We Love Ugly Data! The Deep Analysis Podcast

Alan Pelz-Sharpe
  • 5.0 (1)
  • Technology
  • Updated Bimonthly

This irregular podcast series examines what is happening in the unstructured data automation market. Three topics - Thirty Minutes, that's the format!Topics range from the state of Blockchain, IDP, ECM, and the impact of AI on unstructured data. Deep Analysis provides advisory services, industry research, and M&A guidance. info@deep-analysis.netDeep-Analysis.Net

  1. 2d ago

    SE05E0 - IDP Special (with Pope Encyclical analysis!)

    The fortieth episode of the podcast that you know and love as "We Love Ugly Data!" is out; available in audio form everywhere you get your podcasts from and additionally in video form via YouTube (which we've embedded below). This time, it's Matt and Dan for an IDP-focused episode, talking more about new vendor profiles, where all the cash is going in IDP, and how The Pope is contributing to AI governance (sort of). In this month’s episode: Topic 1: New Vendor Profiles It's time for another set of vendor profiles from the recently published 6. Last time, Matt and Alan talked about 3 of them; this time, Matt and Dan cover the other 3. First up is Box and Box Extract. Dan explains that the company had made some previous efforts around IDP with Box AI, but then acquired the Polish vendor Alphamoon in 2024, the results of which now appear as Box Extract. As a content management/content intelligence, it's one that Box itself is very excited about; CEO Aaron Levie described it as a "killer app," as Dan noted in a related LinkedIn post.  Next up is Indico Data, a specialist in commercial IDP, and Dan discusses how that market focus came about, as well as some details on their technical background. Matt mentions that this market has a bunch of well-understood processes, and that if people are interested, there's a range of other related vendor profiles in the Deep Analysis database that are worth noting. Finally, Retaurus, who didn't come to IDP from a document-processing background but from the messaging EDI space. They've been a big player in messaging and especially fax, which remains an important document exchange format in some industries, for a long time, and have built an IDP platform to help augment their customers, especially in supply chain and logistics. Topic 2: Big money in IDP 🤑 Dan wrote back in April about how money is pouring into IDP (pulling some data from his latest IDP Market Analysis), and here explains how the latest iteration of the technology that we're referring to as IDP came about in 2018/19 when significant money started to be invested in vendors like Hyperscience, Instabase, Indico Data, and others. In May this year, another well-backed IDP vendor, Rossum, was acquired by Coupa (they were already lined up as partners), which validated the value of IDP in this case, the spend management industry (Rossum had focused on selling into accounting teams). Reducto - who sounds like it should be fighting Spiderman, but as far as we know, doesn't - is an LLM-native IDP platform that's raised huge Series A and B rounds, that Dan estimates is valued post-money after the latest of these at $600m. Apparently growing revenue at a clip, selling into other AI peers. Matt points out that it's not uncommon to see a lot of peer selling between venture funding companies with investors in common. Topic 3: What's the Pope on about? At the end of May, the Pope got involved in the AI debate by publishing an encyclical titled "MAGNIFICA HUMANITAS; ON SAFEGUARDING THE HUMAN PERSON IN THE TIME OF ARTIFICIAL INTELLIGENCE," and, being Dan, he decided to discuss this matter with Claude. Matt points out that it was published ahead of Gartner's first MQ on AI Governance, suggesting that where The Pope leads, Gartner follows. In practical terms, Matt and Dan go on to discuss that the primary area of AI governance customers are asking about now is cost control policies.  https://www.youtube.com/watch?v=h3xAcreasy0  Related Links for Series 5 Episode 6 New vendor profiles in a brief blog post: "Six New Reports. No Hype. Just the Good Stuff." Dan's LinkedIn post on Box and Aaron Levie's excited commentary on Box Extract. Investors continue to pour money into IDP - Dan discusses the financial attraction of IDP IDP Leader Rossum is acquired by Coupa - Dan's announcement analysis blog post. How Reducto Became the Fastest-Growing IDP Startup I’ve Ever Seen The Pope's encyclical on AI and Dan's chat with Claude on this subject. Support the show

    39 min

  2. May 28

    SE05E05 - Architecture, Engineering & Construction (AEC)

    The thirty-ninth episode of the podcast that you know and love as "We Love Ugly Data!" is out; available in audio form everywhere you get your podcasts from and additionally in video form via YouTube (which we've embedded below). It's Matt and Alan again this month, talking about new vendor profiles, a brand-new report on the AEC industry, and 4 key areas for agentic automation to focus on. In this month’s episode: Topic 1: New Vendor Profiles It's time for new vendor profiles, and we've just published 6. In this podcast, Matt and Alan talk about 3 of them (when Dan joins Matt for the June edition, they'll cover the other three). This time, they discuss agentic automation and Hyland's content platform (and how it fits into their new strategy under their new leadership). Then, Papyrus's Owlfie conversation AI approach to automation, and finally, novel AI-led retail process automation from European start-up Duvo (whose profile is available to read for free on its website). The pair also hints at a number of related reports, including a market sizing report soon to come. Topic 2: AEC Report On top of those vendor profiles, we've also just released a new report on the AEC (architecture, engineering, and construction) industry and the technological approaches to managing the plethora of complex documents across a wide range of projects undertaken by the industry's piece-parts and processes. There's a blog post that explains the scope of the report both upfront and throughout the discussion. Alan decides that everyone should get access to the report and open-sources it on the fly. Matt also notes that there's a whole wealth of further AEC material to come from Deep Analysis over the next few months. Topic 3: Runtimes and Processes Last month you'll remember that Matt and Alan discussed the latest release of the Work Intelligence research and since, Matt has been writing a little about what the near-future looks like the all those runtimes and how agentic platforms start to fit within that future. Here he talks about 2 recent posts: "What if AI gets stuck at the task and never reaches the process?" and "Enterprises don’t need another automation runtime unless you give them an excellent reason". The second of these contains "4 potential points of value" that agentic automaton platforms need to fit in order to add real value: problem-solving, flexibility, inheritance, and fraternity. https://www.youtube.com/watch?v=bQAElcD9k2w Related Links for Series 5 Episode 5 New vendor profiles in a brief blog post: "Six New Reports. No Hype. Just the Good Stuff." ....and here's the one on Duvo, which you can read on their website for free right now. Here's the blog post Alan wrote about the AEC report, which you can now also download for free. Matt's 2 blog posts on automation, runtimes, and an agentic future: "What if AI gets stuck at the task and never reaches the process?" and "Enterprises don’t need another automation runtime unless you give them an excellent reason". Finally, here's the blog post that Matt mentions in passing on the "elastic zone of plausibility". Support the show

    37 min

  3. Apr 25

    SE5E04 - Salesforce Goes Headless and other stuff :-)

    The thirty-eighth episode of the podcast that you know and love as "We Love Ugly Data!" is out; available in audio form everywhere you get your podcasts from and additionally in video form via YouTube (which we've embedded below). It's Matt and Alan this month, celebrating the 3rd birthday of the pod (even if the episode numbers don't exactly add up), talking Salesforce, Work Intelligence and M&A. In this month’s episode: Topic 1: Salesforce Goes Headless Starting off, Matt has remembered that it's the 3rd birthday of the pod, with the first episode being uploaded n the 19th April 2023 (just don't ask how come it's episode 38 when the pod has been running for 36 months). Alan has been in San Francisco at Salesforce TDX (the conference for Salesforce developers) and whilst he was there, he wrote a blog post about the announcements and Salesforce's technical progress, most notably the newly announced Headless 360. He explains the change in focus at the conference, from the core audience of Salesforce admins and Apex developers to increasingly opening up to developers that are not coming from a purely Salesforce background, in large part driven by Agentforce and Data 360. Matt points out that Headless 360 is an encapsulation of a technical approach, rather than a product and points to the 60+ MCP servers as the increase in potential complexity - as well as openness - for that extended developer community.  Topic 2: Work Intelligence 2026 Last month, we released the 4th edition of the Work Intelligence Market Analysis, covering the period 2026-2031 (it's available right away for Deep Analysis subscribers as part of their cache of content, it's also available for non subscribers for one off purchase). If you're not familiar with Work Intelligence, there's a free primer from 2022 that's free to download for everyone. Here Matt summarises some of the content in the report, the projected value of the markets and the other details available in the report. Alan points out that even though the concept is now a few years old, it's now clear that the market is aligning with the basis of the world that we described in that initial report. Matt talks about some of the market growth data and then mentions that he's recently published an addendum to the report, which compares our market projections year-on-year. Finally Matt has some scary information about M&A data that's contained in certain other market reports - not written by Deep Analysis - that we've come across. Then there's some kind of related "Buffy The Vampire Slayer" chat. Topic 3: M&A(&AI) Whilst we're thinking about M&A, Matt has decided to mention one of the many pieces of content that Deep Analysis publishes, but sometimes forget to tell you all about. One such item is the quarterly M&A round up for the unstructured data automation market (i.e. the part of the market in which we specialise); who bought who and what market dynamics can be observed. This is available to all Deep Analysis subscribers, who get to see all the deals; large and small. In passing Alan mentions his blog post celebrating OpenText's new CEO's first day at his desk. Matt and Alan then go on to discuss how difficult to use the web as a reliable research tool (for rea Support the show

    35 min

  4. Mar 19

    SE5E03 - IDP Special Edition!

    The thirty-seventh episode of the podcast you know and love as "We Love Ugly Data!" is out; available in audio form everywhere you get your podcasts, and in video form via YouTube (which we've embedded below). Dan's back for an IDP special with Matt this month, focusing on the latest edition of the IDP Market Analysis and discussing this year's top trends. In this month’s episode: Topic 1: Agentic AI Will Transform IDP To kick off, Dan explains the background to the IDP Market Analysis - now in its fourth edition - and that it provides an overview of how the market is performing and how it's looking towards the future, and, in the process, uncovers trends from both vendors and customers alike. Dan goes on to explain that the increase in the use of agents - now 66% of vendors saying that they are present in their product set - is helping those vendors move beyond being a single part of a pipeline process and to take on larger parts of those processes (and in the process, increase their potential for value generation). Dan also points out that 80% of vendors are now using foundation LLMs in their products; a state of affairs he predicted and is now pleased to be able to point to as a fact. Topic 2: IDP Is the Low-Hanging Fruit for ROI The second trend from the report is that IDP is ROI gold. Dan points out that people forget the I in IDP really means AI (which has been around much longer than the transformers that underpin LLMs) and also reminds people that IDP's transactional nature makes it well-suited to finding investment models that generate returns. Using data from our 2025 IDP MMI survey, Dan discussed how reducing processing time is the biggest measurement that customers use for IDP project success (51%), with straight-through processing success -  where processes don't require human intervention, the long-time holy grail for IDP - at 47%. Indeed, Dan points out that the simplest way for anyone to find a quick ROI for AI in their organization is to look at IPD projects, especially as Matt points out, transactionally, these processes are often mature, well-understood, and have known cost bases. Finally, Dan points out that there's a potential reckoning coming as true LLM pricing becomes clear once any price discounting or token bundling from the foundation model providers ends.  Topic 3: Mo’ Projects? Mo’ IDP! Cheating a little bit as Dan has 4 trends in his report, but we only ever have 3 topics on the pod, Matt's smooshed 2 of them together as "Mo' Projects? Mo' IDP!"; firstly, looking at the apparent boom in IDP projects, first identified within that 2025 MMI research, where that boom is largely made up of replacement projects for previous IDP platforms (legacy beware!). Matt points out that the double-digit growth projected for IDP through the analysis period should be contrasted with general IT inflation running at 5%; it's a strongly growing market segment. The second part of this topic is that there are more IDP vendors than ever; Dan's now managing a database of 500 individual vendors at the time of recording. The report itself has 24% more vendors than the 2024 edition and a number of interesting insights into revenue trends. https://www.youtube.com/watch?v=A_eCUOJUujY Related Links for Series 5 Episode 3 The announcement blog post for the new IDP Market Analysis and the page where the report can be purchased (Deep Analysis customers receive this as part of their annual subscription). Our 2026 Predictions download (in case you missed it). The two-part (free) agents reports: AI Agents: What They Are, How They Work, and Where Organizations Might Best Use Them, and AI Agents Part 2: The Evolution from Single Systems to Complex Process Orchestration Pod Episode Series 4, Episode 9, where Matt and Dan discuss the MMI IDP survey in detail. ...and here's that MMI which you can download for free from AIIM. Support the show

    40 min

  5. Feb 19

    SE5E02 We Love Ugly Data - Debunking AI Agents

    30 mins of biting the hand that feeds us - demystifying and debunking the enterprise software market.  The thirty-sixth (not the thirty-fifth, as Matt says) episode of the podcast you know and love as "We Love Ugly Data!" is out; available in audio form everywhere you get your podcasts, and in video form via YouTube (which we've embedded below). Matt and Alan are in the chairs, and they catch up on some recent blog posts on how software isn't dead (again), how crazy promises on AI's future are not helping things, and celebrate that Deep Analysis is now in its 10th year of operation. In this month’s episode: Software is Dead. Again. It's been a year, but apparently software is deal again. And only 12 months since it was last dead. Here, Alan and Matt discuss the recent blog post "The 'Death of SaaS' Narrative is a Luxury the Enterprise Cannot Afford" that Alan posted in response to a lot of discussion about how current models for procuring software are over, and also, whilst we were there, software licensing is over too. The reality, of course, is somewhat different, and having discussed this, they touch on Matt's recent post "Beware the promises made by blank sheets of paper," where there are reminders that enterprises are complicated, and also, as a result, some companies are choosing to hold AI at the perimeter to keep things simple. Stupid AI Predictions Noting that software vendors usually rely on industry analysts to make really dumb predictions, but this time it's Microsoft (again). In an interview with the FT, head of Microsoft AI Mustafa Suleyman suggested that many while-collar roles will be entirely automated by AI in 12 to 18 months. Matt in particularly is very annoyed by this suggestion, describing it as "deeply dishonest", and then explaining why. He also points out that Suleyman's hiring was noted in a previous blog post of his from almost 2 years ago, in which he also suggested watching the "cash at hand" figure for software vendors. That seems to be now attracting a bit more attention. Finally, he also points out that Jad Tarifi of Integral AI has said some equally dumb things, such as that doctors (or lawyers) shouldn't be trained because AI will overtake their knowledge during training. Deep Analysis is (almost) 10 Finally, Alan's written about how Deep Analysis is in its 10th year of operation, some of the areas where we've led thinking ahead of the game, and why we do things the way we do.  https://www.youtube.com/watch?v=darqcAv-Agw Related Links for Series 5 Episode 2 Alan's blog post "The 'Death of SaaS' Narrative is a Luxury the Enterprise Cannot Afford". Matt's accompanying blog post: "Beware the promises made by blank sheets of paper." The FT interview with Mustafa Suleyman (subscription required) "Ex–Googler says degrees in law and medicine are Support the show

    37 min

  6. Jan 18

    SE5E01 - Tech Specificity & Paper Mountains

    The thirty-fifth episode of the podcast that you know and love as "We Love Ugly Data!" is out; available in audio form everywhere you get your podcasts from and additionally in video form via YouTube (which we've embedded below). It's a new year and the beginning of series 5 (despite what Matt says in the intro to this episode, because he lost count) and with Matt and Alan in the chairs, they catch-up on some recent blog posts on the need for specificity, not to ignore paper, how there's no new cash in IT and how to make your Information Management skill invaluable in AI projects. In this month’s episode: Specificity & Elephants Alan's kicked off the new year with a blog post "Navigating 2026: The Demand for Hyper-Specificity and the Persistent Paper Elephant", which is a result of him pondering about 2025 over the festive period. First is the need to continue to focus on the specifics of the outcomes from technology, rather that the technology horizontal markets that might provide them. Focus on the fixes, the bottlenecks in the real world. Second is that there remains a lot more paper in the world that technologists. Our own quantitative research data points not only to a lot of paper being used in processes but also to that amount of paper likely to increase (you can read about this in MMI reports that are available here and here). Matt reminds us that the sole superpower that analysts have is the ability to talk to people. Beyond the walled gardens Matt's published a pair of blog posts; at the tail end of last year "Beyond the walled gardens; can business applications break out to be the ultimate agentic orchestrators?" and just this month a follow-up "The lengthy hikes awaiting AI and automation in 2026". As he starts to update the data model for the 2026 edition of the Work Intelligence Market Analysis, he's pondering how AI-native, business application vendors and existing automation vendors are looking to develop their businesses against an tiny inflation in IT budgets (and big renewal numbers being demanded by the core suppliers). The pair also discuss maybe why there's been a lot of big M&A for security software firms lately. How to not lose the AI opportunity Finally, Alan wrote an Information Management focused blog post in December "How to not lose the opportunity that AI offers Information Management". Here he discusses the ways in which IM needs to inveigle itself into AI projects to be the good stewards of data; "own the data, own the data sourcing, the cleansing pipeline, and the metadata framework for the vector store".  https://www.youtube.com/watch?v=5ZKtMA72r5A  Related Links for Series 5 Episode 1 Alan's blog post "Navigating 2026: The Demand for Hyper-Specificity and the Persistent Paper Elephant" on the need for extreme specificity and that paper isn't going away. There a couple of MMI reports that talk about the amount of paper in contemporary processes; Market Momentum Index: Intelligent Document Processing (IDP) Survey 2025 Support the show

    35 min

  7. 12/13/2025

    S04E12 - 2026 Tech Predictions

    The thirty-fourth episode of the podcast that you know and love as "We Love Ugly Data!" is out; available in audio form everywhere you get your podcasts from and additionally in video form via YouTube (which we've embedded below). This month, it's time for the annual Deep Analysis predictions, so Matt is joined by both Alan and Dan to discuss a handful of those in our extensive predictions report and catch up on how well those made 12 months have panned out now that they've made contact with reality. In this month’s episode: Dan: 2025 Review and 2026 Prediction First up, Dan discusses his 2025 prediction, "To grow, intelligent document processing (IDP) companies must cross the border. He's giving himself a tick here, as he's seeing many vendors pushing the edges of both the vertical (industry) and horizontal (business function) markets that they've grown out of. For 2026, Dan is backing the prediction "The unstructured data gold rush will finally begin", which follows on from his 2025 prediction (and also Alan's 2025 prediction recapped below) that the growth in specific vertical and business functions for AI-derived technology like IDP and the broader AI agent market means a need for access to a lot more business data (and all the jeans, shovels and buckets that requires). Alan: 2025 Review and 2026 Prediction Alan starts with a review of his 2025 prediction "Structured data people will stop treating unstructured data like something that got stuck in their shoes" and gives himself a partial tick, in that he believes that there is general move from the big application vendors to recognise unstructured data is really useful for context, but also at the same time that it's really complicated to pick out the different data types within that big unstructured pile (and that's not just a technical challenge). For 2026, he's pitching "Edge computing will re-emerge as a strategic imperative"; specifically, that access to AI-derived applications increasingly need to take advantage of on-device processing, clever caching technology, etc., to enable remote workers (and related use cases) to operate smoothly. Matt: 2025 Review and 2026 Prediction Finally, it's Matt and he's keen up provide an update on his 2025 prediction "The shift to “payment on outcome” is going to lead to some awkward conversations between customers and suppliers". He explains that the business of developing consumption-based pricing models for AI agents has become increasingly complex, moving beyond the success criteria (especially Salesforce's per-conversation pricing, which remains but is joined by other options). In general, determining the economics of AI agent use remains in the early stages of being made easier to calculate. That notwithstanding, Matt is giving himself a tick here. For 2026, Matt is backing "'Doorstep adoption' of AI will be exposed as a counterproductive farce" as his chosen prediction; that vendors enforcing the bundling of AI tools into renewals doesn't make their adoption real. Matt suggests that it's akin to having a trailer (or a caravan if you prefer) welded to the back of your car without your permission, and then a bill arriving for the job for the excellent utility it provides.   https://www.youtube.com/watch?v=LulxRfzf8jk  Related Links for Series 4 Episode 12 Here's Alan's introductory blog for this year's predictions: It’s Time to Separate the Wine from the Hype. The download page for the full awards report (just a few details, nothing terrible). Here's Dan's blog post Support the show

    32 min

  8. 11/21/2025

    S04E11 - Innovation Award 2025 Special

    The thirty-third episode of the podcast that you know and love as "We Love Ugly Data!" is out; available in audio form everywhere you get your podcasts from and additionally in video form via YouTube (which we've embedded below). This month it's time to discuss the annual Deep Analysis Innovation Awards, so Matt is joined by both Alan and Dan to discuss this year's winners; Infrrd and LatticeFlow AI (but as ever, show notes are included just below the embedded video). In this month’s episode: Preamble: Why do we have Innovation Awards? First up, Alan talks about why Deep Analysis has annual Innovations Awards and why they are different to many other awards that are out there. Firstly, you can't apply for them, and if you win one there's no fee to pay, fancy dinner to attend (and buy a table at) or winners speech to rehearse. Also there's no set number of winners each year either; the team discusses which briefings wowed them and meets the awards criteria; Does it solve a problem? Does it apply ingenuity? Does it add value? Does it show flexibility? Winner: Infrrd Dan introduces the first award winner, Infrrd. Having initially annoyed him with their marketing choices, he was won over with the number of patents that have been awarded for their research in document AI and their development team deserves recognition for that work. Additionally, Dan was impressed with a demo involving Autocad drawings and data extraction, retrieval and document understanding (and awarded them a mini award of their own in the Intelligent Document Processing Market Analysis 2025-2028 report for best demo). Finally Dan was also impressed with Infrrd's work on mortgage loan files; a notoriously difficult use case. Alan adds that IDP in general is featuring in virtually every briefing that Deep Analysis receives right now, as it is a vital part of how unstructured data is made available for AI and agentic technology to use and it could be said it's what is driving many of the predominant AI use cases and is fundamental to LLMs themselves. Winner: LatticeFlow AI Matt then talks about second 2025 award winner, LatticeFlow AI. One of the features of 2025's discourse has been that AI projects are not making it out of pilots and into production. One of the reasons that is happening is that often the hurdle that is compliance, risk etc are not baked into the development from the beginning. LatticeFlow AI is focused on ensuring that an organizations AI applications meet both internal (think GRC; governance, risk and compliance) and external (regulations, like ISO, EU AI Act etc), with a focus both on internal data quality as well as how well individual models are likely to work in conjunction with it. Alan adds that that it's another areas that underlines the increasing gap between demos and production realities; as soon as the word governance people tend to switch off, but it's a vital part of ensuring that organizations are insulated against inevitable legal challenges coming down the road for organizations that don't realise its importance. Postamble: What are we looking for in 2026? To wrap things up, the team nominate what they hope they'll be seeing in 2026 that will end up with them pitching briefings in for the nominations pot in 12 month's time. Dan: Hoping to see better support for tabular information in IDP products; there's lots of start-ups working on the 2D table to understanding, some using Visual Language Models (VLM) and there's an expectation Support the show

    35 min
See All (64)

About

This irregular podcast series examines what is happening in the unstructured data automation market. Three topics - Thirty Minutes, that's the format!Topics range from the state of Blockchain, IDP, ECM, and the impact of AI on unstructured data. Deep Analysis provides advisory services, industry research, and M&A guidance. info@deep-analysis.netDeep-Analysis.Net

Information