership: Who will own the customer relationship in an agent-driven commercial ecosystem?
    • Brand visibility: Will brands remain visible to consumers, or will AI intermediaries filter brand selection?
    • Regulatory and trust factors: How will trust, transparency, and regulatory frameworks shape adoption across Europe?
    • Sector-specific disruption: Which sectors will experience disruption first, and which industries will prove more resistant to agentic commerce?
    • Value chain evolution: What role will payments infrastructure, financial institutions, and digital platforms play in the emerging agentic commerce value chain?

    The competitive landscape shift

    The battleground for commerce is shifting from websites and mobile applications to algorithms and AI agents as the primary commerce interface.

    Organisations that understand and anticipate this shift will shape the next era of commerce, while others will need to adapt to leadership established by early movers.

    The question is no longer whether agent trading will become established in Europe, but rather who will provide the agent and what the nature of the service's operational model will be.

    Mung-Ki Woo

    Chief Operation Officer, Financial Services, Sopra Steria

    Explore the full report to understand how AI agents will reshape commerce and how to position your organisation now

    Download the report

    From click to agent: Europe's next retail revolution
    1 01:00:00,000 --> 01:00:00,840 Hello, 2 01:00:00,840 --> 01:00:01,960 I'm Mung Ki Woo, 3 01:00:01,960 --> 01:00:03,920 Chief Digital and Innovation Officer 4 01:00:03,920 --> 01:00:05,840 for Financial Services at Sopra Steria. 5 01:00:05,840 --> 01:00:08,320 I am here today with Laurence Niclosse, 6 01:00:08,599 --> 01:00:11,000 Partner for Customer Excellence at Sopra Steria. 7 01:00:11,000 --> 01:00:12,800 We want to present some of the key 8 01:00:12,800 --> 01:00:14,760 findings from a report we have just 9 01:00:14,760 --> 01:00:16,920 completed on agentic commerce in Europe. 10 01:00:22,119 --> 01:00:23,639 For this report, we worked with a 11 01:00:23,639 --> 01:00:26,239 specialist company, Cluster17, to survey 12 01:00:26,239 --> 01:00:28,559 more than 8,400 consumers 13 01:00:28,559 --> 01:00:30,960 across eight European countries. 14 01:00:31,199 --> 01:00:33,119 So what do we mean by agentic commerce? 15 01:00:33,280 --> 01:00:36,000 Well, in short, it's when a consumer uses 16 01:00:36,000 --> 01:00:39,239 an AI agent such as Chatgpt, Claude or 17 01:00:39,239 --> 01:00:42,480 Gemini, not just to get advice, but to 18 01:00:42,480 --> 01:00:43,639 actually make an 19 01:00:43,639 --> 01:00:45,880 e-commerce purchase on their behalf. 20 01:00:45,880 --> 01:00:49,599 Today, consumers already use AI agents to 21 01:00:49,599 --> 01:00:50,199 help them choose 22 01:00:50,199 --> 01:00:52,360 products or find the best deals. 23 01:00:52,360 --> 01:00:56,079 Tomorrow, they will go one step further 24 01:00:56,079 --> 01:00:58,599 and use agents to take care of 25 01:00:58,599 --> 01:01:00,360 the entire purchase for them. 26 01:01:00,360 --> 01:01:04,320 This is not just a question of a better chatbot. 27 01:01:04,320 --> 01:01:05,800 It is an entirely new 28 01:01:05,800 --> 01:01:07,599 intermediary in the value chain. 29 01:01:07,599 --> 01:01:10,119 So, Laurence, let's start with what 30 01:01:10,119 --> 01:01:12,199 consumers are telling us about their 31 01:01:12,199 --> 01:01:14,360 current e-commerce habits. 32 01:01:14,440 --> 01:01:16,800 Yes, thanks Mung Ki. 33 01:01:16,800 --> 01:01:18,599 The first thing we see 34 01:01:18,599 --> 01:01:20,079 is just how widespread 35 01:01:20,079 --> 01:01:21,840 e-commerce has now become. 36 01:01:22,519 --> 01:01:26,360 Only 2% of people in our survey say that 37 01:01:26,360 --> 01:01:28,079 they never buy online. 38 01:01:28,400 --> 01:01:31,199 The second interesting data point is that 39 01:01:31,199 --> 01:01:35,079 55% of consumers say they are already 40 01:01:35,079 --> 01:01:38,039 aware of the agentic commerce concept. 41 01:01:38,639 --> 01:01:42,000 Next, we found consumers reported 42 01:01:42,000 --> 01:01:44,960 friction points when shopping online. 43 01:01:45,679 --> 01:01:50,480 For example, 46% of people said it takes 44 01:01:50,480 --> 01:01:52,239 too long to find the right product. 45 01:01:52,880 --> 01:01:56,519 44% struggled to compare different offers. 46 01:01:56,519 --> 01:02:01,159 And 19% are frustrated at the time spent 47 01:02:01,159 --> 01:02:03,000 just finding out which 48 01:02:03,000 --> 01:02:04,360 retailer to buy from. 49 01:02:04,920 --> 01:02:09,119 So, there is clearly friction in the process. 50 01:02:09,119 --> 01:02:12,239 Friction agentic commerce could help reduce. 51 01:02:12,239 --> 01:02:16,119 Mung Ki, how interested are consumers in 52 01:02:16,119 --> 01:02:18,320 actually using agentic commerce? 53 01:02:18,559 --> 01:02:21,280 Well, overall, interest is quite high. 54 01:02:21,519 --> 01:02:23,960 Although, this varies on the type of 55 01:02:23,960 --> 01:02:25,000 product being bought. 56 01:02:25,320 --> 01:02:28,599 45% of respondents are happy to delegate 57 01:02:28,599 --> 01:02:31,119 electronics purchases to an AI agent. 58 01:02:31,480 --> 01:02:35,199 But that drops to only 16% for groceries 59 01:02:35,199 --> 01:02:36,840 and health-related products. 60 01:02:37,559 --> 01:02:40,400 And only 22% say they are ready to 61 01:02:40,400 --> 01:02:42,519 delegate purchases in financial services. 62 01:02:42,920 --> 01:02:44,039 So, what is happening here? 63 01:02:44,440 --> 01:02:45,719 Well, consumers are more 64 01:02:45,719 --> 01:02:47,480 comfortable using AI agents 65 01:02:48,039 --> 01:02:50,559 for standardized or low-risk purchases. 66 01:02:51,320 --> 01:02:53,400 But when it comes to personal or 67 01:02:53,400 --> 01:02:55,119 sensitive decisions, 68 01:02:55,119 --> 01:02:57,480 they are more hesitant, at least for now. 69 01:02:57,480 --> 01:02:59,239 One result that is 70 01:02:59,239 --> 01:03:02,400 really striking is that 94% of 71 01:03:02,400 --> 01:03:05,880 consumers want to approve every single 72 01:03:05,880 --> 01:03:08,440 transaction before it goes through. 73 01:03:08,440 --> 01:03:11,760 So, Europeans today are not looking for a 74 01:03:11,760 --> 01:03:14,480 robot to shop for them on their behalf. 75 01:03:15,119 --> 01:03:18,320 What they want is an assistant that does 76 01:03:18,320 --> 01:03:20,800 the hard work, meaning searching, 77 01:03:21,639 --> 01:03:24,679 comparing and shortlisting, and then lets 78 01:03:24,679 --> 01:03:27,280 the consumer make the final call. 79 01:03:27,719 --> 01:03:29,960 Another fascinating point concerns 80 01:03:29,960 --> 01:03:31,360 regional nuances. 81 01:03:31,599 --> 01:03:32,599 Overall, 82 01:03:32,599 --> 01:03:34,400 we have found that Europeans 83 01:03:34,400 --> 01:03:36,519 are basically saying the same thing. 84 01:03:36,519 --> 01:03:38,119 But northern European countries 85 01:03:38,119 --> 01:03:40,800 like Norway and the Netherlands 86 01:03:40,800 --> 01:03:42,840 are more enthusiastic, while southern 87 01:03:42,840 --> 01:03:44,079 European countries like 88 01:03:44,079 --> 01:03:47,360 France, Spain, Italy are more cautious 89 01:03:47,920 --> 01:03:50,079 Laurence, what will it take for 90 01:03:50,079 --> 01:03:52,199 consumers to really adopt 91 01:03:52,199 --> 01:03:53,960 agentic commerce at scale? 92 01:03:54,400 --> 01:03:57,719 Well, it comes down to just one word, 93 01:03:57,719 --> 01:03:58,920 trust. 94 01:03:58,920 --> 01:04:00,119 Consumers are very 95 01:04:00,119 --> 01:04:02,000 clear on what they need to 96 01:04:02,000 --> 01:04:03,360 feel comfortable. 97 01:04:03,360 --> 01:04:05,679 First, transparency. 98 01:04:05,679 --> 01:04:07,679 They want to be sure that the 99 01:04:07,679 --> 01:04:10,280 agent will really act in their interest. 100 01:04:10,719 --> 01:04:13,880 76% of respondents say they 101 01:04:13,880 --> 01:04:15,800 want full transparency on 102 01:04:15,800 --> 01:04:17,800 the product selection criteria. 103 01:04:17,800 --> 01:04:20,000 Second, control. 104 01:04:20,000 --> 01:04:22,599 64% want easy 105 01:04:22,599 --> 01:04:24,239 cancellation or returns. 106 01:04:24,639 --> 01:04:25,519 Basically, 107 01:04:25,519 --> 01:04:27,480 consumer want to be protected 108 01:04:27,480 --> 01:04:29,119 if there is a mistake. 109 01:04:29,119 --> 01:04:30,599 And third, 110 01:04:30,599 --> 01:04:31,480 privacy. 111 01:04:31,480 --> 01:04:33,760 55% have concerns 112 01:04:33,760 --> 01:04:35,840 about how the data will be used. 113 01:04:35,840 --> 01:04:39,039 So if companies want adoption, 114 01:04:39,039 --> 01:04:42,039 they will need to address these concerns. 115 01:04:42,039 --> 01:04:44,639 Mung Ki let's move to another topic. 116 01:04:44,639 --> 01:04:45,480 Can we talk about 117 01:04:45,480 --> 01:04:46,599 118 01:04:46,599 --> 01:04:47,320 Sure. 119 01:04:47,679 --> 01:04:50,199 Consumers are already using AI to help 120 01:04:50,199 --> 01:04:51,039 them choose the right 121 01:04:51,039 --> 01:04:52,239 products and services. 122 01:04:52,400 --> 01:04:54,679 Roughly half of consumers already do this, 123 01:04:54,679 --> 01:04:56,480 according to multiple studies. 124 01:04:56,480 --> 01:04:57,280 Based on the 125 01:04:57,280 --> 01:04:59,400 level of interest we have seen in our 126 01:04:59,400 --> 01:05:02,320 surveys, we estimate that within 127 01:05:02,320 --> 01:05:06,880 10 years, around €310 billion in e-commerce transactions 128 01:05:06,880 --> 01:05:10,519 could be impacted by agentic commerce. 129 01:05:10,519 --> 01:05:12,440 And if agentic commerce 130 01:05:12,440 --> 01:05:15,679 follows a trajectory similar to social commerce, 131 01:05:15,679 --> 01:05:19,280 we could see between €62 and €77 billion 132 01:05:19,280 --> 01:05:22,519 of transactions impacted within just three years. 133 01:05:22,519 --> 01:05:25,199 Looking ahead, there are four types of players 134 01:05:25,199 --> 01:05:27,440 which could become AI purchasing agents. 135 01:05:27,480 --> 01:05:28,840 First, the big tech 136 01:05:28,840 --> 01:05:31,480 players, such as OpenAI or Google. 137 01:05:31,480 --> 01:05:34,440 Second, retailers themselves. 138 01:05:34,440 --> 01:05:35,440 Third, 139 01:05:35,440 --> 01:05:36,440 financial services 140 01:05:36,440 --> 01:05:38,480 players, in particular banks 141 01:05:38,480 --> 01:05:41,039 which already integrate offers and promotions 142 01:05:41,039 --> 01:05:42,840 into their mobile apps. 143 01:05:42,840 --> 01:05:45,039 One can easily imagine how 144 01:05:45,039 --> 01:05:48,000 these banks could take it one step further. 145 01:05:48,000 --> 01:05:49,840 And finally, 146 01:05:49,840 --> 01:05:51,880 specialised vertical players, 147 01:05:51,880 --> 01:05:53,599 such as travel specialists. 148 01:05:53,599 --> 01:05:54,199 Time will 149 01:05:54,199 --> 01:05:56,239 tell what market share will 150 01:05:56,239 --> 01:05:58,480 be captured by each type of player. 151 01:05:58,480 --> 01:06:01,519 When we asked consumers who they 152 01:06:01,519 --> 01:06:04,880 would trust to provide these AI agents, 153 01:06:05,480 --> 01:06:10,239 banks came out on top with 27% of respondents. 154 01:06:10,239 --> 01:06:14,199 14% prefer a European technology company, 155 01:06:14,199 --> 01:06:17,280 and only 6% a major retailer. 156 01:06:17,480 --> 01:06:21,280 Interestingly, 43% are still undecided, 157 01:06:21,280 --> 01:06:23,559 meaning that the market is wide open. 158 01:06:23,559 --> 01:06:25,639 One last thing to note is that 159 01:06:25,639 --> 01:06:27,239 agentic commerce is going to 160 01:06:27,239 --> 01:06:29,199 create new technical functions 161 01:06:29,199 --> 01:06:30,360 in the value chain. 162 01:06:30,360 --> 01:06:32,320 First, retailers will 163 01:06:32,320 --> 01:06:33,480 have to verify the 164 01:06:33,480 --> 01:06:35,559 identity of the agents making the 165 01:06:35,559 --> 01:06:36,320 transactions. 166 01:06:36,320 --> 01:06:38,800 And second, both retailers 167 01:06:38,800 --> 01:06:41,239 and banks will also need 168 01:06:41,239 --> 01:06:43,519 to check that transactions 169 01:06:43,519 --> 01:06:46,320 actually reflect the user's original 170 01:06:46,320 --> 01:06:47,320 intent. 171 01:06:47,639 --> 01:06:47,960 There are 172 01:06:47,960 --> 01:06:49,440 already several protocols being 173 01:06:49,440 --> 01:06:50,920 developed to address this, but they are 174 01:06:50,920 --> 01:06:51,519 developed to address this, but they are 175 01:06:51,519 --> 01:06:52,760 likely to evolve. Laurence, 176 01:06:52,760 --> 01:06:53,599 Laurence, 177 01:06:53,599 --> 01:06:54,920 in the view of all of this, 178 01:06:54,920 --> 01:06:57,760 what should companies actually be doing 179 01:06:57,760 --> 01:06:59,000 right now? 180 01:06:59,000 --> 01:06:59,559 There are 181 01:06:59,559 --> 01:07:01,559 some concrete steps to take. 182 01:07:01,559 --> 01:07:04,760 For retailers, the immediate risk is what 183 01:07:04,760 --> 01:07:07,480 we call silent disintermediation. 184 01:07:07,480 --> 01:07:09,760 If consumers rely on agents, 185 01:07:09,760 --> 01:07:13,039 they stop visiting websites, 186 01:07:13,039 --> 01:07:14,679 stop browsing catalogues, 187 01:07:14,679 --> 01:07:16,760 and stop seeing promotion. 188 01:07:16,760 --> 01:07:19,719 The retailer's brand becomes invisible. 189 01:07:19,719 --> 01:07:21,639 The AI agents' algorithm 190 01:07:21,639 --> 01:07:23,760 will decide whether it will 191 01:07:23,760 --> 01:07:25,760 be selected or ignored. 192 01:07:25,760 --> 01:07:27,360 So, the urgent 193 01:07:27,360 --> 01:07:29,519 action is to develop a strong 194 01:07:29,519 --> 01:07:31,360 strategy for GEO, 195 01:07:31,360 --> 01:07:33,599 Generative Engine Optimisation 196 01:07:33,599 --> 01:07:35,360 In other words, make 197 01:07:35,360 --> 01:07:36,679 sure your products are 198 01:07:36,679 --> 01:07:38,400 correctly interpreted and 199 01:07:38,400 --> 01:07:41,119 recommended by the AI agents. 200 01:07:41,119 --> 01:07:44,320 This is digital marketing in the agentic era. 201 01:07:44,320 --> 01:07:45,760 For banks, 202 01:07:45,760 --> 01:07:48,519 the opportunity is different. They can 203 01:07:48,519 --> 01:07:50,480 move higher in the value chain 204 01:07:50,480 --> 01:07:52,519 to become the agent provider, 205 01:07:52,960 --> 01:07:55,320 not just the invisible payment player in 206 01:07:55,320 --> 01:07:57,119 the background. The agent 207 01:07:57,119 --> 01:07:58,679 can help manage spending, 208 01:07:59,559 --> 01:08:02,360 find better deals, all directly from 209 01:08:02,360 --> 01:08:04,119 within the banking app. 210 01:08:04,719 --> 01:08:07,360 Retailers and banks both need to 211 01:08:07,360 --> 01:08:09,840 ensure they don't treat agentic commerce 212 01:08:09,840 --> 01:08:13,639 as a technology silo. It cuts across 213 01:08:13,639 --> 01:08:15,679 strategy, marketing, 214 01:08:16,279 --> 01:08:19,880 payments, IT, and compliance. 215 01:08:19,880 --> 01:08:23,279 And don't wait for standards to be finalised. 216 01:08:23,279 --> 01:08:25,920 Get involved in defining them. 217 01:08:25,920 --> 01:08:27,640 Mung Ki, back to you. 218 01:08:27,640 --> 01:08:29,000 Thanks Laurence. 219 01:08:29,000 --> 01:08:30,760 If I had to give you three key 220 01:08:30,760 --> 01:08:31,840 takeaways to remember 221 01:08:31,840 --> 01:08:34,479 from this report, they would be, one, 222 01:08:35,399 --> 01:08:36,520 agentic commerce is likely 223 01:08:36,520 --> 01:08:38,239 to be big and to be adopted 224 01:08:38,319 --> 01:08:41,600 quickly by consumers. Two, it will 225 01:08:41,600 --> 01:08:43,159 profoundly change the way 226 01:08:43,159 --> 01:08:45,600 e-commerce works. And three, 227 01:08:46,479 --> 01:08:49,279 the time to act is now to make sure you 228 01:08:49,279 --> 01:08:51,760 are part of shaping what comes next. 229 01:08:52,119 --> 01:08:52,840 Thank you so much for 230 01:08:52,880 --> 01:08:55,239 watching. And if you want to know more, 231 01:08:55,680 --> 01:08:58,399 please download our full report. 232 01:08:58,399 --> 01:09:00,079 Thank you Laurence. And goodbye.