How agentic AI platforms will redefine enterprise applications

We imagine enterprise applications are undergoing a profound change. By next yr, highly capable agentic systems will emerge to create latest application classes and alter the best way organizations take into consideration their backend systems, data platforms and user interfaces. The extent of the transformation, we imagine, can be more impactful to the appliance stack than were the changes led to by innovations seen throughout the modern graphical user interface, Web, cloud and mobile eras.

Though many persons are talking about agents, only a few in our view have thought deeply concerning the potential of deploying and orchestrating armies of lots of or hundreds of agents to more fully automate enterprises. These capabilities are more likely to come from application vendors, data platform providers, the cloud players and a select few innovators.

image here?

This trend will more than likely affect white-collar productivity in the best way that mass production boosted labor productivity. Our expectation is that ultimately, intelligent, agent-based systems will have the ability to finish many common business processes with 1/tenth of the headcount required today.

On this Breaking Evaluation, we construct on our previous work on agentic systems and dig deeper into the subject. We specifically deal with how agent-based automation will alter the best way applications are built, data is accessed and businesses operate.

Today’s applications are islands of automation

As we’ve previously discussed, today’s applications are a group of bespoke systems which have data, business logic and associated metadata locked inside proprietary platforms. Business applications generally address narrow business challenges and connections to data in these systems necessitates a separate data platform that serves because the historical version of the reality. Operationalizing the insights from this data is an asynchronous process that falls wanting industry guarantees to create a real-time, 360-degree view of the business.

True end-to-end automation is virtually unimaginable with today’s legacy applications

Below we show a number of the leading application platforms. They’re most frequently loosely connected systems, which makes it difficult to view the state of a business in real time. We’ve been attempting to bridge these islands of automation for many years. There’s an alphabet soup of acronyms like enterprise application integration or message-oriented middleware, web services, even microservices.

But they’re all hard to take care of and evolve because they’re too low-level. They’re further challenged because they have to translate and navigate between applications that talk fundamentally different languages. As such, they’re brittle they usually automate reach a portion of the appliance estate.

The underside line is, across enterprises, we’ve essentially created software-defined departments, which feed exhaust data to a central repository that can provide us a snapshot of the business at a cut-off date, but lack organicity and the flexibility to affect within the moment actions which can be trusted.

Agentic AI has the potential to unify the silos

We see an emerging opportunity where systems of multiple agents can attack the islands of automation problem. We’ll show you a more full picture in a moment. As a preview, we use the graphic below where we envision agents which have authority and resource access to permit them to work with human supervision toward a set of goals which can be measured by the identical KPIs we now have in our dashboards today.

Agents will work in a coalition with other agents and reply to actions and changes available in the market while optimizing for a top-level business goal comparable to customer satisfaction, market share, profitability or growth. These agents will complement and augment existing automated processes. Importantly, they’ll learn from these existing systems and human actions and do things that may’t effectively be hand-coded.

As such, we see a completely latest stack forming on top of today’s existing systems. We don’t see this as a rip-and-replace but reasonably an evolution that may happen over the subsequent five to 10 years. We imagine those firms which may benefit from the transformation will radically improve their productivity by deploying multiple agents that may work together driving AI-native processes that may disrupt businesses that don’t respond.

Reimagining the enterprise software stack

Below is a more comprehensive view of what the emerging intelligent application stack will appear to be. We’re going to walk through the salient points of this diagram, but as a setup we see five critical areas inside this picture.

1. The connection to existing backend operational apps and analytic systems (bottom layer).
2. The information platform architecture above that, which can turn into increasingly abstracted in our view.
3. A so-called semantic layer that should be built, which harmonizes all of the disparate data elements within the enterprise such that data definitions are consistent, trusted, sharable and, after all, governed.
4. A capability to deploy, manage, orchestrate and optimize multiple agents, as shown within the upper right.
5. The agents are guided by business metrics, represented within the upper left, based on top-down goals.

A bidirectional model of an enterprise

Consider this as a bottom-up, top-down model of the business that we imagine will dramatically improve productivity and alter the appliance landscape. In our view, this can be a profound transition for the industry, unlike any we’ve seen before since it affects each the demand for the software through user interface innovations that make applications accessible in latest and different contexts, but additionally radically changes the productivity of each aspect of the software development lifecycle.

After all, we’ve had top-down “intergalactic” approaches previously which have attempted to integrate the enterprise. But they’ve never succeeded. One example is enterprise data models that were too difficult for in-house developers to take care of. That’s essentially how industrial off-the-shelf software spawned a brand new breed of applications. As well, we’ve seen bottom-up approaches attempting to model every little thing, for instance, in microservices, but that created a “Tower of Babel” that we only partially reconciled downstream by bringing data into a knowledge lake or the fashionable data stack.

All these approaches resulted in the identical islands of automation or analytics that spoke different languages. Agents, that are the subsequent iteration of artificial intelligence large language models (we expect of them as large motion models), can bridge these islands by helping them work together. Eventually, we envision agents from vendors which may exist in numerous applications today to consult with one another essentially in natural language — once we now have the right scaffolding in place, as shown within the diagram above.

Not only will agents have the ability to speak together on this common language, they’ll have the ability to learn what’s occurring within the business by observing the breadcrumbs of business activities, all those activities that may’t be explicitly coded. That brings the bottom-up approach that wasn’t technologically possible before.

Now let’s walk through the above diagram and double-click into the salient points of the five elements we cited above.

Connecting to backend legacy applications and data

Below we show examples, including Salesforce Inc. and its MuleSoft asset, Microsoft Corp.’s Power Platform and automation specialists UiPath Inc. and Celonis Inc., as just a couple of of the firms we see having the potential to contribute to this vision.

It’s necessary to consider that, like all transitions, we don’t sweep aside every little thing that got here before. These technologies must coexist with and construct on the legacy enterprise software stack that has been built over many years investment. To do this, we’d like to begin with these two-way connectors shown above in order that agents can pay attention to what’s occurring within the business by talking to existing systems. The explanation we emphasize two-way connectors is that they’re going to should perform transactions to update what’s in the present systems of record that today are mostly islands.

For instance, let’s consider the idiosyncrasies of product returns that require “tribal knowledge” to adjudicate and have many permutations, making them too cumbersome handy code. An agent that’s assisting a customer support rep talking to a customer about returning a product might want to consult with an order management system, a list system, a logistics system and the like.

We envision that taking place with a number of agents observing the human process, learning over time and ultimately interacting with these connectors that may interrogate or transact with these backend systems. At first, it can only be one agent that may know easy methods to consult with these different systems because they’ll should be trained and even hard-coded on easy methods to connect. But over time, you’ll have agents which can be more adaptable they usually’ll have the ability to consult with and call on one another. Importantly, all of the work currently getting done with existing applications can be critical constructing blocks that agents evolve.

The underside line is these connectors are the place to begin and firms which can be constructing them can have an early advantage.

The evolving data platform

Every 10 years or so we see the emergence of a novel data platform approach — from the unique database management system to the info warehouse, data marts and Hadoop data stores to today’s modern data platform represented below by Snowflake Inc. and Databricks Inc. Innovation in data platforms is constant.

We’ve covered Databricks and Snowflake extensively and the numerous changes which can be occurring in that market. Specifically we’ve covered the shifting point of control from the DBMS to the governance layer, which is becoming more open; and the movement of value up the stack into the appliance layer.

Though there’s plenty of work to do, specifically around governance, above we show those two leaders plus some latest entrants comparable to Starburst Data Inc., which is taking a federated, data mesh approach, and a few existing application vendors comparable to Salesforce with its data cloud and Microsoft constructing abstractions across its data estate and unifying its metadata and governance model. After all, Google LLC and Amazon Web Services Inc. can be participating as well, but we see the five names shown above as actively making moves on this space.

The information layer is the historical source of truth for the enterprise. It captures in a single logical place what has happened, why it happened, what is going to occur and even what should occur. In other words, that is the context that guides agents and their activity, and that’s partly why the move from a DBMS-centric to a metadata-centric approach to the info platform is so necessary, because that’s what allows multiple compute engines or tools to access the analytic data estate at anyone time.

Today, those different engines could be machine learning tools or training your personal LLM, but in the longer term we’ll see multiple agents talking to this data. That’s why we’ve stated that a single DBMS as the purpose of control is a bottleneck. Today’s data platform is unfinished because all the info that it pools has been stripped of the context and meaning that was captured within the apps or those islands of automation.

As Zhamak Dehghani described in detail at Supercloud 7, this business context (metadata, governance, metrics and the like) has been bolted on top of a centralized data store, fed by brittle data pipelines and has created unmanageable complexity for humans.

The underside line is that each one this disparate data must be harmonized, which leads us to the subsequent layer of the stack.

A layer to create common meaning

Let’s keep moving up the stack and deal with what is typically called the semantic layer but we also seek advice from it below because the harmonized data layer.

It is a big missing link today where we’re attempting to rationalize all of the complexity around data pipelines and data products and multiple data types, bolted-on metadata and governance models, metrics layers that each one ultimately feed analytic applications. And we wish to bring the info together in a way that speaks a standard language of the business and is trusted in order that motion might be taken, or what we show here as transactions. This ties back to the work we did last yr with Uber Technologies Inc. and our vision of Uber for all.

The chart above shows firms coming at this from two directions: 1) The information semantics that we’re seeing emerge from Salesforce with customer 360 (for instance); and a pair of) The metrics layers from firms comparable to AtScale Inc., dbt Labs Inc. and Databricks’ Lakehouse Intelligence.

We’re highlighting two approaches above: 1) Salesforce’s Customer 360, which also goes together with Supply Chain 360, Operations 360; and a pair of) The metric layers with dbt and AtScale as examples. We see these two vectors as two sides of the identical coin. They’re all attempting to connect what looks like disparate data in order that they describe the identical thing as a part of our theme of moving from strings of information (that databases understand) to people, places and things which can be meaningful to the business (that’s, human concepts).

Within the case of Customer 360, you possibly can understand a customer journey and their preferences, for instance. With metrics, you get an outline of all the weather for measuring business key performance indicators comparable to bookings, billings and revenue. Again, it’s getting away from strings and moving into business language.

But each of those two approaches only solves a part of the issue of attending to people, places, things and activities. These standardize the language around things or easy measures. They don’t standardize and harmonize the activities or business processes that link and span all this stuff.

That’s a much harder problem. This is the reason we show above RelationalAI Inc., EnterpriseWeb LLC, Palantir Technologies Inc. and Celonis. They’re each coming at the issue from a unique angle and are attempting to unravel this harmonization problem.

As an aside, in our view it’s inaccurate to suggest that the so-called modern data platform is just not going anywhere and can not be necessary within the age of AI. Among the pioneers on this space comparable to Tristan Handy of dbt Labs have written about how interest and valuation from the investment community has moved on from the info platform, which is true from a the attitude of a capital allocator.

But there’s a contrary viewpoint, which is that AI will put way more pressure on the info platform to innovate in order that it may harmonize first the analytic data after which eventually the processes and the operational apps. Agents can be so way more powerful and productive once they can navigate across these modern data and application estates once they have one uniform map of the world of an enterprise and its ecosystem.

What’s not clear yet is whether or not today’s data platform vendors (that’s, Snowflake, Databricks and so forth) will lean in to this layer and be the source of the harmonization technology; or a brand new wave of corporations will attack this problem and suck more value out of today’s data platforms.

Enter the intelligent agent layer of the stack

Large language models turn into large motion models

Let’s move north up the stack into what may emerge as one among the more beneficial pieces of real estate in the subsequent 10 years, the agent operations within the upper right corner, as highlighted below. There’s so much more work that should be done and latest tooling that should be created, but that is where the vision becomes reality.

Importantly, we’re not taking humans out of the equation, reasonably we’re enabling agents to work in concert with one another to grasp a firm’s objectives, adjust plans because the business evolves, and make humans and organizations dramatically more productive by doing repetitive tasks, inferring from human actions and presenting plans to humans for approval.

We predict that the shiny latest toy next yr that everyone seems to be going to be talking about is agents. This in a way is a reincarnation of LLMs from large language models to large motion models, or LAMs. Everyone has been talking about individual agents or copilots. In our view, that’s like talking about individual microservices. They turn into really effective only once they work along with other agents (or services) to perform more sophisticated, compound tasks, which we call business processes, constructing on previous organizational muscle memory.

To do that, we’d like an agent control framework. We see this layer as an important a part of the brand new application model, or what we call scaffolding, that’s emerging. We used to have an application model that existed in three logical tiers: 1) An information model which was stored within the DBMS; 2) Application logic, which was its own type of separate tier; and three) A UI or presentation tier.

This model gave us the islands of automation because each was its own stovepipe, even after we got to the microservices era. Though it’s still evolving in our minds, when this latest model matures, it can allow agents to consult with one another across applications and from different vendors.

Organizational metrics guide the agents

A tree of metrics shown within the upper left quarter can also be critical since it informs and guides the agents. On the very top of that tree are the north star objectives, or objectives and key results, comparable to “grow the business by X,” or “expand the ecosystem by Y” and so forth. At top of the tree, we’re not only financial metrics, but reasonably overall organizational objectives, key results and the actions around them. Beneath that, we envision all of the components that drive or influence those north star metrics. This tree of metrics metaphorically measures and represents a model of how the business runs.

Then we envision an this org chart (shown to the best) of all of the specialized metrics and the associated responsibilities to optimize these (MRR, NRR, market share, ACV, LTV, CAC and so forth). Each agent has specialized expertise, but unlike microservices, they’ve more intelligence to determine easy methods to accomplish a task beyond what’s explicitly specified — easy methods to do product returns, for instance — they usually get well over time as they learn and iterate.

They do that by: 1) Measuring their performance against the outcomes which can be within the metric tree (e.g. customer satisfaction, cost to serve, etc.) and a pair of) By observing their human supervisors and by capturing the knowledge of domain experts who edit the plans that they generate.

The underside line is that this is different from procedural code that we’ve used for many years because now we will capture tacit knowledge, and we will capture the 80% of enterprise activity and processes that previously couldn’t be specified top-down in some Newtonian mechanical algorithm.

A brand new enterprise application paradigm emerges by the top of the last decade

We’ve come full circle here. Let’s bring back the complete picture again and summarize our takeaways.

The evolutionary a part of this method is it’s built on top of existing cloud and on-premises systems with deep connections into these critical operational applications in addition to the historical systems of analytic truth.

A unified and cogent metadata model must emerge from a mix of existing data platforms and a system that ties together multiple data estates right into a single federated data management system.

A layer evolves that harmonizes the info and business logic in addition to the method definitions which can be established by the business. This supports a highly beneficial agent framework that we imagine can be built to orchestrate multiple agents which can be working together toward a business final result, defined by business goals. The progress toward those goals is measured by key performance indicators that comprise OKRs that might be fine-tuned and adjusted as business conditions change.

All of that is transparent to human overseers who can adjust plans and supply feedback from which agents can learn over time.

Moving beyond RAG: How long will it take and where is it starting?

On the timeframe, Salesforce is more than likely to be the primary, a minimum of to announce a military of agents and a framework to make it work, we imagine. Next month at Dreamforce, the corporate will unveil more clarity that may align with our vision. We imagine Microsoft will likely be the second at Ignite, its conference for his or her partners in November. Each of those firms are moving on this direction and we expect they are going to further show a vision for this transition to armies of agents versus all of the type of simmering activity we’ve heard about single agents added within the ecosystem.

That is so way more sophisticated and useful than retrieval-augmented generation, or RAG. On this world we envision agents doing work on behalf of a human supervisor and dealing in concert with other agents. They learn over time and cooperate with other agents to unravel complex tasks by constructing on easy expertise that every of them possesses.

Crucially, these agents construct on the enterprise software that’s been built up over many years. And to be very explicit, what’s in symbolic software, traditional software, is what must be repeatable, precise and accurate, and that might be specified. Consider it as Newtonian mechanics, whereas what’s in agentic AI, there’s a protracted tail of “dark enterprise matter” of activity that you simply couldn’t capture in these Newtonian rules, but you would only learn bottom-up. And so they should coexist.

To be clear, this can be a piece of our vision that’s evolving and is opaque today — specifically, exactly where that layer of coexistence comes from, because within the Newtonian world, you’re still going to wish to align all of your activity with very precise and repeatable analytics. But at the identical time, you’re going to wish to align the activity of your agents.

To perform it will require a really different mechanism. Where those two alignment technologies come together – that’s, how do I align the Newtonian activities and the agentic activities — is the most important open query and can take the higher a part of a decade to evolve.

A key point is it doesn’t make the info platform less relevant. It actually will make it and the layer that it must grow into more necessary. What’s not clear yet is who’s going to innovate and create that layer above the info platform that harmonizes.

To emphasise, agents are going to be the subsequent big thing that captures everyone’s attention, especially armies of agent and the framework that aligns them, but they actually need a harmonized estate underneath to make them work and make them perform to our vision.

Spending data for some aspiring agentic AI players

Let’s herald some of information from Enterprise Technology Research. Within the chart above, we cherry-pick a lot of enterprise application and data platform firms. The information shows Net Rating on the Y-axis and penetration or overlap in the info set on X-axis. That is data from greater than 1,700 accounts that ETR is surveying here. Net Rating is a measure of spending velocity on a platform.

Remember, this technique measures account penetration and has no indication of actual dollars spent.

Salesforce has a dominant presence, as does ServiceNow. OpenAI continues to steer virtually all corporations by way of Net Rating within the enterprise. Snowflake can also be outstanding. Databricks has momentum, as you possibly can see within the vertical axis. Google Workspace and SAP SE are comparably positioned, and we also show Microsoft’s Power Automate as a representative of its overall Power Platform.

We’re also showing a pack of Workday Inc., Salesforce’s MuleSoft, which is strategic for Salesforce, UiPath, NetSuite, which is Oracle Corp., also Oracle Fusion and IBM Corp.’s Watson, which we expect has got a play here. There are also Palantir, which we referenced earlier, in addition to Celonis, after which Infor for context as one other legacy application company to which there are going to be connections.

Each of those firms has a unique approach. The opposite call-out that’s really interesting is the frontier models comparable to OpenAI’s GPTs, Google’s Gemini and Anthropic. They’re all attempting to construct general-purpose agents with these open-ended reasoning and planning capabilities. And that’s why we did an earlier Breaking Evaluation where we said those constructing these open-ended consumer models are like adventurous sailors within the Middle Ages who’re potentially sailing off into the top of the earth and can fall off the top of a flat plane. It’s a really open-ended, difficult problem.

Then again, take someone comparable to Salesforce or Microsoft, and their agents take a look at only a small piece of a well-defined map of the enterprise, they usually realize it’s not an open-ended search and navigation problem. You’ve gotten a alternative of a half-dozen things you would possibly have the ability to do, and you possibly can mix and match easy methods to get them done. It’s a way more tractable and finite problem. So there’s somewhat little bit of mix and match, apples to oranges, on this chart.

And it makes us think back 25 years ago. First, we had business-to-consumer commerce attempt to take off, however it was such a general-purpose problem that the business-to-business commerce had a smaller number of consumers with way more well-defined activities.

That business took off first in a way more meaningful way. And similarly, we expect these enterprise agents are going to get to maturity much faster. So Palantir, for instance, has built a mature foundation across largely the legacy transactional systems within the enterprise. Celonis learns the capabilities across different business systems bottom-up by reading the logs and understanding what’s occurring. UiPath has the chance to learn each from application programming interfaces and screens so you possibly can get to legacy apps that don’t even have APIs.

They’re not semantically harmonized yet, but they’ve the richest and easiest-to-use tools to make it possible to construct armies of agents. They don’t yet have the agent control framework for armies of agents, but we expect we’ll see that this fall. And we expect we’ll see Salesforce put all of the pieces together, a minimum of for working within the context of Salesforce, next month at Dreamforce.

The underside line is we’ll begin to see the several approaches bringing a number of the pieces together and over time filling within the blanks. Nobody firm goes to have all of the pieces.

Agentic timeline and scenarios to look at

Our research indicates that the evolution toward agentic systems is ready to rework the landscape of information management and application development profoundly. As we transition into the 2030s, we imagine the emergence of intelligent, multi-agent orchestration will play a pivotal role in bridging the gap between legacy data systems and modern application needs.

The growing prominence of open table formats and the evolving concept of information mesh illustrate the continued shift toward more flexible, decentralized data architectures. Nonetheless, significant challenges remain, particularly around governance and the combination of disparate data sources right into a unified, harmonized platform.

Looking forward, we expect major cloud providers comparable to AWS, Google and Microsoft increasingly to deal with simplifying and unifying data management to support the event of intelligent applications. The anticipated advancements in low-code and no-code tools, alongside the push toward comprehensive agent frameworks, are more likely to redefine the competitive dynamics in the appliance platform market. As these trends unfold, traditional data and platform platform vendors might want to adapt quickly, moving beyond their current roles to stop being overshadowed by latest harmonization layers that may dictate feature adoption.

In conclusion, though the trail to a totally harmonized and agentic system landscape is fraught with complexities, the potential for enhanced organizational productivity and agility is immense. As we proceed to observe these developments, it is obvious that the flexibility to adapt and innovate can be crucial for players looking to take care of their competitive edge on this rapidly evolving market.

We see the second half of this decade moving beyond the thrill of generative AI, large language models and opaque return on investment within the enterprise. Agents working in concert to dramatically automate businesses have the potential to deliver on the promise of AI and convey productivity improvements that may propel the worldwide economy to latest levels of growth and dynamism.

Though much invention is required to satisfy the promise of AI, the technologies are coming together to make our vision a reality, where an organic and ever-changing digital representation of an enterprise and its ecosystem is delivered as a part of a brand new framework for applications. Built on many years of existing legacy applications and software-as-a-service solutions, we imagine, intelligent agent-based systems can be the subsequent evolution in application architecture.

What do you think that? Are there firms working on this problem that we’ve not mentioned? Please tell us so we will proceed to map the longer term of enterprise data and applications for the community.

Images: theCUBE Research

Disclaimer: All statements made regarding corporations or securities are strictly beliefs, points of view and opinions held by SiliconANGLE Media, Enterprise Technology Research, other guests on theCUBE and guest writers. Such statements aren’t recommendations by these individuals to purchase, sell or hold any security. The content presented doesn’t constitute investment advice and mustn’t be used as the premise for any investment decision. You and only you might be liable for your investment decisions.

Disclosure: Lots of the corporations cited in Breaking Evaluation are sponsors of theCUBE and/or clients of Wikibon. None of those firms or other corporations have any editorial control over or advanced viewing of what’s published in Breaking Evaluation.