CHICAGO \u2014 The market for enterprise AI agents is not being won on model quality anymore, according to Yoav Naveh, founder and CEO of Reindeer, and that shift is reordering where startups plant their competitive bets.<\/p>\n
Speaking with FreightWaves at the Supply Chain AI Symposium in Chicago, Naveh laid out a picture of an industry that has burned through two waves of adoption without unlocking the returns executives expected, and is now chasing a third.<\/p>\n
\u201cThe first wave: assistants. Everybody got a copilot, Claude, ChatGPT,\u201d Naveh said. \u201cWe\u2019ve seen a lot of success with assistants around software engineering. Coding agents work really well. Maybe legal, not as well. But generally we haven\u2019t seen assistants make companies 10x or 100x more efficient.\u201d<\/p>\n
The second wave targeted the call center, i.e. customer support. \u201cI think that has been a very big low-hanging fruit, really,\u201d Naveh said. \u201cOne big workflow, a lot of similar case types in one. I definitely see some success there.\u201d<\/p>\n
Now, Naveh said, enterprises are chasing the wave that actually touches the bulk of their labor cost: core operations. \u201cLet\u2019s take all these core workflows that represent 70, 80% of the work that everybody does,\u201d he said.<\/p>\n
That is a far messier market than customer support ever was. Naveh estimates these workflows number in the dozens, if not hundreds, inside a single enterprise, spanning finance, procurement, treasury, accounting and supply chain, often with no single department owner.<\/p>\n
The Nuance Problem Investors Are Missing<\/p>\n
Naveh\u2019s clearest data point on why this market resists one-size-fits-all software comes from a client he described as one of the largest consumer packaged goods companies in the world.<\/p>\n
\u201cThey say, \u2018Hey, we\u2019re doing accounts payable in my company,\u2019 but your accounts payable is different from their accounts payable,\u201d Naveh said. \u201cWe work with one of the largest CPG companies in the world. Five different departments, five different accounts payable processes within the same company.\u201d<\/p>\n
That fragmentation is the reason Naveh is skeptical of the forward-deployed engineer model gaining traction across the AI startup world, where vendors embed technical staff inside a client to hand-build an agent for one process at a time.<\/p>\n
\u201cI\u2019m going to throw three engineers in and sit with your team and build the agent for that particular process. Are they ever going to come back and be available to be sent to another department, or are they stuck maintaining whatever they built as processes change and evolve?\u201d Naveh said. \u201cI think the math doesn\u2019t work because it doesn\u2019t scale.\u201d<\/p>\n
Why the Model Won\u2019t Be the Moat for Enterprise AI Agents<\/p>\n
Reindeer\u2019s bet runs against a common assumption in enterprise software circles: that owning or fine-tuning a proprietary large language model is a durable edge. Naveh argues the opposite is happening.<\/p>\n
\u201cStrong belief that LLMs are going to be a commodity,\u201d Naveh said. \u201cWe see that open-source models are maybe six months behind the frontier models. But what we\u2019re thinking is, I don\u2019t need the frontier model for everything anymore.\u201d<\/p>\n
\u201cWe work now with a large European insurance company. They decided to build their own LLM. We\u2019re happy to plug in their own LLM,\u201d he said. \u201cWe\u2019re happy to restrict the list of allowed LLMs based on the company\u2019s data privacy policy, and figure out how to run their case with what\u2019s whitelisted.\u201d<\/p>\n
That view puts pressure on any competitor whose pitch centers on model superiority alone. If the underlying model becomes interchangeable, differentiation has to come from somewhere else.<\/p>\n
Building Is Easy. Maintaining Isn\u2019t.<\/p>\n
That somewhere else, in Reindeer\u2019s model, is the maintenance layer: detecting when a workflow drifts, reconciling the change and adjusting the agent without a new engineering ticket.<\/p>\n
\u201cEven after you deploy an agent, the workflow keeps changing,\u201d Naveh said. \u201cIf you don\u2019t have a method to have agents detect these changes, reconcile, self-heal, self-learn, we\u2019re not really going to be able to realize what agents can do for you.\u201d Reindeer built what Naveh calls a two-loop system to handle that drift.<\/p>\n
Naveh said the economics of building have already shifted in the industry\u2019s favor, which is exactly what makes maintenance the bottleneck now.<\/p>\n
\u201cI think everybody wants to build. Building is easy,\u201d he said. He described building an internal dashboard for a colleague in a single evening, then fielding a steady stream of bug fixes and feature requests afterward. \u201cI don\u2019t want to do that,\u201d he said. \u201cThat\u2019s not fun. Maintaining is not fun or easy.\u201d<\/p>\n
\u201cFive years ago you couldn\u2019t build anything, now you can build amazing things,\u201d Naveh said, describing the gap between a novice and an expert engineer as dramatically reduced by tools like live coding assistants. \u201cThe expertise that is needed now is exactly what you said, around all these different things after it\u2019s built, and make sure that what was amazing on the first evening you released it is still amazing six months later.\u201d<\/p>\n
For enterprises evaluating AI vendors, Naveh\u2019s framing suggests a different diligence question than the one most are asking. Instead of which company has the best model at POC stage, the more durable question may be which one can keep an agent accurate long after launch, as the workflow underneath it keeps moving.<\/p>\n","protected":false},"excerpt":{"rendered":"CHICAGO \u2014 The market for enterprise AI agents is not being won on model quality anymore, according to…\n","protected":false},"author":2,"featured_media":128697,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[405,7537,523,12600,7901,64420,64421,6209,64422,134,64423],"class_list":["post-128696","post","type-post","status-publish","format-standard","has-post-thumbnail","category-agentic-ai","tag-ai-agents","tag-artificial-intelligence-agents","tag-enterprise-ai","tag-enterprise-ai-agents","tag-logistics","tag-reindeer","tag-reindeer-ai","tag-supply-chain-ai","tag-supply-chain-ai-symposium","tag-technology","tag-yoav-naveh"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/128696","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=128696"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/128696\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/128697"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=128696"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=128696"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=128696"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}