Articles by Emrecan
Activity
15K followers
-
Emrecan Dogan reacted on thisEmrecan Dogan reacted on thisI spoke to Sohrab Rahimi, Director of AI/ML at Google, at Glean:GO 2026 in San Francisco on The Ravit Show!!!! We went straight at the question enterprise AI is quietly stuck on. Two agents can speak perfect A2A, return a success code, and still be wrong. Because nothing in the protocol says they agree on what a customer is. Here is what stayed with me. 1. The protocol landscape is 3 categories, not 1. Sohrab breaks it into connectivity protocols like MCP and A2A that connect agents to tools and to each other, content protocols like Agents.md and Skills that feed procedural knowledge into agents, and system integration protocols that wire agents into payments, commerce and UIs. They overlap and compete today. He expects convergence, and points to protocols already being donated to neutral foundations as evidence it is under way. 2. The transport layer is solved. The semantic layer is not. This was the sharpest point of the conversation. How agents connect is largely figured out. What they mean is not. Two agents can define "customer" differently, the API call succeeds, the auth passes, and the outcome is still wrong. Nothing in the stack catches it. His fix is unglamorous and correct. Align your own ontology first. No protocol will do that for you. 3. Stop building an army of agents for architectural aesthetics. Referencing Google research, he gave the clearest test I have heard for when to split into multiple agents. Only when the work is parallel rather than sequential, genuinely independent, cognitively different with separate data contexts, and separated by distinct trust boundaries with different permissions. If your workload fails those 4, you wanted 1 agent. 4. Task level authorization is the missing standard. Human SSO is solved. Agents acting on behalf of a user is solved. But whether an agent is authorized to issue a $20,000 refund is still custom code in every enterprise. That is a governance gap sitting directly under every autonomous workflow being demoed right now. 5. He is honest about where his own side is thin. MCP adoption of OAuth in server metadata has been slow. A2A has no designed way for agents to share memory. And if models eventually understand thousands of APIs natively, he thinks the intermediary layer may stop being necessary at all. You do not often hear a protocol's home team argue it might become obsolete. He expects protocols to stabilize into defaults the way HTTP and REST did, and vendors like ServiceNow and Workday to ship pre-built representative agents. You compose workflows from reliable market agents instead of building each one from scratch. We also covered the Glean and Google partnership, including the connectors built into Gemini Enterprise. Thank you for the time at Glean:GO, Sohrab. This one gave me a lot to think about. Full conversation linked in the comments. #dataandai #agenticai #multiagent #mcp #a2a #enterpriseai #aiagents #theravitshow
-
Emrecan Dogan reacted on thisEmrecan Dogan reacted on thisI have a new favourite word from Glean:GO 2026. "Botsitting" I was sitting in the front row when it landed, and the room reacted the way you react when someone names a thing you have been doing for months without a word for it. Finding the file. Pasting the background. Connecting the tool. Correcting the answer. Explaining the same context again the next morning. We have all been calling that "using AI." It is closer to babysitting. And here is the part that matters. Botsitting is not a model problem. It is a context problem. The model has no idea who your customers are, what your team decided last week, or which of those 4 documents is the real one. That single idea sits underneath everything Glean announced today. Glean Tau, a desktop workspace that combines enterprise context with your local files, applications, code and the latest models. It can plan, execute, review, recover and act. Recover is the word I would circle. Executing a plan is table stakes. Recovering from a failed step without a human stepping in is the line between a tool you supervise and a teammate you trust. The benchmark everyone will screenshot. Against Claude Cowork, Glean's response was preferred 78 percent of the time at $0.58 per query versus $2.98. An 81 percent token cost saving. More context does not have to mean more cost when the context is already indexed instead of rebuilt on every request. It is a vendor run benchmark, so treat it as a reason to test on your own queries rather than a verdict. Context that stops getting rebuilt. People, content, processes, customers, infrastructure, all permission aware, so the AI knows what matters, what is connected, what it can access and what happens next. Control over the economics. Usage visibility, auto routing across efficient, balanced and frontier models, and usage controls across users, teams and models. Model choice just became a budget decision instead of an engineering preference. Glean AI Gateway, one common layer for models, context, memory, MCP and security, so all of this does not become sprawl nobody can govern. Team chat, where multiple people work with Glean in 1 shared thread and every response still respects individual permissions. Plus Transform, proactive task management, interactive dashboards and context aware threat detection, which flags risky agent behaviour even when every individual action was permitted. The thread through all of it: models are becoming interchangeable, enterprise understanding is not. We are moving from AI that answers questions to AI that understands the business and gets work done. Swipe through the carousel for the full breakdown. Then tell me which one you would actually put to work on Monday. #glean #gleango #dataandai #enterpriseai #aiagents #aigovernance #futureofwork #theravitshow
-
Emrecan Dogan shared thisModels are getting better and more interchangeable. Understanding the enterprise is not. As AI moves from answering questions to doing work, context tells it what matters, what should happen next, and how to act within the realities of the business. That is the foundation for a new suite of proactive AI capabilities at Glean. New task management, email triage, and meeting coach capabilities help identify what matters and prepare the next step before you ask. A follow-up becomes a first draft. An inbox gets prioritized based on your role, projects, and relationships. In a meeting, you get in-the-moment guidance with clearer next steps that Glean proactively takes on after. Independent agents extend proactive AI across teams. With their own identity, permissions, memory, and context, they can work across tools like Slack, Jira, and GitHub to keep workflows moving and bring in the right teammates. Team chat adds to the multiplayer layer, giving people and AI one shared space to build on the same context together. This is AI that understands the work, takes the next step, and helps everyone move together. It was great to share these updates onstage this morning. We’re just getting started at Glean:GO. https://lnkd.in/epztzAr9
-
Emrecan Dogan shared thisSkills in Glean now have approval flows, versioning, rollback, org-wide sharing, and GitHub sync. A great prompt scales a person. Scaling a company takes everything around the prompt: who approved it, which version is running, and what happens when it turns out to be wrong. I think of a skill as compiled institutional judgment: the exceptions and tacit calls our best people make, captured in a form that runs. Picture the skill that fills out security questionnaires, and what a confidently wrong version might cost. That skill needs what code got decades ago: a reviewed, versioned, revertible lifecycle with a clear owner. The GitHub sync is the tell. Hans Bala tours us what shipped in the video below.
-
Emrecan Dogan shared thisWith 3 of the 104 World Cup matches left to play, FIFA reviewed every variable in the VAR crisis — the calls, the rules, the officials. For the semifinals, it has fixed the seating. :standing_ovation:
-
Emrecan Dogan shared thisVoice gets powerful when it can finish the job. We’re launching voice actions and artifacts in Glean. I can now speak to Glean to send a Slack message, create a Jira ticket, add a Google Calendar invite, or create a document artifact. Voice has been a big leverage for a few months now. But, this launch is a big shift from voice that informs to voice that performs. These are my repeated tasks that I get to run fully on voice now: - Leave a customer meeting and turn fresh thoughts into the right Slack or doc follow-up. - Talk through product review direction on a bike ride and shape it into a one-pager for product, engineering, and design. - Capture a roadmap follow-up and turn it into a Jira ticket or update. - Record interview feedback and draft a structured candidate note. Less drop-off between thinking and follow-through. Spoken intent becomes a completed next step.
-
Emrecan Dogan shared thisAI is expanding what builders can attempt faster than it expands their judgment. That’s a good problem, if leaders design for it and raise their own intensity to match their's teams' pace, engage and calibrate. I wrote about a major upside I’m seeing across many R&D orgs. As builders can build in parallel, with shorter cycle time, leaders need to show up with higher pace as well. More people in the org needs the right calibration on the judgment and taste that leads to building the right things and building the thing right.Expand Scope Aggressively. Make Calibration Unavoidable.Expand Scope Aggressively. Make Calibration Unavoidable.
-
Emrecan Dogan shared thisIn a large enterprise, there are now many front doors to AI. Many models. Many tools. Not enough connective tissue between them. That’s why we've just shipped Glean AI Gateway into private beta: a governed layer between the AI surfaces teams want to use and the models + MCP tools enterprises need to manage. The goal is not to force everyone into one AI app. Claude Code, Codex, ChatGPT, Claude, internal apps, and agents will all have their place. But every new surface should not create a new model connection, permission boundary, cost center, and visibility gap. The pattern I think will define the next phase is the “narrow waist”: many AI experiences above; many models and tools below; and one governed layer in between for context, security, observability, and tokenomics. That means leaders can give teams more freedom while still controlling model and MCP access, enforcing permissions and security policies, tracking token spend by user/app/model, seeing MCP usage, and managing costs with routing and fallback to cheaper models when appropriate. More on Glean AI Gateway here: https://lnkd.in/ek8uv6va
-
Emrecan Dogan shared thisWe just launched GPT-5.6 in Glean. ☀️ Sol for harder reasoning. 🌎 Terra for everyday work. 🌙 Luna for faster, lower-cost agent workflows. Some of my agent runs immediately got better and faster with the 5.6 lineup. But the more nuanced point is about where enterprise AI is going. Model choice is becoming part of the operating layer. The frontier is moving too quickly for companies to anchor every workflow on one model. Different jobs need different tradeoffs: reasoning depth, latency, cost, openness, reliability, and agent-readiness. Users should not have to chase the model frontier. They should get the right model for the job inside the tools, context, permissions, and workflows they already use. That is the value of model choice in Glean: new capability becomes immediately useful, without turning every model shift into another migration project. Models will keep changing. Work will get done. Craft, though, gets a bit better as GPT 5.6 lineup is added to model choice in Glean.Emrecan Dogan shared thisGPT-5.6 is available in Glean ☀️🌎🌔 Now, you have more flexibility to choose the right model for the job: Sol for harder reasoning, Terra for everyday work, and Luna for faster, lower-cost agent workflows. That’s what customers value about model choice in Glean: when the LLM landscape shifts, teams get access to new options in the workflows they already use, without compromising on quality. Try OpenAI GPT-5.6, alongside 30+ frontier and open LLMs in Glean today!
Emrecan Dogan commented on a post
3w
The feeling is very mutual Jayanth Mysore. You and the Iris team have brought exceptional expertise, founder-level urgency, and deep customer instinct from day one. We’re lucky to have you, and I couldn’t be more excited about what we’ll build together at Glean. Welcome!
Emrecan Dogan commented on a post
1mo
Paul Vetere, this made my day.
There’s no better product validation than someone seeing the problem up close at Gartner, then asking during their interview whether Glean would be there to help them ramp.
Really glad it’s helping you get oriented faster at Everpure. Onboarding is one of those visceral moments where company knowledge either becomes leverage or becomes friction. Excited to see the next use cases you uncover.
-
Emrecan Dogan liked thisEmrecan Dogan liked thisA Fortune 50 exec getting on stage to highlight the fact that Glean has been the most transformative platform implemented in his 20+ year career… tells you something about impact + product-market fit. Plus the fact they rolled out to 60K+ users in < 3 months… tells you something about demand, adoption, scalability + trust. Live from SF, GleanGO was quite an impressive showcase of the real enterprise transformation we’re driving in partnership with our customers across industries. Their stories are incredible. Stoked to be building alongside our unbelievable team, customers and partners as we continue to lead the shift from AI experimentation to real business impact + execution at scale. Intelligence era = proactively turning context into action + maximizing token yield. Couldn't be more excited to be here. Quite a few heavy hitting product announcements coming out of GO... 👀👇 + was fortunate to cap off the week with a few days in Carmel ! 🌊
-
Emrecan Dogan liked thisEmrecan Dogan liked thisDon’t post much on LinkedIn but I feel like this one is worth it. One of my favorite parts of my job is building programs like Glean:GO. I get to work with some incredible people from all across Glean for months to make something our customers can actually experience. There’s a lot that goes into pulling something like this off, and I’m really proud of everyone who had a hand in creating it. A special thank you to the entire marketing team and Glean’s global events team - Melissa Munro, Anna Chibukhchyan, Jade Hook and Emily Franklin + a great partner in CG CREATIVE STUDIOS.
-
Emrecan Dogan reacted on thisEmrecan Dogan reacted on thisAs August folds, I can’t let it pass without a note of gratitude. Thank you, Tim, for a masterclass in leadership and integrity. We will miss having you around, and I hope to catch you on the next hike. #Timcook
-
Emrecan Dogan reacted on thisI was interviewed by Ravit Jain at the Glean GO conference, where we covered topics ranging from product building, to AI in data and the inventions we are creating at Glean to help business teams make better decisions. Check it out !Emrecan Dogan reacted on thisYour dashboard tells you revenue dropped 8% last quarter. It cannot tell you why. The why is sitting in a Slack thread, 3 support tickets, and a deal review nobody wrote up properly!!!! That split is the most underrated problem in enterprise AI, and it is what I got into with Jayanth Mysore, Product Manager at Glean, at Glean:GO on The Ravit Show in San Francisco. Jayanth co-founded Iris, an AI analyst that sat directly on the cloud data warehouse and answered questions in plain language with SQL behind it. Before that he was an early product leader on Google Analytics and VP of Product at Sigma Computing. He has spent a career on the structured side of this problem. Iris is now part of Glean, whose strength has been the opposite half, the documents, conversations, and tickets where the reasoning lives. Most enterprise AI still picks one side. You get a tool that queries your warehouse but has no idea what your company decided last month, or a tool that reads every document but cannot count. Neither one answers the question an executive actually asks, which is what happened and why. We talked about what it really takes for AI to understand a company rather than generate plausible answers, where enterprises still underestimate context, how to decide what AI automates versus where a human stays in control, and what customers are asking for now that they were not asking for a year ago. Full conversation on The Ravit Show. Link in the comments. #glean #enterpriseai #ai #dataandai #artificialintelligence #analytics #cdo #theravitshow
-
Emrecan Dogan reacted on thisEmrecan Dogan reacted on thisJust back from Glean:GO 2026 at Fort Mason in San Francisco. Great two days of conversations with people building agentic AI for the enterprise, including Emrecan Dogan, Arun Kumar, and Anuraag Gupta! 🙌 Here are three things that stuck with me: 🚀 Independent agents as AI coworkers: The conversation wasn't about better answers to prompts, but an agent that notices something needs attention and proactively suggests a fix or just handles it. It's a big shift from "AI you use" to "AI that works alongside you". 🧠 Auto-routing as the sustainability lever: The first wave of enterprise AI was all about getting people to use it. The next wave is about proving it's sustainable. Now that the bills are coming in, companies are realizing not every task needs the biggest, most expensive model. Automatically routing each prompt to the right model is the real cost and performance strategy. 🤝 Interactive artifacts and team collaboration features: I'm genuinely excited about turning AI output into something a team can actually work with together, instead of a static answer one person got, which is where a lot of real workflow value will show up next. 🛡️ Trust and governance are the real unlock: Every customer story came back to the same tension: teams want AI that acts on their behalf, but only with security and guardrails baked in. Human-in-the-loop isn't a limitation to design around, it's the principle that lets us move fast without betting the company. Thank you Glean team for putting together two days of engaging content. Looking forward to what's next! #GleanGO2026 #EnterpriseAI #AIAgents #ai #glean #sanfrancisco
-
Emrecan Dogan reacted on thisEmrecan Dogan reacted on thisThanks Amazon for all the fun (and the fish!) for these last five years. I’ll always remember my first product strategy leadership review where I projected 3 year revenue of $500M and was told to come back with a path to $1bln so it’d be worth investing in. Since then, I’ve had the good luck to lead global products with multi-billion dollar revenue, working alongside some of the brightest and most passionate people I’ve known and strategizing with executives from largest enterprises and startups. What’s next? I’ll be working with some physical AI goodness. 🦾🤖🦾
Experience & Education
Licenses & Certifications
Patents
-
Assessment-Based Opportunity Exploration
Issued 11205144 16/359,793
The disclosed embodiments provide a system for processing data. During operation, the system obtains a first assessment result related to a first qualification of a candidate. Next, the system filters a set of opportunities based on a first comparison of the first assessment result with qualifications for the opportunities to produce a first subset of opportunities. The system then selects a second qualification associated with some or all of the first subset of opportunities to assess for the…
The disclosed embodiments provide a system for processing data. During operation, the system obtains a first assessment result related to a first qualification of a candidate. Next, the system filters a set of opportunities based on a first comparison of the first assessment result with qualifications for the opportunities to produce a first subset of opportunities. The system then selects a second qualification associated with some or all of the first subset of opportunities to assess for the candidate. The system also filters the first subset of opportunities based on a second comparison of a second assessment result related to the second qualification of the candidate with additional qualifications for the first subset of opportunities to produce a second subset of opportunities. Finally, the system outputs recommendations related to applying to the second subset of opportunities by the candidate.
Other inventorsSee patent -
Assessment Deployment Auto-Pilot
Filed US 16/577,241
See patentTechniques for automatically deploying assessments over a computer network
-
Assessing Job Metadata Quality And Automatically Suggesting Corrective Actions
Filed US 16/556,682
See patentTechnologies for suggesting changes to job posting data to increase performance of a job posting for a job opportunity.
-
Mapping Assessment Results To Levels Of Experience
Filed US 16/359,777
See patentA system that, based on the assessment result, approximates a position of the proficiency rating in a distribution of assessment results associated with the qualification and one or more attributes of the job opportunity.
-
Screening-Based Opportunity Enrichment
Filed US 16/359,757
See patentApplication of a machine learning model to attributes of a job opportunity to generate a set of confidence scores between the opportunity and a set of screening questions.
-
Assessment-Based Qualified Candidate Delivery
Filed US 16/359,761
See patentA machine learning model applied to a job seeker's assessment answers and additional qualifications for an additional opportunity to produce a match score between the candidate and the additional opportunity.
Languages
-
Turkish
Native or bilingual proficiency
-
English
Native or bilingual proficiency
-
Spanish
Limited working proficiency
Other similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top contentOthers named Emrecan Dogan
14 others named Emrecan Dogan are on LinkedIn
See others named Emrecan Dogan